Release Notes

v10.0.0

This is the first development release for the Jewel cycle.

Notable Changes

• build: cmake tweaks (pr#6254, John Spray)
• build: more CMake package check fixes (pr#6108, Daniel Gryniewicz)
• ceph-disk: get Nonetype when ceph-disk list with –format plain on single device. (pr#6410, Vicente Cheng)
• ceph: fix tell behavior (pr#6329, David Zafman)
• ceph-fuse: While starting ceph-fuse, start the log thread first (issue#13443, pr#6224, Wenjun Huang)
• client: don’t mark_down on command reply (pr#6204, John Spray)
• client: drop prefix from ints (pr#6275, John Coyle)
• client: sys/file.h includes for flock operations (pr#6282, John Coyle)
• cls_rbd: change object_map_update to return 0 on success, add logging (pr#6467, Douglas Fuller)
• cmake: Use uname instead of arch. (pr#6358, John Coyle)
• common: assert: __STRING macro is not defined by musl libc. (pr#6210, John Coyle)
• common: fix OpTracker age histogram calculation (pr#5065, Zhiqiang Wang)
• common/MemoryModel: Added explicit feature check for mallinfo(). (pr#6252, John Coyle)
• common/obj_bencher.cc: fix verification crashing when there’s no objects (pr#5853, Piotr Dałek)
• common: optimize debug logging (pr#6307, Adam Kupczyk)
• common: Thread: move copy constructor and assignment op (pr#5133, Michal Jarzabek)
• common: WorkQueue: new PointerWQ base class for ContextWQ (issue#13636, pr#6525, Jason Dillaman)
• compat: use prefixed typeof extension (pr#6216, John Coyle)
• crush: validate bucket id before indexing buckets array (issue#13477, pr#6246, Sage Weil)
• doc: download GPG key from download.ceph.com (issue#13603, pr#6384, Ken Dreyer)
• doc: fix outdated content in cache tier (pr#6272, Yuan Zhou)
• doc/release-notes: v9.1.0 (pr#6281, Loic Dachary)
• doc/releases-notes: fix build error (pr#6483, Kefu Chai)
• doc: remove toctree items under Create CephFS (pr#6241, Jevon Qiao)
• doc: rename the “Create a Ceph User” section and add verbage about… (issue#13502, pr#6297, ritz303)
• docs: Fix styling of newly added mirror docs (pr#6127, Wido den Hollander)
• doc, tests: update all http://ceph.com/ to download.ceph.com (pr#6435, Alfredo Deza)
• doc: update doc for with new pool settings (pr#5951, Guang Yang)
• doc: update radosgw-admin example (pr#6256, YankunLi)
• doc: update the OS recommendations for newer Ceph releases (pr#6355, ritz303)
• drop envz.h includes (pr#6285, John Coyle)
• libcephfs: Improve portability by replacing loff_t type usage with off_t (pr#6301, John Coyle)
• libcephfs: only check file offset on glibc platforms (pr#6288, John Coyle)
• librados: fix examples/librados/Makefile error. (pr#6320, You Ji)
• librados: init crush_location from config file. (issue#13473, pr#6243, Wei Luo)
• librados: wrongly passed in argument for stat command (issue#13703, pr#6476, xie xingguo)
• librbd: deadlock while attempting to flush AIO requests (issue#13726, pr#6508, Jason Dillaman)
• librbd: fix enable objectmap feature issue (issue#13558, pr#6339, xinxin shu)
• librbd: remove duplicate read_only test in librbd::async_flatten (pr#5856, runsisi)
• mailmap: modify member info (pr#6468, Xiaowei Chen)
• mailmap: updates (pr#6258, M Ranga Swami Reddy)
• mailmap: Xie Xingguo affiliation (pr#6409, Loic Dachary)
• mds: implement snapshot rename (pr#5645, xinxin shu)
• mds: messages/MOSDOp: cast in assert to eliminate warnings (issue#13625, pr#6414, David Zafman)
• mds: new filtered MDS tell commands for sessions (pr#6180, John Spray)
• mds/Session: use projected parent for auth path check (issue#13364, pr#6200, Sage Weil)
• mon: should not set isvalid = true when cephx_verify_authorizer return false (issue#13525, pr#6306, Ruifeng Yang)
• osd: Add config option osd_read_ec_check_for_errors for testing (pr#5865, David Zafman)
• osd: add pin/unpin support to cache tier (11066) (pr#6326, Zhiqiang Wang)
• osd: auto repair EC pool (issue#12754, pr#6196, Guang Yang)
• osd: drop the interim set from load_pgs() (pr#6277, Piotr Dałek)
• osd: FileJournal: _fdump wrongly returns if journal is currently unreadable. (issue#13626, pr#6406, xie xingguo)
• osd: FileStore: add a field indicate xattr only one chunk for set xattr. (pr#6244, Jianpeng Ma)
• osd: FileStore: LFNIndex: remove redundant local variable ‘obj’. (issue#13552, pr#6333, xiexingguo)
• osd: FileStore: potential memory leak if _fgetattrs fails (issue#13597, pr#6377, xie xingguo)
• osd: FileStore: remove unused local variable ‘handle’ (pr#6381, xie xingguo)
• osd: fix bogus scrub results when missing a clone (issue#12738, issue#12740, pr#5783, David Zafman)
• osd: fix debug message in OSD::is_healthy (pr#6226, Xiaoxi Chen)
• osd: fix MOSDOp encoding (pr#6174, Sage Weil)
• osd: init started to 0 (issue#13206, pr#6107, Sage Weil)
• osd: KeyValueStore: fix the name’s typo of keyvaluestore_default_strip_size (pr#6375, Zhi Zhang)
• osd: new and delete ObjectStore::Transaction in a function is not necessary (pr#6299, Ruifeng Yang)
• osd: optimize get_object_context (pr#6305, Jianpeng Ma)
• osd: optimize MOSDOp/do_op/handle_op (pr#5211, Jacek J. Lakis)
• osd: os/chain_xattr: On linux use linux/limits.h for XATTR_NAME_MAX. (pr#6343, John Coyle)
• osd: reorder bool fields in PGLog struct (pr#6279, Piotr Dałek)
• osd: ReplicatedPG: remove unused local variables (issue#13575, pr#6360, xiexingguo)
• osd: reset primary and up_primary when building a new past_interval. (issue#13471, pr#6240, xiexingguo)
• radosgw-admin: Checking the legality of the parameters (issue#13018, pr#5879, Qiankun Zheng)
• radosgw-admin: Create –secret-key alias for –secret (issue#5821, pr#5335, Yuan Zhou)
• radosgw-admin: metadata list user should return an empty list when user pool is empty (issue#13596, pr#6465, Orit Wasserman)
• rados: new options for write benchmark (pr#6340, Joaquim Rocha)
• rbd: fix clone isssue (issue#13553, pr#6334, xinxin shu)
• rbd: fix init-rbdmap CMDPARAMS (issue#13214, pr#6109, Sage Weil)
• rbdmap: systemd support (issue#13374, pr#6479, Boris Ranto)
• rbd: rbdmap improvements (pr#6445, Boris Ranto)
• release-notes: draft v0.94.4 release notes (pr#5907, Loic Dachary)
• release-notes: draft v0.94.4 release notes (pr#6195, Loic Dachary)
• release-notes: draft v0.94.4 release notes (pr#6238, Loic Dachary)
• rgw: add compat header for TEMP_FAILURE_RETRY (pr#6294, John Coyle)
• rgw: add default quota config (pr#6400, Daniel Gryniewicz)
• rgw: add support for getting Swift’s DLO without manifest handling (pr#6206, Radoslaw Zarzynski)
• rgw: clarify the error message when trying to create an existed user (pr#5938, Zeqiang Zhuang)
• rgw: fix objects can not be displayed which object name does not cont… (issue#12963, pr#5738, Weijun Duan)
• rgw: fix typo in RGWHTTPClient::process error message (pr#6424, Brad Hubbard)
• rgw: fix wrong etag calculation during POST on S3 bucket. (issue#11241, pr#6030, Radoslaw Zarzynski)
• rgw: mdlog trim add usage prompt (pr#6059, Weijun Duan)
• rgw: modify the conditional statement in parse_metadata_key method. (pr#5875, Zengran Zhang)
• rgw: refuse to calculate digest when the s3 secret key is empty (issue#13133, pr#6045, Sangdi Xu)
• rgw: remove extra check in RGWGetObj::execute (issue#12352, pr#5262, Javier M. Mellid)
• rgw: support core file limit for radosgw daemon (pr#6346, Guang Yang)
• rgw: swift use Civetweb ssl can not get right url (issue#13628, pr#6408, Weijun Duan)
• rocksdb: build with PORTABLE=1 (pr#6311, Sage Weil)
• rocksdb: remove rdb source files from dist tarball (issue#13554, pr#6379, Kefu Chai)
• rocksdb: use native rocksdb makefile (and our autotools) (pr#6290, Sage Weil)
• rpm: ceph.spec.in: correctly declare systemd dependency for SLE/openSUSE (pr#6114, Nathan Cutler)
• rpm: ceph.spec.in: fix libs-compat / devel-compat conditional (issue#12315, pr#5219, Ken Dreyer)
• rpm: rhel 5.9 librados compile fix, moved blkid to RBD check/compilation (issue#13177, pr#5954, Rohan Mars)
• scripts: release_notes can track original issue (pr#6009, Abhishek Lekshmanan)
• test/libcephfs/flock: add sys/file.h include for flock operations (pr#6310, John Coyle)
• test_rgw_admin: use freopen for output redirection. (pr#6303, John Coyle)
• tests: allow docker-test.sh to run under root (issue#13355, pr#6173, Loic Dachary)
• tests: ceph-disk workunit uses configobj (pr#6342, Loic Dachary)
• tests: destroy testprofile before creating one (issue#13664, pr#6446, Loic Dachary)
• tests: port uniqueness reminder (pr#6387, Loic Dachary)
• tests: test/librados/test.cc must create profile (issue#13664, pr#6452, Loic Dachary)
• tools/cephfs: fix overflow writing header to fixed size buffer (#13816) (pr#6617, John Spray)
• tools: ceph-monstore-update-crush: add “–test” when testing crushmap (pr#6418, Kefu Chai)
• tools:remove duplicate references (pr#5917, Bo Cai)
• vstart: grant full access to Swift testing account (pr#6239, Yuan Zhou)
• vstart: set cephfs root uid/gid to caller (pr#6255, John Spray)

v9.2.0 Infernalis

This major release will be the foundation for the next stable series.There have been some major changes since v0.94.x Hammer, and theupgrade process is non-trivial. Please read these release notes carefully.

Major Changes from Hammer

• General:
  • Ceph daemons are now managed via systemd (with the exception ofUbuntu Trusty, which still uses upstart).
  • Ceph daemons run as ‘ceph’ user instead root.
  • On Red Hat distros, there is also an SELinux policy.
• RADOS:
  • The RADOS cache tier can now proxy write operations to the basetier, allowing writes to be handled without forcing migration ofan object into the cache.
  • The SHEC erasure coding support is no longer flagged asexperimental. SHEC trades some additional storage space for fasterrepair.
  • There is now a unified queue (and thus prioritization) of clientIO, recovery, scrubbing, and snapshot trimming.
  • There have been many improvements to low-level repair tooling(ceph-objectstore-tool).
  • The internal ObjectStore API has been significantly cleaned up in orderto faciliate new storage backends like NewStore.
• RGW:
  • The Swift API now supports object expiration.
  • There are many Swift API compatibility improvements.
• RBD:
  • The rbddu command shows actual usage (quickly, whenobject-map is enabled).
  • The object-map feature has seen many stability improvements.
  • Object-map and exclusive-lock features can be enabled or disableddynamically.
  • You can now store user metadata and set persistent librbd optionsassociated with individual images.
  • The new deep-flatten features allows flattening of a clone and allof its snapshots. (Previously snapshots could not be flattened.)
  • The export-diff command command is now faster (it uses aio). There is alsoa new fast-diff feature.
  • The –size argument can be specified with a suffix for units(e.g., —size64G).
  • There is a new rbdstatus command that, for now, shows who hasthe image open/mapped.
• CephFS:
  • You can now rename snapshots.
  • There have been ongoing improvements around administration, diagnostics,and the check and repair tools.
  • The caching and revocation of client cache state due to unusedinodes has been dramatically improved.
  • The ceph-fuse client behaves better on 32-bit hosts.

Distro compatibility

We have decided to drop support for many older distributions so that we canmove to a newer compiler toolchain (e.g., C++11). Although it is still possibleto build Ceph on older distributions by installing backported development tools,we are not building and publishing release packages for ceph.com.

We now build packages for:

• CentOS 7 or later. We have dropped support for CentOS 6 (and otherRHEL 6 derivatives, like Scientific Linux 6).
• Debian Jessie 8.x or later. Debian Wheezy 7.x’s g++ has incompletesupport for C++11 (and no systemd).
• Ubuntu Trusty 14.04 or later. Ubuntu Precise 12.04 is no longersupported.
• Fedora 22 or later.

Upgrading from Firefly

Upgrading directly from Firefly v0.80.z is not recommended. It ispossible to do a direct upgrade, but not without downtime. Werecommend that clusters are first upgraded to Hammer v0.94.4 or alater v0.94.z release; only then is it possible to upgrade toInfernalis 9.2.z for an online upgrade (see below).

To do an offline upgrade directly from Firefly, all Firefly OSDs mustbe stopped and marked down before any Infernalis OSDs will be allowedto start up. This fencing is enforced by the Infernalis monitor, souse an upgrade procedure like:

1. Upgrade Ceph on monitor hosts

2. Restart all ceph-mon daemons

3. Upgrade Ceph on all OSD hosts

4. Stop all ceph-osd daemons

5. Mark all OSDs down with something like::

   ceph osd down seq 0 1000

6. Start all ceph-osd daemons

7. Upgrade and restart remaining daemons (ceph-mds, radosgw)

Upgrading from Hammer

• All cluster nodes must first upgrade to Hammer v0.94.4 or a later v0.94.z release; onlythen is it possible to upgrade to Infernalis 9.2.z.

• For all distributions that support systemd (CentOS 7, Fedora, DebianJessie 8.x, OpenSUSE), ceph daemons are now managed using native systemdfiles instead of the legacy sysvinit scripts. For example,:

systemctl start ceph.target       # start all daemons
systemctl status ceph-osd@12      # check status of osd.12

The main notable distro that is not yet using systemd is Ubuntu trusty14.04. (The next Ubuntu LTS, 16.04, will use systemd instead of upstart.)

• Ceph daemons now run as user and group ceph by default. Theceph user has a static UID assigned by Fedora and Debian (also usedby derivative distributions like RHEL/CentOS and Ubuntu). On SUSEthe ceph user will currently get a dynamically assigned UID when theuser is created.

If your systems already have a ceph user, upgrading the package will causeproblems. We suggest you first remove or rename the existing ‘ceph’ userand ‘ceph’ group before upgrading.

When upgrading, administrators have two options:

1. Add the following line to ceph.conf on all hosts:

   setuser match path = /var/lib/ceph/$type/$cluster-$id

   This will make the Ceph daemons run as root (i.e., not dropprivileges and switch to user ceph) if the daemon’s datadirectory is still owned by root. Newly deployed daemons willbe created with data owned by user ceph and will run withreduced privileges, but upgraded daemons will continue to run asroot.

2. Fix the data ownership during the upgrade. This is thepreferred option, but it is more work and can be very timeconsuming. The process for each host is to:

   1. Upgrade the ceph package. This creates the ceph user and group. Forexample:

      ceph-deploy install —stable infernalis HOST

   2. Stop the daemon(s).:

      service ceph stop           # fedora, centos, rhel, debianstop ceph-all               # ubuntu

   3. Fix the ownership:

      chown -R ceph:ceph /var/lib/ceph

   4. Restart the daemon(s).:

      start ceph-all                # ubuntusystemctl start ceph.target   # debian, centos, fedora, rhel

   Alternatively, the same process can be done with a single daemontype, for example by stopping only monitors and chowning only/var/lib/ceph/mon.

• The on-disk format for the experimental KeyValueStore OSD backend haschanged. You will need to remove any OSDs using that backend before youupgrade any test clusters that use it.

• When a pool quota is reached, librados operations now block indefinitely,the same way they do when the cluster fills up. (Previously they would return-ENOSPC). By default, a full cluster or pool will now block. If yourlibrados application can handle ENOSPC or EDQUOT errors gracefully, you canget error returns instead by using the new librados OPERATION_FULL_TRY flag.

• The return code for librbd’s rbd_aio_read and Image::aio_read API methods nolonger returns the number of bytes read upon success. Instead, it returns 0upon success and a negative value upon failure.

• ‘ceph scrub’, ‘ceph compact’ and ‘ceph sync force are now DEPRECATED. Usersshould instead use ‘ceph mon scrub’, ‘ceph mon compact’ and‘ceph mon sync force’.

• ‘ceph mon_metadata’ should now be used as ‘ceph mon metadata’. There is noneed to deprecate this command (same major release since it was firstintroduced).

• The –dump-json option of “osdmaptool” is replaced by –dump json.

• The commands of “pg ls-by-{pool,primary,osd}” and “pg ls” now take “recovering”instead of “recovery”, to include the recovering pgs in the listed pgs.

Notable Changes since Hammer

• aarch64: add optimized version of crc32c (Yazen Ghannam, Steve Capper)
• auth: cache/reuse crypto lib key objects, optimize msg signature check (Sage Weil)
• auth: reinit NSS after fork() (#11128 Yan, Zheng)
• autotools: fix out of tree build (Krxysztof Kosinski)
• autotools: improve make check output (Loic Dachary)
• buffer: add invalidate_crc() (Piotr Dalek)
• buffer: fix zero bug (#12252 Haomai Wang)
• buffer: some cleanup (Michal Jarzabek)
• build: allow tcmalloc-minimal (Thorsten Behrens)
• build: C++11 now supported
• build: cmake: fix nss linking (Danny Al-Gaaf)
• build: cmake: misc fixes (Orit Wasserman, Casey Bodley)
• build: disable LTTNG by default (#11333 Josh Durgin)
• build: do not build ceph-dencoder with tcmalloc (#10691 Boris Ranto)
• build: fix junit detection on Fedora 22 (Ira Cooper)
• build: fix pg ref disabling (William A. Kennington III)
• build: fix ppc build (James Page)
• build: install-deps: misc fixes (Loic Dachary)
• build: install-deps.sh improvements (Loic Dachary)
• build: install-deps: support OpenSUSE (Loic Dachary)
• build: make_dist_tarball.sh (Sage Weil)
• build: many cmake improvements
• build: misc cmake fixes (Matt Benjamin)
• build: misc fixes (Boris Ranto, Ken Dreyer, Owen Synge)
• build: OSX build fixes (Yan, Zheng)
• build: remove rest-bench
• ceph-authtool: fix return code on error (Gerhard Muntingh)
• ceph-detect-init: added Linux Mint (Michal Jarzabek)
• ceph-detect-init: robust init system detection (Owen Synge)
• ceph-disk: ensure ‘zap’ only operates on a full disk (#11272 Loic Dachary)
• ceph-disk: fix zap sgdisk invocation (Owen Synge, Thorsten Behrens)
• ceph-disk: follow ceph-osd hints when creating journal (#9580 Sage Weil)
• ceph-disk: handle re-using existing partition (#10987 Loic Dachary)
• ceph-disk: improve parted output parsing (#10983 Loic Dachary)
• ceph-disk: install pip > 6.1 (#11952 Loic Dachary)
• ceph-disk: make suppression work for activate-all and activate-journal (Dan van der Ster)
• ceph-disk: many fixes (Loic Dachary, Alfredo Deza)
• ceph-disk: fixes to respect init system (Loic Dachary, Owen Synge)
• ceph-disk: pass –cluster arg on prepare subcommand (Kefu Chai)
• ceph-disk: support for multipath devices (Loic Dachary)
• ceph-disk: support NVMe device partitions (#11612 Ilja Slepnev)
• ceph: fix ‘df’ units (Zhe Zhang)
• ceph: fix parsing in interactive cli mode (#11279 Kefu Chai)
• cephfs-data-scan: many additions, improvements (John Spray)
• ceph-fuse: do not require successful remount when unmounting (#10982 Greg Farnum)
• ceph-fuse, libcephfs: don’t clear COMPLETE when trimming null (Yan, Zheng)
• ceph-fuse, libcephfs: drop inode when rmdir finishes (#11339 Yan, Zheng)
• ceph-fuse,libcephfs: fix uninline (#11356 Yan, Zheng)
• ceph-fuse, libcephfs: hold exclusive caps on dirs we “own” (#11226 Greg Farnum)
• ceph-fuse: mostly behave on 32-bit hosts (Yan, Zheng)
• ceph: improve error output for ‘tell’ (#11101 Kefu Chai)
• ceph-monstore-tool: fix store-copy (Huangjun)
• ceph: new ‘ceph daemonperf’ command (John Spray, Mykola Golub)
• ceph-objectstore-tool: many many improvements (David Zafman)
• ceph-objectstore-tool: refactoring and cleanup (John Spray)
• ceph-post-file: misc fixes (Joey McDonald, Sage Weil)
• ceph_test_rados: test pipelined reads (Zhiqiang Wang)
• client: avoid sending unnecessary FLUSHSNAP messages (Yan, Zheng)
• client: exclude setfilelock when calculating oldest tid (Yan, Zheng)
• client: fix error handling in check_pool_perm (John Spray)
• client: fsync waits only for inode’s caps to flush (Yan, Zheng)
• client: invalidate kernel dcache when cache size exceeds limits (Yan, Zheng)
• client: make fsync wait for unsafe dir operations (Yan, Zheng)
• client: pin lookup dentry to avoid inode being freed (Yan, Zheng)
• common: add descriptions to perfcounters (Kiseleva Alyona)
• common: add perf counter descriptions (Alyona Kiseleva)
• common: bufferlist performance tuning (Piotr Dalek, Sage Weil)
• common: detect overflow of int config values (#11484 Kefu Chai)
• common: fix bit_vector extent calc (#12611 Jason Dillaman)
• common: fix json parsing of utf8 (#7387 Tim Serong)
• common: fix leak of pthread_mutexattr (#11762 Ketor Meng)
• common: fix LTTNG vs fork issue (Josh Durgin)
• common: fix throttle max change (Henry Chang)
• common: make mutex more efficient
• common: make work queue addition/removal thread safe (#12662 Jason Dillaman)
• common: optracker improvements (Zhiqiang Wang, Jianpeng Ma)
• common: PriorityQueue tests (Kefu Chai)
• common: some async compression infrastructure (Haomai Wang)
• crush: add –check to validate dangling names, max osd id (Kefu Chai)
• crush: cleanup, sync with kernel (Ilya Dryomov)
• crush: fix crash from invalid ‘take’ argument (#11602 Shiva Rkreddy, Sage Weil)
• crush: fix divide-by-2 in straw2 (#11357 Yann Dupont, Sage Weil)
• crush: fix has_v4_buckets (#11364 Sage Weil)
• crush: fix subtree base weight on adjust_subtree_weight (#11855 Sage Weil)
• crush: respect default replicated ruleset config on map creation (Ilya Dryomov)
• crushtool: fix order of operations, usage (Sage Weil)
• crypto: fix NSS leak (Jason Dillaman)
• crypto: fix unbalanced init/shutdown (#12598 Zheng Yan)
• deb: fix rest-bench-dbg and ceph-test-dbg dependendies (Ken Dreyer)
• debian: minor package reorg (Ken Dreyer)
• deb, rpm: move ceph-objectstore-tool to ceph (Ken Dreyer)
• doc: docuemnt object corpus generation (#11099 Alexis Normand)
• doc: document region hostnames (Robin H. Johnson)
• doc: fix gender neutrality (Alexandre Maragone)
• doc: fix install doc (#10957 Kefu Chai)
• doc: fix sphinx issues (Kefu Chai)
• doc: man page updates (Kefu Chai)
• doc: mds data structure docs (Yan, Zheng)
• doc: misc updates (Fracois Lafont, Ken Dreyer, Kefu Chai, Owen Synge, Gael Fenet-Garde, Loic Dachary, Yannick Atchy-Dalama, Jiaying Ren, Kevin Caradant, Robert Maxime, Nicolas Yong, Germain Chipaux, Arthur Gorjux, Gabriel Sentucq, Clement Lebrun, Jean-Remi Deveaux, Clair Massot, Robin Tang, Thomas Laumondais, Jordan Dorne, Yuan Zhou, Valentin Thomas, Pierre Chaumont, Benjamin Troquereau, Benjamin Sesia, Vikhyat Umrao, Nilamdyuti Goswami, Vartika Rai, Florian Haas, Loic Dachary, Simon Guinot, Andy Allan, Alistair Israel, Ken Dreyer, Robin Rehu, Lee Revell, Florian Marsylle, Thomas Johnson, Bosse Klykken, Travis Rhoden, Ian Kelling)
• doc: swift tempurls (#10184 Abhishek Lekshmanan)
• doc: switch doxygen integration back to breathe (#6115 Kefu Chai)
• doc: update release schedule docs (Loic Dachary)
• erasure-code: cleanup (Kefu Chai)
• erasure-code: improve tests (Loic Dachary)
• erasure-code: shec: fix recovery bugs (Takanori Nakao, Shotaro Kawaguchi)
• erasure-code: update ISA-L to 2.13 (Yuan Zhou)
• gmock: switch to submodule (Danny Al-Gaaf, Loic Dachary)
• hadoop: add terasort test (Noah Watkins)
• init-radosgw: merge with sysv version; fix enumeration (Sage Weil)
• java: fix libcephfs bindings (Noah Watkins)
• libcephfs: add pread, pwrite (Jevon Qiao)
• libcephfs,ceph-fuse: cache cleanup (Zheng Yan)
• libcephfs,ceph-fuse: fix request resend on cap reconnect (#10912 Yan, Zheng)
• librados: add config observer (Alistair Strachan)
• librados: add FULL_TRY and FULL_FORCE flags for dealing with full clusters or pools (Sage Weil)
• librados: add src_fadvise_flags for copy-from (Jianpeng Ma)
• librados: define C++ flags from C constants (Josh Durgin)
• librados: fadvise flags per op (Jianpeng Ma)
• librados: fix last_force_resent handling (#11026 Jianpeng Ma)
• librados: fix memory leak from C_TwoContexts (Xiong Yiliang)
• librados: fix notify completion race (#13114 Sage Weil)
• librados: fix striper when stripe_count = 1 and stripe_unit != object_size (#11120 Yan, Zheng)
• librados, libcephfs: randomize client nonces (Josh Durgin)
• librados: op perf counters (John Spray)
• librados: pybind: fix binary omap values (Robin H. Johnson)
• librados: pybind: fix write() method return code (Javier Guerra)
• librados: respect default_crush_ruleset on pool_create (#11640 Yuan Zhou)
• libradosstriper: fix leak (Danny Al-Gaaf)
• librbd: add const for single-client-only features (Josh Durgin)
• librbd: add deep-flatten operation (Jason Dillaman)
• librbd: add purge_on_error cache behavior (Jianpeng Ma)
• librbd: allow additional metadata to be stored with the image (Haomai Wang)
• librbd: avoid blocking aio API methods (#11056 Jason Dillaman)
• librbd: better handling for dup flatten requests (#11370 Jason Dillaman)
• librbd: cancel in-flight ops on watch error (#11363 Jason Dillaman)
• librbd: default new images to format 2 (#11348 Jason Dillaman)
• librbd: fadvise for copy, export, import (Jianpeng Ma)
• librbd: fast diff implementation that leverages object map (Jason Dillaman)
• librbd: fix fast diff bugs (#11553 Jason Dillaman)
• librbd: fix image format detection (Zhiqiang Wang)
• librbd: fix lock ordering issue (#11577 Jason Dillaman)
• librbd: fix reads larger than the cache size (Lu Shi)
• librbd: fix snapshot creation when other snap is active (#11475 Jason Dillaman)
• librbd: flatten/copyup fixes (Jason Dillaman)
• librbd: handle NOCACHE fadvise flag (Jinapeng Ma)
• librbd: lockdep, helgrind validation (Jason Dillaman, Josh Durgin)
• librbd: metadata filter fixes (Haomai Wang)
• librbd: misc aio fixes (#5488 Jason Dillaman)
• librbd: misc rbd fixes (#11478 #11113 #11342 #11380 Jason Dillaman, Zhiqiang Wang)
• librbd: new diff_iterate2 API (Jason Dillaman)
• librbd: object map rebuild support (Jason Dillaman)
• librbd: only update image flags while hold exclusive lock (#11791 Jason Dillaman)
• librbd: optionally disable allocation hint (Haomai Wang)
• librbd: prevent race between resize requests (#12664 Jason Dillaman)
• librbd: readahead fixes (Zhiqiang Wang)
• librbd: return result code from close (#12069 Jason Dillaman)
• librbd: store metadata, including config options, in image (Haomai Wang)
• librbd: tolerate old osds when getting image metadata (#11549 Jason Dillaman)
• librbd: use write_full when possible (Zhiqiang Wang)
• log: fix data corruption race resulting from log rotation (#12465 Samuel Just)
• logrotate.d: prefer service over invoke-rc.d (#11330 Win Hierman, Sage Weil)
• mds: add ‘damaged’ state to MDSMap (John Spray)
• mds: add nicknames for perfcounters (John Spray)
• mds: avoid emitting cap warnigns before evicting session (John Spray)
• mds: avoid getting stuck in XLOCKDONE (#11254 Yan, Zheng)
• mds: disable problematic rstat propagation into snap parents (Yan, Zheng)
• mds: do not add snapped items to bloom filter (Yan, Zheng)
• mds: expose frags via asok (John Spray)
• mds: fix expected holes in journal objects (#13167 Yan, Zheng)
• mds: fix handling for missing mydir dirfrag (#11641 John Spray)
• mds: fix integer truncateion on large client ids (Henry Chang)
• mds: fix mydir replica issue with shutdown (#10743 John Spray)
• mds: fix out-of-order messages (#11258 Yan, Zheng)
• mds: fix rejoin (Yan, Zheng)
• mds: fix setting entire file layout in one setxattr (John Spray)
• mds: fix shutdown (John Spray)
• mds: fix shutdown with strays (#10744 John Spray)
• mds: fix SnapServer crash on deleted pool (John Spray)
• mds: fix snapshot bugs (Yan, Zheng)
• mds: fix stray reintegration (Yan, Zheng)
• mds: fix stray handling (John Spray)
• mds: fix suicide beacon (John Spray)
• mds: flush immediately in do_open_truncate (#11011 John Spray)
• mds: handle misc corruption issues (John Spray)
• mds: improve dump methods (John Spray)
• mds: many fixes (Yan, Zheng, John Spray, Greg Farnum)
• mds: many snapshot and stray fixes (Yan, Zheng)
• mds: misc fixes (Jianpeng Ma, Dan van der Ster, Zhang Zhi)
• mds: misc journal cleanups and fixes (#10368 John Spray)
• mds: misc repair improvements (John Spray)
• mds: misc snap fixes (Zheng Yan)
• mds: misc snapshot fixes (Yan, Zheng)
• mds: new SessionMap storage using omap (#10649 John Spray)
• mds: persist completed_requests reliably (#11048 John Spray)
• mds: reduce memory consumption (Yan, Zheng)
• mds: respawn instead of suicide on blacklist (John Spray)
• mds: separate safe_pos in Journaler (#10368 John Spray)
• mds: snapshot rename support (#3645 Yan, Zheng)
• mds: store layout on header object (#4161 John Spray)
• mds: throttle purge stray operations (#10390 John Spray)
• mds: tolerate clock jumping backwards (#11053 Yan, Zheng)
• mds: warn when clients fail to advance oldest_client_tid (#10657 Yan, Zheng)
• misc cleanups and fixes (Danny Al-Gaaf)
• misc coverity fixes (Danny Al-Gaaf)
• misc performance and cleanup (Nathan Cutler, Xinxin Shu)
• mon: add cache over MonitorDBStore (Kefu Chai)
• mon: add ‘mon_metadata ’ command (Kefu Chai)
• mon: add ‘node ls …’ command (Kefu Chai)
• mon: add NOFORWARD, OBSOLETE, DEPRECATE flags for mon commands (Joao Eduardo Luis)
• mon: add PG count to ‘ceph osd df’ output (Michal Jarzabek)
• mon: ‘ceph osd metadata’ can dump all osds (Haomai Wang)
• mon: clean up, reorg some mon commands (Joao Eduardo Luis)
• monclient: flush_log (John Spray)
• mon: detect kv backend failures (Sage Weil)
• mon: disallow >2 tiers (#11840 Kefu Chai)
• mon: disallow ec pools as tiers (#11650 Samuel Just)
• mon: do not deactivate last mds (#10862 John Spray)
• mon: fix average utilization calc for ‘osd df’ (Mykola Golub)
• mon: fix CRUSH map test for new pools (Sage Weil)
• mon: fix log dump crash when debugging (Mykola Golub)
• mon: fix mds beacon replies (#11590 Kefu Chai)
• mon: fix metadata update race (Mykola Golub)
• mon: fix min_last_epoch_clean tracking (Kefu Chai)
• mon: fix ‘pg ls’ sort order, state names (#11569 Kefu Chai)
• mon: fix refresh (#11470 Joao Eduardo Luis)
• mon: fix variance calc in ‘osd df’ (Sage Weil)
• mon: improve callout to crushtool (Mykola Golub)
• mon: make blocked op messages more readable (Jianpeng Ma)
• mon: make osd get pool ‘all’ only return applicable fields (#10891 Michal Jarzabek)
• mon: misc scaling fixes (Sage Weil)
• mon: normalize erasure-code profile for storage and comparison (Loic Dachary)
• mon: only send mon metadata to supporting peers (Sage Weil)
• mon: optionally specify osd id on ‘osd create’ (Mykola Golub)
• mon: ‘osd tree’ fixes (Kefu Chai)
• mon: periodic background scrub (Joao Eduardo Luis)
• mon: prevent bucket deletion when referenced by a crush rule (#11602 Sage Weil)
• mon: prevent pgp_num > pg_num (#12025 Xinxin Shu)
• mon: prevent pool with snapshot state from being used as a tier (#11493 Sage Weil)
• mon: prime pg_temp when CRUSH map changes (Sage Weil)
• mon: refine check_remove_tier checks (#11504 John Spray)
• mon: reject large max_mds values (#12222 John Spray)
• mon: remove spurious who arg from ‘mds rm …’ (John Spray)
• mon: streamline session handling, fix memory leaks (Sage Weil)
• mon: upgrades must pass through hammer (Sage Weil)
• mon: warn on bogus cache tier config (Jianpeng Ma)
• msgr: add ceph_perf_msgr tool (Hoamai Wang)
• msgr: async: fix seq handling (Haomai Wang)
• msgr: async: many many fixes (Haomai Wang)
• msgr: simple: fix clear_pipe (#11381 Haomai Wang)
• msgr: simple: fix connect_seq assert (Haomai Wang)
• msgr: xio: fastpath improvements (Raju Kurunkad)
• msgr: xio: fix ip and nonce (Raju Kurunkad)
• msgr: xio: improve lane assignment (Vu Pham)
• msgr: xio: sync with accellio v1.4 (Vu Pham)
• msgr: xio: misc fixes (#10735 Matt Benjamin, Kefu Chai, Danny Al-Gaaf, Raju Kurunkad, Vu Pham, Casey Bodley)
• msg: unit tests (Haomai Wang)
• objectcacher: misc bug fixes (Jianpeng Ma)
• osd: add latency perf counters for tier operations (Xinze Chi)
• osd: add misc perfcounters (Xinze Chi)
• osd: add simple sleep injection in recovery (Sage Weil)
• osd: allow SEEK_HOLE/SEEK_DATA for sparse read (Zhiqiang Wang)
• osd: avoid dup omap sets for in pg metadata (Sage Weil)
• osd: avoid multiple hit set insertions (Zhiqiang Wang)
• osd: avoid transaction append in some cases (Sage Weil)
• osd: break PG removal into multiple iterations (#10198 Guang Yang)
• osd: cache proxy-write support (Zhiqiang Wang, Samuel Just)
• osd: check scrub state when handling map (Jianpeng Ma)
• osd: clean up some constness, privateness (Kefu Chai)
• osd: clean up temp object if promotion fails (Jianpeng Ma)
• osd: configure promotion based on write recency (Zhiqiang Wang)
• osd: constrain collections to meta and PGs (normal and temp) (Sage Weil)
• osd: don’t send dup MMonGetOSDMap requests (Sage Weil, Kefu Chai)
• osd: EIO injection (David Zhang)
• osd: elminiate txn apend, ECSubWrite copy (Samuel Just)
• osd: erasure-code: drop entries according to LRU (Andreas-Joachim Peters)
• osd: erasure-code: fix SHEC floating point bug (#12936 Loic Dachary)
• osd: erasure-code: update to ISA-L 2.14 (Yuan Zhou)
• osd: filejournal: cleanup (David Zafman)
• osd: filestore: clone using splice (Jianpeng Ma)
• osd: filestore: fix recursive lock (Xinxin Shu)
• osd: fix check_for_full (Henry Chang)
• osd: fix dirty accounting in make_writeable (Zhiqiang Wang)
• osd: fix dup promotion lost op bug (Zhiqiang Wang)
• osd: fix endless repair when object is unrecoverable (Jianpeng Ma, Kefu Chai)
• osd: fix hitset object naming to use GMT (Kefu Chai)
• osd: fix misc memory leaks (Sage Weil)
• osd: fix negative degraded stats during backfill (Guang Yang)
• osd: fix osdmap dump of blacklist items (John Spray)
• osd: fix peek_queue locking in FileStore (Xinze Chi)
• osd: fix pg resurrection (#11429 Samuel Just)
• osd: fix promotion vs full cache tier (Samuel Just)
• osd: fix replay requeue when pg is still activating (#13116 Samuel Just)
• osd: fix scrub stat bugs (Sage Weil, Samuel Just)
• osd: fix snap flushing from cache tier (again) (#11787 Samuel Just)
• osd: fix snap handling on promotion (#11296 Sam Just)
• osd: fix temp-clearing (David Zafman)
• osd: force promotion for ops EC can’t handle (Zhiqiang Wang)
• osd: handle log split with overlapping entries (#11358 Samuel Just)
• osd: ignore non-existent osds in unfound calc (#10976 Mykola Golub)
• osd: improve behavior on machines with large memory pages (Steve Capper)
• osd: include a temp namespace within each collection/pgid (Sage Weil)
• osd: increase default max open files (Owen Synge)
• osd: keyvaluestore: misc fixes (Varada Kari)
• osd: low and high speed flush modes (Mingxin Liu)
• osd: make suicide timeouts individually configurable (Samuel Just)
• osd: merge multiple setattr calls into a setattrs call (Xinxin Shu)
• osd: misc fixes (Ning Yao, Kefu Chai, Xinze Chi, Zhiqiang Wang, Jianpeng Ma)
• osd: move scrub in OpWQ (Samuel Just)
• osd: newstore prototype (Sage Weil)
• osd: ObjectStore internal API refactor (Sage Weil)
• osd: peer_features includes self (David Zafman)
• osd: pool size change triggers new interval (#11771 Samuel Just)
• osd: prepopulate needs_recovery_map when only one peer has missing (#9558 Guang Yang)
• osd: randomize scrub times (#10973 Kefu Chai)
• osd: recovery, peering fixes (#11687 Samuel Just)
• osd: refactor scrub and digest recording (Sage Weil)
• osd: refuse first write to EC object at non-zero offset (Jianpeng Ma)
• osd: relax reply order on proxy read (#11211 Zhiqiang Wang)
• osd: require firefly features (David Zafman)
• osd: set initial crush weight with more precision (Sage Weil)
• osd: SHEC no longer experimental
• osd: skip promotion for flush/evict op (Zhiqiang Wang)
• osd: stripe over small xattrs to fit in XFS’s 255 byte inline limit (Sage Weil, Ning Yao)
• osd: sync object_map on syncfs (Samuel Just)
• osd: take excl lock of op is rw (Samuel Just)
• osd: throttle evict ops (Yunchuan Wen)
• osd: upgrades must pass through hammer (Sage Weil)
• osd: use a temp object for recovery (Sage Weil)
• osd: use blkid to collection partition information (Joseph Handzik)
• osd: use SEEK_HOLE / SEEK_DATA for sparse copy (Xinxin Shu)
• osd: WBThrottle cleanups (Jianpeng Ma)
• osd: write journal header on clean shutdown (Xinze Chi)
• osdc/Objecter: allow per-pool calls to op_cancel_writes (John Spray)
• os/filestore: enlarge getxattr buffer size (Jianpeng Ma)
• pybind: pep8 cleanups (Danny Al-Gaaf)
• pycephfs: many fixes for bindings (Haomai Wang)
• qa: fix filelock_interrupt.py test (Yan, Zheng)
• qa: improve ceph-disk tests (Loic Dachary)
• qa: improve docker build layers (Loic Dachary)
• qa: run-make-check.sh script (Loic Dachary)
• rados: add –striper option to use libradosstriper (#10759 Sebastien Ponce)
• rados: bench: add –no-verify option to improve performance (Piotr Dalek)
• rados bench: misc fixes (Dmitry Yatsushkevich)
• rados: fix error message on failed pool removal (Wido den Hollander)
• radosgw-admin: add ‘bucket check’ function to repair bucket index (Yehuda Sadeh)
• radosgw-admin: fix subuser modify output (#12286 Guce)
• rados: handle –snapid arg properly (Abhishek Lekshmanan)
• rados: improve bench buffer handling, performance (Piotr Dalek)
• rados: misc bench fixes (Dmitry Yatsushkevich)
• rados: new pool import implementation (John Spray)
• rados: translate errno to string in CLI (#10877 Kefu Chai)
• rbd: accept map options config option (Ilya Dryomov)
• rbd: add disk usage tool (#7746 Jason Dillaman)
• rbd: allow unmapping by spec (Ilya Dryomov)
• rbd: cli: fix arg parsing with –io-pattern (Dmitry Yatsushkevich)
• rbd: deprecate –new-format option (Jason Dillman)
• rbd: fix error messages (#2862 Rajesh Nambiar)
• rbd: fix link issues (Jason Dillaman)
• rbd: improve CLI arg parsing, usage (Ilya Dryomov)
• rbd: rbd-replay-prep and rbd-replay improvements (Jason Dillaman)
• rbd: recognize queue_depth kernel option (Ilya Dryomov)
• rbd: support G and T units for CLI (Abhishek Lekshmanan)
• rbd: update rbd man page (Ilya Dryomov)
• rbd: update xfstests tests (Douglas Fuller)
• rbd: use image-spec and snap-spec in help (Vikhyat Umrao, Ilya Dryomov)
• rest-bench: misc fixes (Shawn Chen)
• rest-bench: support https (#3968 Yuan Zhou)
• rgw: add max multipart upload parts (#12146 Abshishek Dixit)
• rgw: add missing headers to Swift container details (#10666 Ahmad Faheem, Dmytro Iurchenko)
• rgw: add stats to headers for account GET (#10684 Yuan Zhou)
• rgw: add Trasnaction-Id to response (Abhishek Dixit)
• rgw: add X-Timestamp for Swift containers (#10938 Radoslaw Zarzynski)
• rgw: always check if token is expired (#11367 Anton Aksola, Riku Lehto)
• rgw: conversion tool to repair broken multipart objects (#12079 Yehuda Sadeh)
• rgw: document layout of pools and objects (Pete Zaitcev)
• rgw: do not enclose bucket header in quotes (#11860 Wido den Hollander)
• rgw: do not prefetch data for HEAD requests (Guang Yang)
• rgw: do not preserve ACLs when copying object (#12370 Yehuda Sadeh)
• rgw: do not set content-type if length is 0 (#11091 Orit Wasserman)
• rgw: don’t clobber bucket/object owner when setting ACLs (#10978 Yehuda Sadeh)
• rgw: don’t use end_marker for namespaced object listing (#11437 Yehuda Sadeh)
• rgw: don’t use rgw_socket_path if frontend is configured (#11160 Yehuda Sadeh)
• rgw: enforce Content-Length for POST on Swift cont/obj (#10661 Radoslaw Zarzynski)
• rgw: error out if frontend did not send all data (#11851 Yehuda Sadeh)
• rgw: expose the number of unhealthy workers through admin socket (Guang Yang)
• rgw: fail if parts not specified on multipart upload (#11435 Yehuda Sadeh)
• rgw: fix assignment of copy obj attributes (#11563 Yehuda Sadeh)
• rgw: fix broken stats in container listing (#11285 Radoslaw Zarzynski)
• rgw: fix bug in domain/subdomain splitting (Robin H. Johnson)
• rgw: fix casing of Content-Type header (Robin H. Johnson)
• rgw: fix civetweb max threads (#10243 Yehuda Sadeh)
• rgw: fix Connection: header handling (#12298 Wido den Hollander)
• rgw: fix copy metadata, support X-Copied-From for swift (#10663 Radoslaw Zarzynski)
• rgw: fix data corruptions race condition (#11749 Wuxingyi)
• rgw: fix decoding of X-Object-Manifest from GET on Swift DLO (Radslow Rzarzynski)
• rgw: fix GET on swift account when limit == 0 (#10683 Radoslaw Zarzynski)
• rgw: fix handling empty metadata items on Swift container (#11088 Radoslaw Zarzynski)
• rgw: fix JSON response when getting user quota (#12117 Wuxingyi)
• rgw: fix locator for objects starting with _ (#11442 Yehuda Sadeh)
• rgw: fix log rotation (Wuxingyi)
• rgw: fix mulitipart upload in retry path (#11604 Yehuda Sadeh)
• rgw: fix quota enforcement on POST (#11323 Sergey Arkhipov)
• rgw: fix reset_loc (#11974 Yehuda Sadeh)
• rgw: fix return code on missing upload (#11436 Yehuda Sadeh)
• rgw: fix sysvinit script
• rgw: fix sysvinit script w/ multiple instances (Sage Weil, Pavan Rallabhandi)
• rgw: force content_type for swift bucket stats requests (#12095 Orit Wasserman)
• rgw: force content type header on responses with no body (#11438 Orit Wasserman)
• rgw: generate Date header for civetweb (#10873 Radoslaw Zarzynski)
• rgw: generate new object tag when setting attrs (#11256 Yehuda Sadeh)
• rgw: improve content-length env var handling (#11419 Robin H. Johnson)
• rgw: improved support for swift account metadata (Radoslaw Zarzynski)
• rgw: improve handling of already removed buckets in expirer (Radoslaw Rzarzynski)
• rgw: issue aio for first chunk before flush cached data (#11322 Guang Yang)
• rgw: log to /var/log/ceph instead of /var/log/radosgw
• rgw: make init script wait for radosgw to stop (#11140 Dmitry Yatsushkevich)
• rgw: make max put size configurable (#6999 Yuan Zhou)
• rgw: make quota/gc threads configurable (#11047 Guang Yang)
• rgw: make read user buckets backward compat (#10683 Radoslaw Zarzynski)
• rgw: merge manifests properly with prefix override (#11622 Yehuda Sadeh)
• rgw: only scan for objects not in a namespace (#11984 Yehuda Sadeh)
• rgw: orphan detection tool (Yehuda Sadeh)
• rgw: pass in civetweb configurables (#10907 Yehuda Sadeh)
• rgw: rectify 202 Accepted in PUT response (#11148 Radoslaw Zarzynski)
• rgw: remove meta file after deleting bucket (#11149 Orit Wasserman)
• rgw: remove trailing :port from HTTP_HOST header (Sage Weil)
• rgw: return 412 on bad limit when listing buckets (#11613 Yehuda Sadeh)
• rgw: rework X-Trans-Id header to conform with Swift API (Radoslaw Rzarzynski)
• rgw: s3 encoding-type for get bucket (Jeff Weber)
• rgw: send ETag, Last-Modified for swift (#11087 Radoslaw Zarzynski)
• rgw: set content length on container GET, PUT, DELETE, HEAD (#10971, #11036 Radoslaw Zarzynski)
• rgw: set max buckets per user in ceph.conf (Vikhyat Umrao)
• rgw: shard work over multiple librados instances (Pavan Rallabhandi)
• rgw: support end marker on swift container GET (#10682 Radoslaw Zarzynski)
• rgw: support for Swift expiration API (Radoslaw Rzarzynski, Yehuda Sadeh)
• rgw: swift: allow setting attributes with COPY (#10662 Ahmad Faheem, Dmytro Iurchenko)
• rgw: swift: do not override sent content type (#12363 Orit Wasserman)
• rgw: swift: enforce Content-Type in response (#12157 Radoslaw Zarzynski)
• rgw: swift: fix account listing (#11501 Radoslaw Zarzynski)
• rgw: swift: fix metadata handling on copy (#10645 Radoslaw Zarzynski)
• rgw: swift: send Last-Modified header (#10650 Radoslaw Zarzynski)
• rgw: swift: set Content-Length for account GET (#12158 Radoslav Zarzynski)
• rgw: swift: set content-length on keystone tokens (#11473 Herv Rousseau)
• rgw: update keystone cache with token info (#11125 Yehuda Sadeh)
• rgw: update to latest civetweb, enable config for IPv6 (#10965 Yehuda Sadeh)
• rgw: use attrs from source bucket on copy (#11639 Javier M. Mellid)
• rgw: use correct oid for gc chains (#11447 Yehuda Sadeh)
• rgw: user rm is idempotent (Orit Wasserman)
• rgw: use unique request id for civetweb (#10295 Orit Wasserman)
• rocksdb: add perf counters for get/put latency (Xinxin Shu)
• rocksdb, leveldb: fix compact_on_mount (Xiaoxi Chen)
• rocksdb: pass options as single string (Xiaoxi Chen)
• rocksdb: update to latest (Xiaoxi Chen)
• rpm: add suse firewall files (Tim Serong)
• rpm: always rebuild and install man pages for rpm (Owen Synge)
• rpm: loosen ceph-test dependencies (Ken Dreyer)
• rpm: many spec file fixes (Owen Synge, Ken Dreyer)
• rpm: misc fixes (Boris Ranto, Owen Synge, Ken Dreyer, Ira Cooper)
• rpm: misc systemd and SUSE fixes (Owen Synge, Nathan Cutler)
• selinux policy (Boris Ranto, Milan Broz)
• systemd: logrotate fixes (Tim Serong, Lars Marowsky-Bree, Nathan Cutler)
• systemd: many fixes (Sage Weil, Owen Synge, Boris Ranto, Dan van der Ster)
• systemd: run daemons as user ceph
• sysvinit compat: misc fixes (Owen Synge)
• test: misc fs test improvements (John Spray, Loic Dachary)
• test: python tests, linter cleanup (Alfredo Deza)
• tests: fixes for rbd xstests (Douglas Fuller)
• tests: fix tiering health checks (Loic Dachary)
• tests for low-level performance (Haomai Wang)
• tests: many ec non-regression improvements (Loic Dachary)
• tests: many many ec test improvements (Loic Dachary)
• upstart: throttle restarts (#11798 Sage Weil, Greg Farnum)

v9.1.0 Infernalis release candidate

This is the first Infernalis release candidate. There have been somemajor changes since Hammer, and the upgrade process is non-trivial.Please read carefully.

Getting the release candidate

The v9.1.0 packages are pushed to the development release repositories:

http://download.ceph.com/rpm-testing
http://download.ceph.com/debian-testing

For for info, see:

http://docs.ceph.com/docs/master/install/get-packages/

Or install with ceph-deploy via:

ceph-deploy install --testing HOST

Known issues

• librbd and librados ABI compatibility is broken. Be carefulinstalling this RC on client machines (e.g., those running qemu).It will be fixed in the final v9.2.0 release.

Major Changes from Hammer

• General:
  • Ceph daemons are now managed via systemd (with the exception ofUbuntu Trusty, which still uses upstart).
  • Ceph daemons run as ‘ceph’ user instead of root.
  • On Red Hat distros, there is also an SELinux policy.
• RADOS:
  • The RADOS cache tier can now proxy write operations to the basetier, allowing writes to be handled without forcing migration ofan object into the cache.
  • The SHEC erasure coding support is no longer flagged asexperimental. SHEC trades some additional storage space for fasterrepair.
  • There is now a unified queue (and thus prioritization) of clientIO, scrubbing, and snapshot trimming.
  • There have been many improvements to low-level repair tooling(ceph-objectstore-tool).
  • The internal ObjectStore API has been significantly cleaned up in orderto faciliate new storage backends like NewStore.
• RGW:
  • The Swift API now supports object expiration.
  • There are many Swift API compatibility improvements.
• RBD:
  • The rbddu command shows actual usage (quickly, whenobject-map is enabled).
  • The object-map feature has seen many stability improvements.
  • Object-map and exclusive-lock features can be enabled or disableddynamically.
  • You can now store user metadata and set persistent librbd optionsassociated with individual images.
  • The new deep-flatten features allows flattening of a clone and allof its snapshots. (Previously snapshots could not be flattened.)
  • The export-diff command command is now faster (it uses aio). There is alsoa new fast-diff feature.
  • The –size argument can be specified with a suffix for units(e.g., —size64G).
  • There is a new rbdstatus command that, for now, shows who hasthe image open/mapped.
• CephFS:
  • You can now rename snapshots.
  • There have been ongoing improvements around administration, diagnostics,and the check and repair tools.
  • The caching and revocation of client cache state due to unusedinodes has been dramatically improved.
  • The ceph-fuse client behaves better on 32-bit hosts.

Distro compatibility

We have decided to drop support for many older distributions so that we canmove to a newer compiler toolchain (e.g., C++11). Although it is still possibleto build Ceph on older distributions by installing backported development tools,we are not building and publishing release packages for them on ceph.com.

In particular,

• CentOS 7 or later; we have dropped support for CentOS 6 (and otherRHEL 6 derivatives, like Scientific Linux 6).
• Debian Jessie 8.x or later; Debian Wheezy 7.x’s g++ has incompletesupport for C++11 (and no systemd).
• Ubuntu Trusty 14.04 or later; Ubuntu Precise 12.04 is no longersupported.
• Fedora 22 or later.

Upgrading from Firefly

Upgrading directly from Firefly v0.80.z is not possible. All clustersmust first upgrade to Hammer v0.94.4 or a later v0.94.z release; onlythen is it possible to do online upgrade to Infernalis 9.2.z.

User can upgrade to latest hammer v0.94.zfrom gitbuilder with(also refer the hammer release notes for more details):

ceph-deploy install --release hammer HOST

Upgrading from Hammer

• All cluster nodes must first upgrade to Hammer v0.94.4 or a later v0.94.z release; onlythen is it possible to do online upgrade to Infernalis 9.2.z.

• For all distributions that support systemd (CentOS 7, Fedora, DebianJessie 8.x, OpenSUSE), ceph daemons are now managed using native systemdfiles instead of the legacy sysvinit scripts. For example,:

systemctl start ceph.target       # start all daemons
systemctl status ceph-osd@12      # check status of osd.12

The main notable distro that is not yet using systemd is Ubuntu trusty14.04. (The next Ubuntu LTS, 16.04, will use systemd instead of upstart.)

• Ceph daemons now run as user and group ceph by default. Theceph user has a static UID assigned by Fedora and Debian (also usedby derivative distributions like RHEL/CentOS and Ubuntu). On SUSEthe ceph user will currently get a dynamically assigned UID when theuser is created.

If your systems already have a ceph user, the package upgradeprocess will usually fail with an error. We suggest you firstremove or rename the existing ‘ceph’ user and then upgrade.

When upgrading, administrators have two options:

1. Add the following line to ceph.conf on all hosts:

   setuser match path = /var/lib/ceph/$type/$cluster-$id

   This will make the Ceph daemons run as root (i.e., not dropprivileges and switch to user ceph) if the daemon’s datadirectory is still owned by root. Newly deployed daemons willbe created with data owned by user ceph and will run withreduced privileges, but upgraded daemons will continue to run asroot.

2. Fix the data ownership during the upgrade. This is the preferred option,but is more work. The process for each host would be to:

   1. Upgrade the ceph package. This creates the ceph user and group. Forexample:

      ceph-deploy install —stable infernalis HOST

   2. Stop the daemon(s).:

      service ceph stop           # fedora, centos, rhel, debianstop ceph-all               # ubuntu

   3. Fix the ownership:

      chown -R ceph:ceph /var/lib/ceph

   4. Restart the daemon(s).:

      start ceph-all                # ubuntusystemctl start ceph.target   # debian, centos, fedora, rhel

• The on-disk format for the experimental KeyValueStore OSD backend haschanged. You will need to remove any OSDs using that backend before youupgrade any test clusters that use it.

Upgrade notes

• When a pool quota is reached, librados operations now block indefinitely,the same way they do when the cluster fills up. (Previously they would return-ENOSPC). By default, a full cluster or pool will now block. If yourlibrados application can handle ENOSPC or EDQUOT errors gracefully, you canget error returns instead by using the new librados OPERATION_FULL_TRY flag.

Notable changes

NOTE: These notes are somewhat abbreviated while we find a lesstime-consuming process for generating them.

• build: C++11 now supported
• build: many cmake improvements
• build: OSX build fixes (Yan, Zheng)
• build: remove rest-bench
• ceph-disk: many fixes (Loic Dachary)
• ceph-disk: support for multipath devices (Loic Dachary)
• ceph-fuse: mostly behave on 32-bit hosts (Yan, Zheng)
• ceph-objectstore-tool: many improvements (David Zafman)
• common: bufferlist performance tuning (Piotr Dalek, Sage Weil)
• common: make mutex more efficient
• common: some async compression infrastructure (Haomai Wang)
• librados: add FULL_TRY and FULL_FORCE flags for dealing with full clusters or pools (Sage Weil)
• librados: fix notify completion race (#13114 Sage Weil)
• librados, libcephfs: randomize client nonces (Josh Durgin)
• librados: pybind: fix binary omap values (Robin H. Johnson)
• librbd: fix reads larger than the cache size (Lu Shi)
• librbd: metadata filter fixes (Haomai Wang)
• librbd: use write_full when possible (Zhiqiang Wang)
• mds: avoid emitting cap warnigns before evicting session (John Spray)
• mds: fix expected holes in journal objects (#13167 Yan, Zheng)
• mds: fix SnapServer crash on deleted pool (John Spray)
• mds: many fixes (Yan, Zheng, John Spray, Greg Farnum)
• mon: add cache over MonitorDBStore (Kefu Chai)
• mon: ‘ceph osd metadata’ can dump all osds (Haomai Wang)
• mon: detect kv backend failures (Sage Weil)
• mon: fix CRUSH map test for new pools (Sage Weil)
• mon: fix min_last_epoch_clean tracking (Kefu Chai)
• mon: misc scaling fixes (Sage Weil)
• mon: streamline session handling, fix memory leaks (Sage Weil)
• mon: upgrades must pass through hammer (Sage Weil)
• msg/async: many fixes (Haomai Wang)
• osd: cache proxy-write support (Zhiqiang Wang, Samuel Just)
• osd: configure promotion based on write recency (Zhiqiang Wang)
• osd: don’t send dup MMonGetOSDMap requests (Sage Weil, Kefu Chai)
• osd: erasure-code: fix SHEC floating point bug (#12936 Loic Dachary)
• osd: erasure-code: update to ISA-L 2.14 (Yuan Zhou)
• osd: fix hitset object naming to use GMT (Kefu Chai)
• osd: fix misc memory leaks (Sage Weil)
• osd: fix peek_queue locking in FileStore (Xinze Chi)
• osd: fix promotion vs full cache tier (Samuel Just)
• osd: fix replay requeue when pg is still activating (#13116 Samuel Just)
• osd: fix scrub stat bugs (Sage Weil, Samuel Just)
• osd: force promotion for ops EC can’t handle (Zhiqiang Wang)
• osd: improve behavior on machines with large memory pages (Steve Capper)
• osd: merge multiple setattr calls into a setattrs call (Xinxin Shu)
• osd: newstore prototype (Sage Weil)
• osd: ObjectStore internal API refactor (Sage Weil)
• osd: SHEC no longer experimental
• osd: throttle evict ops (Yunchuan Wen)
• osd: upgrades must pass through hammer (Sage Weil)
• osd: use SEEK_HOLE / SEEK_DATA for sparse copy (Xinxin Shu)
• rbd: rbd-replay-prep and rbd-replay improvements (Jason Dillaman)
• rgw: expose the number of unhealthy workers through admin socket (Guang Yang)
• rgw: fix casing of Content-Type header (Robin H. Johnson)
• rgw: fix decoding of X-Object-Manifest from GET on Swift DLO (Radslow Rzarzynski)
• rgw: fix sysvinit script
• rgw: fix sysvinit script w/ multiple instances (Sage Weil, Pavan Rallabhandi)
• rgw: improve handling of already removed buckets in expirer (Radoslaw Rzarzynski)
• rgw: log to /var/log/ceph instead of /var/log/radosgw
• rgw: rework X-Trans-Id header to be conform with Swift API (Radoslaw Rzarzynski)
• rgw: s3 encoding-type for get bucket (Jeff Weber)
• rgw: set max buckets per user in ceph.conf (Vikhyat Umrao)
• rgw: support for Swift expiration API (Radoslaw Rzarzynski, Yehuda Sadeh)
• rgw: user rm is idempotent (Orit Wasserman)
• selinux policy (Boris Ranto, Milan Broz)
• systemd: many fixes (Sage Weil, Owen Synge, Boris Ranto, Dan van der Ster)
• systemd: run daemons as user ceph

v9.0.3

This is the second to last batch of development work for theInfernalis cycle. The most intrusive change is an internal (nonuser-visible) change to the OSD’s ObjectStore interface. Many fixes andimprovements elsewhere across RGW, RBD, and another big pile of CephFSscrub/repair improvements.

Upgrading

• The return code for librbd’s rbd_aio_read and Image::aio_read API methods nolonger returns the number of bytes read upon success. Instead, it returns 0upon success and a negative value upon failure.
• ‘ceph scrub’, ‘ceph compact’ and ‘ceph sync force’ are now deprecated. Usersshould instead use ‘ceph mon scrub’, ‘ceph mon compact’ and‘ceph mon sync force’.
• ‘ceph mon_metadata’ should now be used as ‘ceph mon metadata’.
• The –dump-json option of “osdmaptool” is replaced by –dump json.
• The commands of ‘pg ls-by-{pool,primary,osd}’ and ‘pg ls’ now take ‘recovering’instead of ‘recovery’ to include the recovering pgs in the listed pgs.

Notable Changes

• autotools: fix out of tree build (Krxysztof Kosinski)
• autotools: improve make check output (Loic Dachary)
• buffer: add invalidate_crc() (Piotr Dalek)
• buffer: fix zero bug (#12252 Haomai Wang)
• build: fix junit detection on Fedora 22 (Ira Cooper)
• ceph-disk: install pip > 6.1 (#11952 Loic Dachary)
• cephfs-data-scan: many additions, improvements (John Spray)
• ceph: improve error output for ‘tell’ (#11101 Kefu Chai)
• ceph-objectstore-tool: misc improvements (David Zafman)
• ceph-objectstore-tool: refactoring and cleanup (John Spray)
• ceph_test_rados: test pipelined reads (Zhiqiang Wang)
• common: fix bit_vector extent calc (#12611 Jason Dillaman)
• common: make work queue addition/removal thread safe (#12662 Jason Dillaman)
• common: optracker improvements (Zhiqiang Wang, Jianpeng Ma)
• crush: add –check to validate dangling names, max osd id (Kefu Chai)
• crush: cleanup, sync with kernel (Ilya Dryomov)
• crush: fix subtree base weight on adjust_subtree_weight (#11855 Sage Weil)
• crypo: fix NSS leak (Jason Dillaman)
• crypto: fix unbalanced init/shutdown (#12598 Zheng Yan)
• doc: misc updates (Kefu Chai, Owen Synge, Gael Fenet-Garde, Loic Dachary, Yannick Atchy-Dalama, Jiaying Ren, Kevin Caradant, Robert Maxime, Nicolas Yong, Germain Chipaux, Arthur Gorjux, Gabriel Sentucq, Clement Lebrun, Jean-Remi Deveaux, Clair Massot, Robin Tang, Thomas Laumondais, Jordan Dorne, Yuan Zhou, Valentin Thomas, Pierre Chaumont, Benjamin Troquereau, Benjamin Sesia, Vikhyat Umrao)
• erasure-code: cleanup (Kefu Chai)
• erasure-code: improve tests (Loic Dachary)
• erasure-code: shec: fix recovery bugs (Takanori Nakao, Shotaro Kawaguchi)
• libcephfs: add pread, pwrite (Jevon Qiao)
• libcephfs,ceph-fuse: cache cleanup (Zheng Yan)
• librados: add src_fadvise_flags for copy-from (Jianpeng Ma)
• librados: respect default_crush_ruleset on pool_create (#11640 Yuan Zhou)
• librbd: fadvise for copy, export, import (Jianpeng Ma)
• librbd: handle NOCACHE fadvise flag (Jinapeng Ma)
• librbd: optionally disable allocation hint (Haomai Wang)
• librbd: prevent race between resize requests (#12664 Jason Dillaman)
• log: fix data corruption race resulting from log rotation (#12465 Samuel Just)
• mds: expose frags via asok (John Spray)
• mds: fix setting entire file layout in one setxattr (John Spray)
• mds: fix shutdown (John Spray)
• mds: handle misc corruption issues (John Spray)
• mds: misc fixes (Jianpeng Ma, Dan van der Ster, Zhang Zhi)
• mds: misc snap fixes (Zheng Yan)
• mds: store layout on header object (#4161 John Spray)
• misc performance and cleanup (Nathan Cutler, Xinxin Shu)
• mon: add NOFORWARD, OBSOLETE, DEPRECATE flags for mon commands (Joao Eduardo Luis)
• mon: add PG count to ‘ceph osd df’ output (Michal Jarzabek)
• mon: clean up, reorg some mon commands (Joao Eduardo Luis)
• mon: disallow >2 tiers (#11840 Kefu Chai)
• mon: fix log dump crash when debugging (Mykola Golub)
• mon: fix metadata update race (Mykola Golub)
• mon: fix refresh (#11470 Joao Eduardo Luis)
• mon: make blocked op messages more readable (Jianpeng Ma)
• mon: only send mon metadata to supporting peers (Sage Weil)
• mon: periodic background scrub (Joao Eduardo Luis)
• mon: prevent pgp_num > pg_num (#12025 Xinxin Shu)
• mon: reject large max_mds values (#12222 John Spray)
• msgr: add ceph_perf_msgr tool (Hoamai Wang)
• msgr: async: fix seq handling (Haomai Wang)
• msgr: xio: fastpath improvements (Raju Kurunkad)
• msgr: xio: sync with accellio v1.4 (Vu Pham)
• osd: clean up temp object if promotion fails (Jianpeng Ma)
• osd: constrain collections to meta and PGs (normal and temp) (Sage Weil)
• osd: filestore: clone using splice (Jianpeng Ma)
• osd: filestore: fix recursive lock (Xinxin Shu)
• osd: fix dup promotion lost op bug (Zhiqiang Wang)
• osd: fix temp-clearing (David Zafman)
• osd: include a temp namespace within each collection/pgid (Sage Weil)
• osd: low and high speed flush modes (Mingxin Liu)
• osd: peer_features includes self (David Zafman)
• osd: recovery, peering fixes (#11687 Samuel Just)
• osd: require firefly features (David Zafman)
• osd: set initial crush weight with more precision (Sage Weil)
• osd: use a temp object for recovery (Sage Weil)
• osd: use blkid to collection partition information (Joseph Handzik)
• rados: add –striper option to use libradosstriper (#10759 Sebastien Ponce)
• radosgw-admin: fix subuser modify output (#12286 Guce)
• rados: handle –snapid arg properly (Abhishek Lekshmanan)
• rados: improve bench buffer handling, performance (Piotr Dalek)
• rados: new pool import implementation (John Spray)
• rbd: fix link issues (Jason Dillaman)
• rbd: improve CLI arg parsing, usage (Ilya Dryomov)
• rbd: recognize queue_depth kernel option (Ilya Dryomov)
• rbd: support G and T units for CLI (Abhishek Lekshmanan)
• rbd: use image-spec and snap-spec in help (Vikhyat Umrao, Ilya Dryomov)
• rest-bench: misc fixes (Shawn Chen)
• rest-bench: support https (#3968 Yuan Zhou)
• rgw: add max multipart upload parts (#12146 Abshishek Dixit)
• rgw: add Trasnaction-Id to response (Abhishek Dixit)
• rgw: document layout of pools and objects (Pete Zaitcev)
• rgw: do not preserve ACLs when copying object (#12370 Yehuda Sadeh)
• rgw: fix Connection: header handling (#12298 Wido den Hollander)
• rgw: fix data corruptions race condition (#11749 Wuxingyi)
• rgw: fix JSON response when getting user quota (#12117 Wuxingyi)
• rgw: force content_type for swift bucket stats requests (#12095 Orit Wasserman)
• rgw: improved support for swift account metadata (Radoslaw Zarzynski)
• rgw: make max put size configurable (#6999 Yuan Zhou)
• rgw: orphan detection tool (Yehuda Sadeh)
• rgw: swift: do not override sent content type (#12363 Orit Wasserman)
• rgw: swift: set Content-Length for account GET (#12158 Radoslav Zarzynski)
• rpm: always rebuild and install man pages for rpm (Owen Synge)
• rpm: misc fixes (Boris Ranto, Owen Synge, Ken Dreyer, Ira Cooper)
• systemd: logrotate fixes (Tim Seron, Lars Marowsky-Bree, Nathan Cutler)
• sysvinit compat: misc fixes (Owen Synge)
• test: misc fs test improvements (John Spray, Loic Dachary)
• test: python tests, linter cleanup (Alfredo Deza)

v9.0.2

This development release features more of the OSD work queueunification, randomized osd scrub times, a huge pile of librbd fixes,more MDS repair and snapshot fixes, and a significant amount of workon the tests and build infrastructure.

Notable Changes

• buffer: some cleanup (Michal Jarzabek)
• build: cmake: fix nss linking (Danny Al-Gaaf)
• build: cmake: misc fixes (Orit Wasserman, Casey Bodley)
• build: install-deps: misc fixes (Loic Dachary)
• build: make_dist_tarball.sh (Sage Weil)
• ceph-detect-init: added Linux Mint (Michal Jarzabek)
• ceph-detect-init: robust init system detection (Owen Synge, Loic Dachary)
• ceph-disk: ensure ‘zap’ only operates on a full disk (#11272 Loic Dachary)
• ceph-disk: misc fixes to respect init system (Loic Dachary, Owen Synge)
• ceph-disk: support NVMe device partitions (#11612 Ilja Slepnev)
• ceph: fix ‘df’ units (Zhe Zhang)
• ceph: fix parsing in interactive cli mode (#11279 Kefu Chai)
• ceph-objectstore-tool: many many changes (David Zafman)
• ceph-post-file: misc fixes (Joey McDonald, Sage Weil)
• client: avoid sending unnecessary FLUSHSNAP messages (Yan, Zheng)
• client: exclude setfilelock when calculating oldest tid (Yan, Zheng)
• client: fix error handling in check_pool_perm (John Spray)
• client: fsync waits only for inode’s caps to flush (Yan, Zheng)
• client: invalidate kernel dcache when cache size exceeds limits (Yan, Zheng)
• client: make fsync wait for unsafe dir operations (Yan, Zheng)
• client: pin lookup dentry to avoid inode being freed (Yan, Zheng)
• common: detect overflow of int config values (#11484 Kefu Chai)
• common: fix json parsing of utf8 (#7387 Tim Serong)
• common: fix leak of pthread_mutexattr (#11762 Ketor Meng)
• crush: respect default replicated ruleset config on map creation (Ilya Dryomov)
• deb, rpm: move ceph-objectstore-tool to ceph (Ken Dreyer)
• doc: man page updates (Kefu Chai)
• doc: misc updates (#11396 Nilamdyuti, Fracois Lafont, Ken Dreyer, Kefu Chai)
• init-radosgw: merge with sysv version; fix enumeration (Sage Weil)
• librados: add config observer (Alistair Strachan)
• librbd: add const for single-client-only features (Josh Durgin)
• librbd: add deep-flatten operation (Jason Dillaman)
• librbd: avoid blocking aio API methods (#11056 Jason Dillaman)
• librbd: fix fast diff bugs (#11553 Jason Dillaman)
• librbd: fix image format detection (Zhiqiang Wang)
• librbd: fix lock ordering issue (#11577 Jason Dillaman)
• librbd: flatten/copyup fixes (Jason Dillaman)
• librbd: lockdep, helgrind validation (Jason Dillaman, Josh Durgin)
• librbd: only update image flags while hold exclusive lock (#11791 Jason Dillaman)
• librbd: return result code from close (#12069 Jason Dillaman)
• librbd: tolerate old osds when getting image metadata (#11549 Jason Dillaman)
• mds: do not add snapped items to bloom filter (Yan, Zheng)
• mds: fix handling for missing mydir dirfrag (#11641 John Spray)
• mds: fix rejoin (Yan, Zheng)
• mds: fix stra reintegration (Yan, Zheng)
• mds: fix suicide beason (John Spray)
• mds: misc repair improvements (John Spray)
• mds: misc snapshot fixes (Yan, Zheng)
• mds: respawn instead of suicide on blacklist (John Spray)
• misc coverity fixes (Danny Al-Gaaf)
• mon: add ‘mon_metadata ’ command (Kefu Chai)
• mon: add ‘node ls …’ command (Kefu Chai)
• mon: disallow ec pools as tiers (#11650 Samuel Just)
• mon: fix mds beacon replies (#11590 Kefu Chai)
• mon: fix ‘pg ls’ sort order, state names (#11569 Kefu Chai)
• mon: normalize erasure-code profile for storage and comparison (Loic Dachary)
• mon: optionally specify osd id on ‘osd create’ (Mykola Golub)
• mon: ‘osd tree’ fixes (Kefu Chai)
• mon: prevent pool with snapshot state from being used as a tier (#11493 Sage Weil)
• mon: refine check_remove_tier checks (#11504 John Spray)
• mon: remove spurious who arg from ‘mds rm …’ (John Spray)
• msgr: async: misc fixes (Haomai Wang)
• msgr: xio: fix ip and nonce (Raju Kurunkad)
• msgr: xio: improve lane assignment (Vu Pham)
• msgr: xio: misc fixes (Vu Pham, Cosey Bodley)
• osd: avoid transaction append in some cases (Sage Weil)
• osdc/Objecter: allow per-pool calls to op_cancel_writes (John Spray)
• osd: elminiate txn apend, ECSubWrite copy (Samuel Just)
• osd: filejournal: cleanup (David Zafman)
• osd: fix check_for_full (Henry Chang)
• osd: fix dirty accounting in make_writeable (Zhiqiang Wang)
• osd: fix osdmap dump of blacklist items (John Spray)
• osd: fix snap flushing from cache tier (again) (#11787 Samuel Just)
• osd: fix snap handling on promotion (#11296 Sam Just)
• osd: handle log split with overlapping entries (#11358 Samuel Just)
• osd: keyvaluestore: misc fixes (Varada Kari)
• osd: make suicide timeouts individually configurable (Samuel Just)
• osd: move scrub in OpWQ (Samuel Just)
• osd: pool size change triggers new interval (#11771 Samuel Just)
• osd: randomize scrub times (#10973 Kefu Chai)
• osd: refactor scrub and digest recording (Sage Weil)
• osd: refuse first write to EC object at non-zero offset (Jianpeng Ma)
• osd: stripe over small xattrs to fit in XFS’s 255 byte inline limit (Sage Weil, Ning Yao)
• osd: sync object_map on syncfs (Samuel Just)
• osd: take excl lock of op is rw (Samuel Just)
• osd: WBThrottle cleanups (Jianpeng Ma)
• pycephfs: many fixes for bindings (Haomai Wang)
• rados: bench: add –no-verify option to improve performance (Piotr Dalek)
• rados: misc bench fixes (Dmitry Yatsushkevich)
• rbd: add disk usage tool (#7746 Jason Dillaman)
• rgw: alwasy check if token is expired (#11367 Anton Aksola, Riku Lehto)
• rgw: conversion tool to repair broken multipart objects (#12079 Yehuda Sadeh)
• rgw: do not enclose bucket header in quotes (#11860 Wido den Hollander)
• rgw: error out if frontend did not send all data (#11851 Yehuda Sadeh)
• rgw: fix assignment of copy obj attributes (#11563 Yehuda Sadeh)
• rgw: fix reset_loc (#11974 Yehuda Sadeh)
• rgw: improve content-length env var handling (#11419 Robin H. Johnson)
• rgw: only scan for objects not in a namespace (#11984 Yehuda Sadeh)
• rgw: remove trailing :port from HTTP_HOST header (Sage Weil)
• rgw: shard work over multiple librados instances (Pavan Rallabhandi)
• rgw: swift: enforce Content-Type in response (#12157 Radoslaw Zarzynski)
• rgw: use attrs from source bucket on copy (#11639 Javier M. Mellid)
• rocksdb: pass options as single string (Xiaoxi Chen)
• rpm: many spec file fixes (Owen Synge, Ken Dreyer)
• tests: fixes for rbd xstests (Douglas Fuller)
• tests: fix tiering health checks (Loic Dachary)
• tests for low-level performance (Haomai Wang)
• tests: many ec non-regression improvements (Loic Dachary)
• tests: many many ec test improvements (Loic Dachary)
• upstart: throttle restarts (#11798 Sage Weil, Greg Farnum)

v9.0.1

This development release is delayed a bit due to tooling changes in the buildenvironment. As a result the next one (v9.0.2) will have a bit more work thanis usual.

Highlights here include lots of RGW Swift fixes, RBD feature worksurrounding the new object map feature, more CephFS snapshot fixes,and a few important CRUSH fixes.

Notable Changes

• auth: cache/reuse crypto lib key objects, optimize msg signature check (Sage Weil)
• build: allow tcmalloc-minimal (Thorsten Behrens)
• build: do not build ceph-dencoder with tcmalloc (#10691 Boris Ranto)
• build: fix pg ref disabling (William A. Kennington III)
• build: install-deps.sh improvements (Loic Dachary)
• build: misc fixes (Boris Ranto, Ken Dreyer, Owen Synge)
• ceph-authtool: fix return code on error (Gerhard Muntingh)
• ceph-disk: fix zap sgdisk invocation (Owen Synge, Thorsten Behrens)
• ceph-disk: pass –cluster arg on prepare subcommand (Kefu Chai)
• ceph-fuse, libcephfs: drop inode when rmdir finishes (#11339 Yan, Zheng)
• ceph-fuse,libcephfs: fix uninline (#11356 Yan, Zheng)
• ceph-monstore-tool: fix store-copy (Huangjun)
• common: add perf counter descriptions (Alyona Kiseleva)
• common: fix throttle max change (Henry Chang)
• crush: fix crash from invalid ‘take’ argument (#11602 Shiva Rkreddy, Sage Weil)
• crush: fix divide-by-2 in straw2 (#11357 Yann Dupont, Sage Weil)
• deb: fix rest-bench-dbg and ceph-test-dbg dependendies (Ken Dreyer)
• doc: document region hostnames (Robin H. Johnson)
• doc: update release schedule docs (Loic Dachary)
• init-radosgw: run radosgw as root (#11453 Ken Dreyer)
• librados: fadvise flags per op (Jianpeng Ma)
• librbd: allow additional metadata to be stored with the image (Haomai Wang)
• librbd: better handling for dup flatten requests (#11370 Jason Dillaman)
• librbd: cancel in-flight ops on watch error (#11363 Jason Dillaman)
• librbd: default new images to format 2 (#11348 Jason Dillaman)
• librbd: fast diff implementation that leverages object map (Jason Dillaman)
• librbd: fix snapshot creation when other snap is active (#11475 Jason Dillaman)
• librbd: new diff_iterate2 API (Jason Dillaman)
• librbd: object map rebuild support (Jason Dillaman)
• logrotate.d: prefer service over invoke-rc.d (#11330 Win Hierman, Sage Weil)
• mds: avoid getting stuck in XLOCKDONE (#11254 Yan, Zheng)
• mds: fix integer truncateion on large client ids (Henry Chang)
• mds: many snapshot and stray fixes (Yan, Zheng)
• mds: persist completed_requests reliably (#11048 John Spray)
• mds: separate safe_pos in Journaler (#10368 John Spray)
• mds: snapshot rename support (#3645 Yan, Zheng)
• mds: warn when clients fail to advance oldest_client_tid (#10657 Yan, Zheng)
• misc cleanups and fixes (Danny Al-Gaaf)
• mon: fix average utilization calc for ‘osd df’ (Mykola Golub)
• mon: fix variance calc in ‘osd df’ (Sage Weil)
• mon: improve callout to crushtool (Mykola Golub)
• mon: prevent bucket deletion when referenced by a crush rule (#11602 Sage Weil)
• mon: prime pg_temp when CRUSH map changes (Sage Weil)
• monclient: flush_log (John Spray)
• msgr: async: many many fixes (Haomai Wang)
• msgr: simple: fix clear_pipe (#11381 Haomai Wang)
• osd: add latency perf counters for tier operations (Xinze Chi)
• osd: avoid multiple hit set insertions (Zhiqiang Wang)
• osd: break PG removal into multiple iterations (#10198 Guang Yang)
• osd: check scrub state when handling map (Jianpeng Ma)
• osd: fix endless repair when object is unrecoverable (Jianpeng Ma, Kefu Chai)
• osd: fix pg resurrection (#11429 Samuel Just)
• osd: ignore non-existent osds in unfound calc (#10976 Mykola Golub)
• osd: increase default max open files (Owen Synge)
• osd: prepopulate needs_recovery_map when only one peer has missing (#9558 Guang Yang)
• osd: relax reply order on proxy read (#11211 Zhiqiang Wang)
• osd: skip promotion for flush/evict op (Zhiqiang Wang)
• osd: write journal header on clean shutdown (Xinze Chi)
• qa: run-make-check.sh script (Loic Dachary)
• rados bench: misc fixes (Dmitry Yatsushkevich)
• rados: fix error message on failed pool removal (Wido den Hollander)
• radosgw-admin: add ‘bucket check’ function to repair bucket index (Yehuda Sadeh)
• rbd: allow unmapping by spec (Ilya Dryomov)
• rbd: deprecate –new-format option (Jason Dillman)
• rgw: do not set content-type if length is 0 (#11091 Orit Wasserman)
• rgw: don’t use end_marker for namespaced object listing (#11437 Yehuda Sadeh)
• rgw: fail if parts not specified on multipart upload (#11435 Yehuda Sadeh)
• rgw: fix GET on swift account when limit == 0 (#10683 Radoslaw Zarzynski)
• rgw: fix broken stats in container listing (#11285 Radoslaw Zarzynski)
• rgw: fix bug in domain/subdomain splitting (Robin H. Johnson)
• rgw: fix civetweb max threads (#10243 Yehuda Sadeh)
• rgw: fix copy metadata, support X-Copied-From for swift (#10663 Radoslaw Zarzynski)
• rgw: fix locator for objects starting with _ (#11442 Yehuda Sadeh)
• rgw: fix mulitipart upload in retry path (#11604 Yehuda Sadeh)
• rgw: fix quota enforcement on POST (#11323 Sergey Arkhipov)
• rgw: fix return code on missing upload (#11436 Yehuda Sadeh)
• rgw: force content type header on responses with no body (#11438 Orit Wasserman)
• rgw: generate new object tag when setting attrs (#11256 Yehuda Sadeh)
• rgw: issue aio for first chunk before flush cached data (#11322 Guang Yang)
• rgw: make read user buckets backward compat (#10683 Radoslaw Zarzynski)
• rgw: merge manifests properly with prefix override (#11622 Yehuda Sadeh)
• rgw: return 412 on bad limit when listing buckets (#11613 Yehuda Sadeh)
• rgw: send ETag, Last-Modified for swift (#11087 Radoslaw Zarzynski)
• rgw: set content length on container GET, PUT, DELETE, HEAD (#10971, #11036 Radoslaw Zarzynski)
• rgw: support end marker on swift container GET (#10682 Radoslaw Zarzynski)
• rgw: swift: fix account listing (#11501 Radoslaw Zarzynski)
• rgw: swift: set content-length on keystone tokens (#11473 Herv Rousseau)
• rgw: use correct oid for gc chains (#11447 Yehuda Sadeh)
• rgw: use unique request id for civetweb (#10295 Orit Wasserman)
• rocksdb, leveldb: fix compact_on_mount (Xiaoxi Chen)
• rocksdb: add perf counters for get/put latency (Xinxin Shu)
• rpm: add suse firewall files (Tim Serong)
• rpm: misc systemd and suse fixes (Owen Synge, Nathan Cutler)

v9.0.0

This is the first development release for the Infernalis cycle, andthe first Ceph release to sport a version number from the newnumbering scheme. The “9” indicates this is the 9th release cycle–I(for Infernalis) is the 9th letter. The first “0” indicates this is adevelopment release (“1” will mean release candidate and “2” will meanstable release), and the final “0” indicates this is the first suchdevelopment release.

A few highlights include:

• a new ‘ceph daemonperf’ command to watch perfcounter stats in realtime
• reduced MDS memory usage
• many MDS snapshot fixes
• librbd can now store options in the image itself
• many fixes for RGW Swift API support
• OSD performance improvements
• many doc updates and misc bug fixes

Notable Changes

• aarch64: add optimized version of crc32c (Yazen Ghannam, Steve Capper)
• auth: reinit NSS after fork() (#11128 Yan, Zheng)
• build: disable LTTNG by default (#11333 Josh Durgin)
• build: fix ppc build (James Page)
• build: install-deps: support OpenSUSE (Loic Dachary)
• build: misc cmake fixes (Matt Benjamin)
• ceph-disk: follow ceph-osd hints when creating journal (#9580 Sage Weil)
• ceph-disk: handle re-using existing partition (#10987 Loic Dachary)
• ceph-disk: improve parted output parsing (#10983 Loic Dachary)
• ceph-disk: make suppression work for activate-all and activate-journal (Dan van der Ster)
• ceph-disk: misc fixes (Alfredo Deza)
• ceph-fuse, libcephfs: don’t clear COMPLETE when trimming null (Yan, Zheng)
• ceph-fuse, libcephfs: hold exclusive caps on dirs we “own” (#11226 Greg Farnum)
• ceph-fuse: do not require successful remount when unmounting (#10982 Greg Farnum)
• ceph: new ‘ceph daemonperf’ command (John Spray, Mykola Golub)
• common: PriorityQueue tests (Kefu Chai)
• common: add descriptions to perfcounters (Kiseleva Alyona)
• common: fix LTTNG vs fork issue (Josh Durgin)
• crush: fix has_v4_buckets (#11364 Sage Weil)
• crushtool: fix order of operations, usage (Sage Weil)
• debian: minor package reorg (Ken Dreyer)
• doc: docuemnt object corpus generation (#11099 Alexis Normand)
• doc: fix gender neutrality (Alexandre Maragone)
• doc: fix install doc (#10957 Kefu Chai)
• doc: fix sphinx issues (Kefu Chai)
• doc: mds data structure docs (Yan, Zheng)
• doc: misc updates (Nilamdyuti Goswami, Vartika Rai, Florian Haas, Loic Dachary, Simon Guinot, Andy Allan, Alistair Israel, Ken Dreyer, Robin Rehu, Lee Revell, Florian Marsylle, Thomas Johnson, Bosse Klykken, Travis Rhoden, Ian Kelling)
• doc: swift tempurls (#10184 Abhishek Lekshmanan)
• doc: switch doxygen integration back to breathe (#6115 Kefu Chai)
• erasure-code: update ISA-L to 2.13 (Yuan Zhou)
• gmock: switch to submodule (Danny Al-Gaaf, Loic Dachary)
• hadoop: add terasort test (Noah Watkins)
• java: fix libcephfs bindings (Noah Watkins)
• libcephfs,ceph-fuse: fix request resend on cap reconnect (#10912 Yan, Zheng)
• librados: define C++ flags from C constants (Josh Durgin)
• librados: fix last_force_resent handling (#11026 Jianpeng Ma)
• librados: fix memory leak from C_TwoContexts (Xiong Yiliang)
• librados: fix striper when stripe_count = 1 and stripe_unit != object_size (#11120 Yan, Zheng)
• librados: op perf counters (John Spray)
• librados: pybind: fix write() method return code (Javier Guerra)
• libradosstriper: fix leak (Danny Al-Gaaf)
• librbd: add purge_on_error cache behavior (Jianpeng Ma)
• librbd: misc aio fixes (#5488 Jason Dillaman)
• librbd: misc rbd fixes (#11478 #11113 #11342 #11380 Jason Dillaman, Zhiqiang Wang)
• librbd: readahead fixes (Zhiqiang Wang)
• librbd: store metadata, including config options, in image (Haomai Wang)
• mds: add ‘damaged’ state to MDSMap (John Spray)
• mds: add nicknames for perfcounters (John Spray)
• mds: disable problematic rstat propagation into snap parents (Yan, Zheng)
• mds: fix mydir replica issue with shutdown (#10743 John Spray)
• mds: fix out-of-order messages (#11258 Yan, Zheng)
• mds: fix shutdown with strays (#10744 John Spray)
• mds: fix snapshot fixes (Yan, Zheng)
• mds: fix stray handling (John Spray)
• mds: flush immediately in do_open_truncate (#11011 John Spray)
• mds: improve dump methods (John Spray)
• mds: misc journal cleanups and fixes (#10368 John Spray)
• mds: new SessionMap storage using omap (#10649 John Spray)
• mds: reduce memory consumption (Yan, Zheng)
• mds: throttle purge stray operations (#10390 John Spray)
• mds: tolerate clock jumping backwards (#11053 Yan, Zheng)
• misc coverity fixes (Danny Al-Gaaf)
• mon: do not deactivate last mds (#10862 John Spray)
• mon: make osd get pool ‘all’ only return applicable fields (#10891 Michal Jarzabek)
• mon: warn on bogus cache tier config (Jianpeng Ma)
• msg/async: misc bug fixes and updates (Haomai Wang)
• msg/simple: fix connect_seq assert (Haomai Wang)
• msg/xio: misc fixes (#10735 Matt Benjamin, Kefu Chai, Danny Al-Gaaf, Raju Kurunkad, Vu Pham)
• msg: unit tests (Haomai Wang)
• objectcacher: misc bug fixes (Jianpeng Ma)
• os/filestore: enlarge getxattr buffer size (Jianpeng Ma)
• osd: EIO injection (David Zhang)
• osd: add misc perfcounters (Xinze Chi)
• osd: add simple sleep injection in recovery (Sage Weil)
• osd: allow SEEK_HOLE/SEEK_DATA for sparse read (Zhiqiang Wang)
• osd: avoid dup omap sets for in pg metadata (Sage Weil)
• osd: clean up some constness, privateness (Kefu Chai)
• osd: erasure-code: drop entries according to LRU (Andreas-Joachim Peters)
• osd: fix negative degraded stats during backfill (Guang Yang)
• osd: misc fixes (Ning Yao, Kefu Chai, Xinze Chi, Zhiqiang Wang, Jianpeng Ma)
• pybind: pep8 cleanups (Danny Al-Gaaf)
• qa: fix filelock_interrupt.py test (Yan, Zheng)
• qa: improve ceph-disk tests (Loic Dachary)
• qa: improve docker build layers (Loic Dachary)
• rados: translate erno to string in CLI (#10877 Kefu Chai)
• rbd: accept map options config option (Ilya Dryomov)
• rbd: cli: fix arg parsing with –io-pattern (Dmitry Yatsushkevich)
• rbd: fix error messages (#2862 Rajesh Nambiar)
• rbd: update rbd man page (Ilya Dryomov)
• rbd: update xfstests tests (Douglas Fuller)
• rgw: add X-Timestamp for Swift containers (#10938 Radoslaw Zarzynski)
• rgw: add missing headers to Swift container details (#10666 Ahmad Faheem, Dmytro Iurchenko)
• rgw: add stats to headers for account GET (#10684 Yuan Zhou)
• rgw: do not prefecth data for HEAD requests (Guang Yang)
• rgw: don’t clobber bucket/object owner when setting ACLs (#10978 Yehuda Sadeh)
• rgw: don’t use rgw_socket_path if frontend is configured (#11160 Yehuda Sadeh)
• rgw: enforce Content-Lenth for POST on Swift cont/obj (#10661 Radoslaw Zarzynski)
• rgw: fix handling empty metadata items on Swift container (#11088 Radoslaw Zarzynski)
• rgw: fix log rotation (Wuxingyi)
• rgw: generate Date header for civetweb (#10873 Radoslaw Zarzynski)
• rgw: make init script wait for radosgw to stop (#11140 Dmitry Yatsushkevich)
• rgw: make quota/gc threads configurable (#11047 Guang Yang)
• rgw: pass in civetweb configurables (#10907 Yehuda Sadeh)
• rgw: rectify 202 Accepted in PUT response (#11148 Radoslaw Zarzynski)
• rgw: remove meta file after deleting bucket (#11149 Orit Wasserman)
• rgw: swift: allow setting attributes with COPY (#10662 Ahmad Faheem, Dmytro Iurchenko)
• rgw: swift: fix metadata handling on copy (#10645 Radoslaw Zarzynski)
• rgw: swift: send Last-Modified header (#10650 Radoslaw Zarzynski)
• rgw: update keystone cache with token info (#11125 Yehuda Sadeh)
• rgw: update to latest civetweb, enable config for IPv6 (#10965 Yehuda Sadeh)
• rocksdb: update to latest (Xiaoxi Chen)
• rpm: loosen ceph-test dependencies (Ken Dreyer)

v0.94.5 Hammer

This Hammer point release fixes a critical regression in librbd that can causeQemu/KVM to crash when caching is enabled on images that have been cloned.

All v0.94.4 Hammer users are strongly encouraged to upgrade.

Notable Changes

• librbd: potential assertion failure during cache read (issue#13559, pr#6348, Jason Dillaman)
• osd: osd/ReplicatedPG: remove stray debug line (issue#13455, pr#6362, Sage Weil)
• tests: qemu workunit refers to apt-mirror.front.sepia.ceph.com (issue#13420, pr#6330, Yuan Zhou)

For more detailed information, see thecompletechangelog.

v0.94.4 Hammer

This Hammer point release fixes several important bugs in Hammer, as well asfixing interoperability issues that are required before an upgrade toInfernalis. That is, all users of earlier version of Hammer or anyversion of Firefly will first need to upgrade to hammer v0.94.4 orlater before upgrading to Infernalis (or future releases).

All v0.94.x Hammer users are strongly encouraged to upgrade.

Notable Changes

• build/ops: ceph.spec.in: 50-rbd.rules conditional is wrong (issue#12166, pr#5207, Nathan Cutler)
• build/ops: ceph.spec.in: ceph-common needs python-argparse on older distros, but doesn’t require it (issue#12034, pr#5216, Nathan Cutler)
• build/ops: ceph.spec.in: radosgw requires apache for SUSE only – makes no sense (issue#12358, pr#5411, Nathan Cutler)
• build/ops: ceph.spec.in: rpm: cephfs_java not fully conditionalized (issue#11991, pr#5202, Nathan Cutler)
• build/ops: ceph.spec.in: rpm: not possible to turn off Java (issue#11992, pr#5203, Owen Synge)
• build/ops: ceph.spec.in: running fdupes unnecessarily (issue#12301, pr#5223, Nathan Cutler)
• build/ops: ceph.spec.in: snappy-devel for all supported distros (issue#12361, pr#5264, Nathan Cutler)
• build/ops: ceph.spec.in: SUSE/openSUSE builds need libbz2-devel (issue#11629, pr#5204, Nathan Cutler)
• build/ops: ceph.spec.in: useless %py_requires breaks SLE11-SP3 build (issue#12351, pr#5412, Nathan Cutler)
• build/ops: error in ext_mime_map_init() when /etc/mime.types is missing (issue#11864, pr#5385, Ken Dreyer)
• build/ops: upstart: limit respawn to 3 in 30 mins (instead of 5 in 30s) (issue#11798, pr#5930, Sage Weil)
• build/ops: With root as default user, unable to have multiple RGW instances running (issue#10927, pr#6161, Sage Weil)
• build/ops: With root as default user, unable to have multiple RGW instances running (issue#11140, pr#6161, Sage Weil)
• build/ops: With root as default user, unable to have multiple RGW instances running (issue#11686, pr#6161, Sage Weil)
• build/ops: With root as default user, unable to have multiple RGW instances running (issue#12407, pr#6161, Sage Weil)
• cli: ceph: cli throws exception on unrecognized errno (issue#11354, pr#5368, Kefu Chai)
• cli: ceph tell: broken error message / misleading hinting (issue#11101, pr#5371, Kefu Chai)
• common: arm: all programs that link to librados2 hang forever on startup (issue#12505, pr#5366, Boris Ranto)
• common: buffer: critical bufferlist::zero bug (issue#12252, pr#5365, Haomai Wang)
• common: ceph-object-corpus: add 0.94.2-207-g88e7ee7 hammer objects (issue#13070, pr#5551, Sage Weil)
• common: do not insert emtpy ptr when rebuild emtpy bufferlist (issue#12775, pr#5764, Xinze Chi)
• common: [ FAILED ] TestLibRBD.BlockingAIO (issue#12479, pr#5768, Jason Dillaman)
• common: LibCephFS.GetPoolId failure (issue#12598, pr#5887, Yan, Zheng)
• common: Memory leak in Mutex.cc, pthread_mutexattr_init without pthread_mutexattr_destroy (issue#11762, pr#5378, Ketor Meng)
• common: object_map_update fails with -EINVAL return code (issue#12611, pr#5559, Jason Dillaman)
• common: Pipe: Drop connect_seq increase line (issue#13093, pr#5908, Haomai Wang)
• common: recursive lock of md_config_t (0) (issue#12614, pr#5759, Josh Durgin)
• crush: ceph osd crush reweight-subtree does not reweight parent node (issue#11855, pr#5374, Sage Weil)
• doc: update docs to point to download.ceph.com (issue#13162, pr#6156, Alfredo Deza)
• fs: ceph-fuse 0.94.2-1trusty segfaults / aborts (issue#12297, pr#5381, Greg Farnum)
• fs: segfault launching ceph-fuse with bad –name (issue#12417, pr#5382, John Spray)
• librados: Change radosgw pools default crush ruleset (issue#11640, pr#5754, Yuan Zhou)
• librbd: correct issues discovered via lockdep / helgrind (issue#12345, pr#5296, Jason Dillaman)
• librbd: Crash during TestInternal.MultipleResize (issue#12664, pr#5769, Jason Dillaman)
• librbd: deadlock during cooperative exclusive lock transition (issue#11537, pr#5319, Jason Dillaman)
• librbd: Possible crash while concurrently writing and shrinking an image (issue#11743, pr#5318, Jason Dillaman)
• mon: add a cache layer over MonitorDBStore (issue#12638, pr#5697, Kefu Chai)
• mon: fix crush testing for new pools (issue#13400, pr#6192, Sage Weil)
• mon: get pools health’info have error (issue#12402, pr#5369, renhwztetecs)
• mon: implicit erasure code crush ruleset is not validated (issue#11814, pr#5276, Loic Dachary)
• mon: PaxosService: call post_refresh() instead of post_paxos_update() (issue#11470, pr#5359, Joao Eduardo Luis)
• mon: pgmonitor: wrong at/near target max“ reporting (issue#12401, pr#5370, huangjun)
• mon: register_new_pgs() should check ruleno instead of its index (issue#12210, pr#5377, Xinze Chi)
• mon: Show osd as NONE in ceph osd map For more detailed information, see thecompletechangelog.

v0.94.3 Hammer

This Hammer point release fixes a critical (though rare) datacorruption bug that could be triggered when logs are rotated viaSIGHUP. It also fixes a range of other important bugs in the OSD,monitor, RGW, RGW, and CephFS.

All v0.94.x Hammer users are strongly encouraged to upgrade.

Upgrading

• The pgls-by-{pool,primary,osd} commands and pgls now takethe argument recovering instead of recovery in order toinclude the recovering pgs in the listed pgs.

Notable Changes

• librbd: aio calls may block (issue#11770, pr#4875, Jason Dillaman)
• osd: make the all osd/filestore thread pool suicide timeouts separately configurable (issue#11701, pr#5159, Samuel Just)
• mon: ceph fails to compile with boost 1.58 (issue#11982, pr#5122, Kefu Chai)
• tests: TEST_crush_reject_empty must not run a mon (issue#12285,11975, pr#5208, Kefu Chai)
• osd: FAILED assert(!old_value.deleted()) in upgrade:giant-x-hammer-distro-basic-multi run (issue#11983, pr#5121, Samuel Just)
• build/ops: linking ceph to tcmalloc causes segfault on SUSE SLE11-SP3 (issue#12368, pr#5265, Thorsten Behrens)
• common: utf8 and old gcc breakage on RHEL6.5 (issue#7387, pr#4687, Kefu Chai)
• crush: take crashes due to invalid arg (issue#11740, pr#4891, Sage Weil)
• rgw: need conversion tool to handle fixes following #11974 (issue#12502, pr#5384, Yehuda Sadeh)
• rgw: Swift API: support for 202 Accepted response code on container creation (issue#12299, pr#5214, Radoslaw Zarzynski)
• common: Log::reopen_log_file: take m_flush_mutex (issue#12520, pr#5405, Samuel Just)
• rgw: Properly respond to the Connection header with Civetweb (issue#12398, pr#5284, Wido den Hollander)
• rgw: multipart list part response returns incorrect field (issue#12399, pr#5285, Henry Chang)
• build/ops: ceph.spec.in: 95-ceph-osd.rules, mount.ceph, and mount.fuse.ceph not installed properly on SUSE (issue#12397, pr#5283, Nathan Cutler)
• rgw: radosgw-admin dumps user info twice (issue#12400, pr#5286, guce)
• doc: fix doc build (issue#12180, pr#5095, Kefu Chai)
• tests: backport 11493 fixes, and test, preventing ec cache pools (issue#12314, pr#4961, Samuel Just)
• rgw: does not send Date HTTP header when civetweb frontend is used (issue#11872, pr#5228, Radoslaw Zarzynski)
• mon: pg ls is broken (issue#11910, pr#5160, Kefu Chai)
• librbd: A client opening an image mid-resize can result in the object map being invalidated (issue#12237, pr#5279, Jason Dillaman)
• doc: missing man pages for ceph-create-keys, ceph-disk-* (issue#11862, pr#4846, Nathan Cutler)
• tools: ceph-post-file fails on rhel7 (issue#11876, pr#5038, Sage Weil)
• build/ops: rcceph script is buggy (issue#12090, pr#5028, Owen Synge)
• rgw: Bucket header is enclosed by quotes (issue#11874, pr#4862, Wido den Hollander)
• build/ops: packaging: add SuSEfirewall2 service files (issue#12092, pr#5030, Tim Serong)
• rgw: Keystone PKI token expiration is not enforced (issue#11722, pr#4884, Anton Aksola)
• build/ops: debian/control: ceph-common (>> 0.94.2) must be >= 0.94.2-2 (issue#12529,11998, pr#5417, Loic Dachary)
• mon: Clock skew causes missing summary and confuses Calamari (issue#11879, pr#4868, Thorsten Behrens)
• rgw: rados objects wronly deleted (issue#12099, pr#5117, wuxingyi)
• tests: kernel_untar_build fails on EL7 (issue#12098, pr#5119, Greg Farnum)
• fs: Fh ref count will leak if readahead does not need to do read from osd (issue#12319, pr#5427, Zhi Zhang)
• mon: OSDMonitor: allow addition of cache pool with non-empty snaps with co… (issue#12595, pr#5252, Samuel Just)
• mon: MDSMonitor: handle MDSBeacon messages properly (issue#11979, pr#5123, Kefu Chai)
• tools: ceph-disk: get_partition_type fails on /dev/cciss… (issue#11760, pr#4892, islepnev)
• build/ops: max files open limit for OSD daemon is too low (issue#12087, pr#5026, Owen Synge)
• mon: add an “osd crush tree” command (issue#11833, pr#5248, Kefu Chai)
• mon: mon crashes when “ceph osd tree 85 –format json” (issue#11975, pr#4936, Kefu Chai)
• build/ops: ceph / ceph-dbg steal ceph-objecstore-tool from ceph-test / ceph-test-dbg (issue#11806, pr#5069, Loic Dachary)
• rgw: DragonDisk fails to create directories via S3: MissingContentLength (issue#12042, pr#5118, Yehuda Sadeh)
• build/ops: /usr/bin/ceph from ceph-common is broken without installing ceph (issue#11998, pr#5206, Ken Dreyer)
• build/ops: systemd: Increase max files open limit for OSD daemon (issue#11964, pr#5040, Owen Synge)
• build/ops: rgw/logrotate.conf calls service with wrong init script name (issue#12044, pr#5055, wuxingyi)
• common: OPT_INT option interprets 3221225472 as -1073741824, and crashes in Throttle::Throttle() (issue#11738, pr#4889, Kefu Chai)
• doc: doc/release-notes: v0.94.2 (issue#11492, pr#4934, Sage Weil)
• common: admin_socket: close socket descriptor in destructor (issue#11706, pr#4657, Jon Bernard)
• rgw: Object copy bug (issue#11755, pr#4885, Javier M. Mellid)
• rgw: empty json response when getting user quota (issue#12245, pr#5237, wuxingyi)
• fs: cephfs Dumper tries to load whole journal into memory at once (issue#11999, pr#5120, John Spray)
• rgw: Fix tool for #11442 does not correctly fix objects created via multipart uploads (issue#12242, pr#5229, Yehuda Sadeh)
• rgw: Civetweb RGW appears to report full size of object as downloaded when only partially downloaded (issue#12243, pr#5231, Yehuda Sadeh)
• osd: stuck incomplete (issue#12362, pr#5269, Samuel Just)
• osd: start_flush: filter out removed snaps before determining snapc’s (issue#11911, pr#4899, Samuel Just)
• librbd: internal.cc: 1967: FAILED assert(watchers.size() == 1) (issue#12239, pr#5243, Jason Dillaman)
• librbd: new QA client upgrade tests (issue#12109, pr#5046, Jason Dillaman)
• librbd: [ FAILED ] TestLibRBD.ExclusiveLockTransition (issue#12238, pr#5241, Jason Dillaman)
• rgw: Swift API: XML document generated in response for GET on account does not contain account name (issue#12323, pr#5227, Radoslaw Zarzynski)
• rgw: keystone does not support chunked input (issue#12322, pr#5226, Hervé Rousseau)
• mds: MDS is crashed (mds/CDir.cc: 1391: FAILED assert(!is_complete())) (issue#11737, pr#4886, Yan, Zheng)
• cli: ceph: cli interactive mode does not understand quotes (issue#11736, pr#4776, Kefu Chai)
• librbd: add valgrind memory checks for unit tests (issue#12384, pr#5280, Zhiqiang Wang)
• build/ops: admin/build-doc: script fails silently under certain circumstances (issue#11902, pr#4877, John Spray)
• osd: Fixes for rados ops with snaps (issue#11908, pr#4902, Samuel Just)
• build/ops: ceph.spec.in: ceph-common subpackage def needs tweaking for SUSE/openSUSE (issue#12308, pr#4883, Nathan Cutler)
• fs: client: reference counting ‘struct Fh’ (issue#12088, pr#5222, Yan, Zheng)
• build/ops: ceph.spec: update OpenSUSE BuildRequires (issue#11611, pr#4667, Loic Dachary)

For more detailed information, see thecompletechangelog.

v0.94.2 Hammer

This Hammer point release fixes a few critical bugs in RGW that canprevent objects starting with underscore from behaving properly andthat prevent garbage collection of deleted objects when using theCivetweb standalone mode.

All v0.94.x Hammer users are strongly encouraged to upgrade, and tomake note of the repair procedure below if RGW is in use.

Upgrading from previous Hammer release

Bug #11442 introduced a change that made rgw objects that start with underscoreincompatible with previous versions. The fix to that bug reverts to theprevious behavior. In order to be able to access objects that start with anunderscore and were created in prior Hammer releases, following the upgrade itis required to run (for each affected bucket):

$ radosgw-admin bucket check --check-head-obj-locator \
                             --bucket=<bucket> [--fix]

Notable changes

• build: compilation error: No high-precision counter available (armhf, powerpc..) (#11432, James Page)
• ceph-dencoder links to libtcmalloc, and shouldn’t (#10691, Boris Ranto)
• ceph-disk: disk zap sgdisk invocation (#11143, Owen Synge)
• ceph-disk: use a new disk as journal disk,ceph-disk prepare fail (#10983, Loic Dachary)
• ceph-objectstore-tool should be in the ceph server package (#11376, Ken Dreyer)
• librados: can get stuck in redirect loop if osdmap epoch == last_force_op_resend (#11026, Jianpeng Ma)
• librbd: A retransmit of proxied flatten request can result in -EINVAL (Jason Dillaman)
• librbd: ImageWatcher should cancel in-flight ops on watch error (#11363, Jason Dillaman)
• librbd: Objectcacher setting max object counts too low (#7385, Jason Dillaman)
• librbd: Periodic failure of TestLibRBD.DiffIterateStress (#11369, Jason Dillaman)
• librbd: Queued AIO reference counters not properly updated (#11478, Jason Dillaman)
• librbd: deadlock in image refresh (#5488, Jason Dillaman)
• librbd: notification race condition on snap_create (#11342, Jason Dillaman)
• mds: Hammer uclient checking (#11510, John Spray)
• mds: remove caps from revoking list when caps are voluntarily released (#11482, Yan, Zheng)
• messenger: double clear of pipe in reaper (#11381, Haomai Wang)
• mon: Total size of OSDs is a maginitude less than it is supposed to be. (#11534, Zhe Zhang)
• osd: don’t check order in finish_proxy_read (#11211, Zhiqiang Wang)
• osd: handle old semi-deleted pgs after upgrade (#11429, Samuel Just)
• osd: object creation by write cannot use an offset on an erasure coded pool (#11507, Jianpeng Ma)
• rgw: Improve rgw HEAD request by avoiding read the body of the first chunk (#11001, Guang Yang)
• rgw: civetweb is hitting a limit (number of threads 1024) (#10243, Yehuda Sadeh)
• rgw: civetweb should use unique request id (#10295, Orit Wasserman)
• rgw: critical fixes for hammer (#11447, #11442, Yehuda Sadeh)
• rgw: fix swift COPY headers (#10662, #10663, #11087, #10645, Radoslaw Zarzynski)
• rgw: improve performance for large object (multiple chunks) GET (#11322, Guang Yang)
• rgw: init-radosgw: run RGW as root (#11453, Ken Dreyer)
• rgw: keystone token cache does not work correctly (#11125, Yehuda Sadeh)
• rgw: make quota/gc thread configurable for starting (#11047, Guang Yang)
• rgw: make swift responses of RGW return last-modified, content-length, x-trans-id headers.(#10650, Radoslaw Zarzynski)
• rgw: merge manifests correctly when there’s prefix override (#11622, Yehuda Sadeh)
• rgw: quota not respected in POST object (#11323, Sergey Arkhipov)
• rgw: restore buffer of multipart upload after EEXIST (#11604, Yehuda Sadeh)
• rgw: shouldn’t need to disable rgw_socket_path if frontend is configured (#11160, Yehuda Sadeh)
• rgw: swift: Response header of GET request for container does not contain X-Container-Object-Count, X-Container-Bytes-Used and x-trans-id headers (#10666, Dmytro Iurchenko)
• rgw: swift: Response header of POST request for object does not contain content-length and x-trans-id headers (#10661, Radoslaw Zarzynski)
• rgw: swift: response for GET/HEAD on container does not contain the X-Timestamp header (#10938, Radoslaw Zarzynski)
• rgw: swift: response for PUT on /container does not contain the mandatory Content-Length header when FCGI is used (#11036, #10971, Radoslaw Zarzynski)
• rgw: swift: wrong handling of empty metadata on Swift container (#11088, Radoslaw Zarzynski)
• tests: TestFlatIndex.cc races with TestLFNIndex.cc (#11217, Xinze Chi)
• tests: ceph-helpers kill_daemons fails when kill fails (#11398, Loic Dachary)

For more detailed information, see thecompletechangelog.

v0.94.1 Hammer

This bug fix release fixes a few critical issues with CRUSH. The mostimportant addresses a bug in feature bit enforcement that may preventpre-hammer clients from communicating with the cluster during anupgrade. This only manifests in some cases (for example, when the‘rack’ type is in use in the CRUSH map, and possibly other cases), but forsafety we strongly recommend that all users use 0.94.1 instead of 0.94 whenupgrading.

There is also a fix in the new straw2 buckets when OSD weights are 0.

We recommend that all v0.94 users upgrade.

Notable changes

• crush: fix divide-by-0 in straw2 (#11357 Sage Weil)
• crush: fix has_v4_buckets (#11364 Sage Weil)
• osd: fix negative degraded objects during backfilling (#7737 Guang Yang)

For more detailed information, see thecompletechangelog.

v0.94 Hammer

This major release is expected to form the basis of the next long-termstable series. It is intended to supersede v0.80.x Firefly.

Highlights since Giant include:

• RADOS Performance: a range of improvements have been made in theOSD and client-side librados code that improve the throughput onflash backends and improve parallelism and scaling on fast machines.
• Simplified RGW deployment: the ceph-deploy tool now has a new‘ceph-deploy rgw create HOST’ command that quickly deploys ainstance of the S3/Swift gateway using the embedded Civetweb server.This is vastly simpler than the previous Apache-based deployment.There are a few rough edges (e.g., around SSL support) but weencourage users to try the new method.
• RGW object versioning: RGW now supports the S3 object versioningAPI, which preserves old version of objects instead of overwritingthem.
• RGW bucket sharding: RGW can now shard the bucket index for largebuckets across, improving performance for very large buckets.
• RBD object maps: RBD now has an object map function that trackswhich parts of the image are allocating, improving performance forclones and for commands like export and delete.
• RBD mandatory locking: RBD has a new mandatory locking framework(still disabled by default) that adds additional safeguards toprevent multiple clients from using the same image at the same time.
• RBD copy-on-read: RBD now supports copy-on-read for image clones,improving performance for some workloads.
• CephFS snapshot improvements: Many many bugs have been fixed withCephFS snapshots. Although they are still disabled by default,stability has improved significantly.
• CephFS Recovery tools: We have built some journal recovery anddiagnostic tools. Stability and performance of single-MDS systems isvastly improved in Giant, and more improvements have been made nowin Hammer. Although we still recommend caution when storingimportant data in CephFS, we do encourage testing for non-criticalworkloads so that we can better guage the feature, usability,performance, and stability gaps.
• CRUSH improvements: We have added a new straw2 bucket algorithmthat reduces the amount of data migration required when changes aremade to the cluster.
• Shingled erasure codes (SHEC): The OSDs now have experimentalsupport for shingled erasure codes, which allow a small amount ofadditional storage to be traded for improved recovery performance.
• RADOS cache tiering: A series of changes have been made in thecache tiering code that improve performance and reduce latency.
• RDMA support: There is now experimental support the RDMA via theAccelio (libxio) library.
• New administrator commands: The ‘ceph osd df’ command showspertinent details on OSD disk utilizations. The ‘ceph pg ls …’command makes it much simpler to query PG states while diagnosingcluster issues.

Other highlights since Firefly include:

• CephFS: we have fixed a raft of bugs in CephFS and built somebasic journal recovery and diagnostic tools. Stability andperformance of single-MDS systems is vastly improved in Giant.Although we do not yet recommend CephFS for production deployments,we do encourage testing for non-critical workloads so that we canbetter guage the feature, usability, performance, and stabilitygaps.
• Local Recovery Codes: the OSDs now support an erasure-coding schemethat stores some additional data blocks to reduce the IO required torecover from single OSD failures.
• Degraded vs misplaced: the Ceph health reports from ‘ceph -s’ andrelated commands now make a distinction between data that isdegraded (there are fewer than the desired number of copies) anddata that is misplaced (stored in the wrong location in thecluster). The distinction is important because the latter does notcompromise data safety.
• Tiering improvements: we have made several improvements to thecache tiering implementation that improve performance. Mostnotably, objects are not promoted into the cache tier by a singleread; they must be found to be sufficiently hot before that happens.
• Monitor performance: the monitors now perform writes to the localdata store asynchronously, improving overall responsiveness.
• Recovery tools: the ceph-objectstore-tool is greatly expanded toallow manipulation of an individual OSDs data store for debuggingand repair purposes. This is most heavily used by our QAinfrastructure to exercise recovery code.

I would like to take this opportunity to call out the amazing growthin contributors to Ceph beyond the core development team from Inktank.Hammer features major new features and improvements from Intel, Fujitsu,UnitedStack, Yahoo, UbuntuKylin, CohortFS, Mellanox, CERN, DeutscheTelekom, Mirantis, and SanDisk.

Dedication

This release is dedicated in memoriam to Sandon Van Ness, akaHoukouonchi, who unexpectedly passed away a few weeks ago. Sandon wasresponsible for maintaining the large and complex Sepia lab thathouses the Ceph project’s build and test infrastructure. His effortshave made an important impact on our ability to reliably test Cephwith a relatively small group of people. He was a valued member ofthe team and we will miss him. H is also for Houkouonchi.

Upgrading

• If your existing cluster is running a version older than v0.80.xFirefly, please first upgrade to the latest Firefly release beforemoving on to Giant. We have not tested upgrades directly fromEmperor, Dumpling, or older releases.

We have tested:

• Firefly to Hammer
• Giant to Hammer
• Dumpling to Firefly to Hammer

• Please upgrade daemons in the following order:

1. Monitors
2. OSDs
3. MDSs and/or radosgw

Note that the relative ordering of OSDs and monitors should not matter, butwe primarily tested upgrading monitors first.

• The ceph-osd daemons will perform a disk-format upgrade improve thePG metadata layout and to repair a minor bug in the on-disk format.It may take a minute or two for this to complete, depending on howmany objects are stored on the node; do not be alarmed if they donot marked “up” by the cluster immediately after starting.

• • If upgrading from v0.93, set
  • osd enable degraded writes = false

on all osds prior to upgrading. The degraded writes feature hasbeen reverted due to 11155.

• The LTTNG tracing in librbd and librados is disabled in the release packagesuntil we find a way to avoid violating distro security policies when linkinglibust.

Upgrading from v0.87.x Giant

• librbd and librados include lttng tracepoints on distros withliblttng 2.4 or later (only Ubuntu Trusty for the ceph.compackages). When running a daemon that uses these libraries, i.e. anapplication that calls fork(2) or clone(2) without exec(3), you mustset LD_PRELOAD=liblttng-ust-fork.so.0 to prevent a crash in thelttng atexit handler when the process exits. The only ceph tool thatrequires this is rbd-fuse.

• If rgw_socket_path is defined and rgw_frontends defines asocket_port and socket_host, we now allow the rgw_frontends settingsto take precedence. This change should only affect users who havemade non-standard changes to their radosgw configuration.

• If you are upgrading specifically from v0.92, you must stop all OSDdaemons and flush their journals (ceph-osd-iNNN—flush-journal) before upgrading. There was a transactionencoding bug in v0.92 that broke compatibility. Upgrading from v0.93,v0.91, or anything earlier is safe.

• The experimental ‘keyvaluestore-dev’ OSD backend has been renamed‘keyvaluestore’ (for simplicity) and marked as experimental. Toenable this untested feature and acknowledge that you understandthat it is untested and may destroy data, you need to add thefollowing to your ceph.conf:

enable experimental unrecoverable data corrupting featuers = keyvaluestore

• The following librados C API function calls take a ‘flags’ argument whose valueis now correctly interpreted:

rados_write_op_operate()rados_aio_write_op_operate()rados_read_op_operate()rados_aio_read_op_operate()

The flags were not correctly being translated from the librados constants to theinternal values. Now they are. Any code that is passing flags to these methodsshould be audited to ensure that they are using the correct LIBRADOSOP_FLAG*constants.

• The ‘rados’ CLI ‘copy’ and ‘cppool’ commands now use the copy-from operation,which means the latest CLI cannot run these commands against pre-firefly OSDs.

• The librados watch/notify API now includes a watch_flush() operation to flushthe async queue of notify operations. This should be called by any watch/notifyuser prior to rados_shutdown().

• The ‘category’ field for objects has been removed. This was originally addedto track PG stat summations over different categories of objects for use byradosgw. It is no longer has any known users and is prone to abuse because itcan lead to a pg_stat_t structure that is unbounded. The librados API callsthat accept this field now ignore it, and the OSD no longers tracks theper-category summations.

• The output for ‘rados df’ has changed. The ‘category’ level has beeneliminated, so there is now a single stat object per pool. The structure ofthe JSON output is different, and the plaintext output has one less column.

• The ‘rados create [category]’ optional category argument is nolonger supported or recognized.

• rados.py’s Rados class no longer has a del method; it was causingproblems on interpreter shutdown and use of threads. If your code hasRados objects with limited lifetimes and you’re concerned about lockedresources, call Rados.shutdown() explicitly.

• There is a new version of the librados watch/notify API with vastlyimproved semantics. Any applications using this interface areencouraged to migrate to the new API. The old API calls are markedas deprecated and will eventually be removed.

• The librados rados_unwatch() call used to be safe to call on aninvalid handle. The new version has undefined behavior when passeda bogus value (for example, when rados_watch() returns an error andhandle is not defined).

• The structure of the formatted ‘pg stat’ command is changed for theportion that counts states by name to avoid using the ‘+’ character(which appears in state names) as part of the XML token (it is notlegal).

• Previously, the formatted output of ‘ceph pg stat -f …’ was a fullpg dump that included all metadata about all PGs in the system. Itis now a concise summary of high-level PG stats, just like theunformatted ‘ceph pg stat’ command.

• All JSON dumps of floating point values were incorrecting surrounding thevalue with quotes. These quotes have been removed. Any consumer of structuredJSON output that was consuming the floating point values was previously havingto interpret the quoted string and will most likely need to be fixed to takethe unquoted number.

• New ability to list all objects from all namespaces that can fail orreturn incomplete results when not all OSDs have been upgraded.Features rados –all ls, rados cppool, rados export, radoscache-flush-evict-all and rados cache-try-flush-evict-all can alsofail or return incomplete results.

• Due to a change in the Linux kernel version 3.18 and the limits of the FUSEinterface, ceph-fuse needs be mounted as root on at least some systems. Seeissues #9997, #10277, and #10542 for details.

Upgrading from v0.80x Firefly (additional notes)

• The client-side caching for librbd is now enabled by default (rbdcache = true). A safety option (rbd cache writethrough until flush= true) is also enabled so that writeback caching is not used untilthe library observes a ‘flush’ command, indicating that the librbdusers is passing that operation through from the guest VM. Thisavoids potential data loss when used with older versions of qemuthat do not support flush.

leveldb_write_buffer_size = 810241024 = 33554432 // 8MBleveldb_cache_size = 51210241204 = 536870912 // 512MBleveldb_block_size = 64*1024 = 65536 // 64KBleveldb_compression = falseleveldb_log = “”

OSDs will still maintain the following osd-specific defaults:

leveldb_log = “”

• The ‘rados getxattr …’ command used to add a gratuitous newline to the attrvalue; it now does not.

• The _kbperf counters on the monitor have been removed. These arereplaced with a new set of _bytes counters (e.g., cluster_osd_kb isreplaced by cluster_osd_bytes).

• The rd_kb and wr_kb fields in the JSON dumps for pool stats (accessedvia the cephdfdetail-fjson-pretty and related commands) have beenreplaced with corresponding *_bytes fields. Similarly, thetotal_space, total_used, and total_avail fields are replaced withtotal_bytes, total_used_bytes, and total_avail_bytes fields.

• The radosdf—format=json output read_bytes and write_bytesfields were incorrectly reporting ops; this is now fixed.

• The radosdf—format=json output previously included read_kb andwrite_kb fields; these have been removed. Please use read_bytes andwrite_bytes instead (and divide by 1024 if appropriate).

• The experimental keyvaluestore-dev OSD backend had an on-disk formatchange that prevents existing OSD data from being upgraded. Thisaffects developers and testers only.

• mon-specific and osd-specific leveldb options have been removed.From this point onward users should use the leveldb*_ genericoptions and add the options in the appropriate sections of theirconfiguration files. Monitors will still maintain the followingmonitor-specific defaults:

leveldb_write_buffer_size = 810241024 = 33554432 // 8MBleveldb_cache_size = 51210241204 = 536870912 // 512MBleveldb_block_size = 64*1024 = 65536 // 64KBleveldb_compression = falseleveldb_log = “”

OSDs will still maintain the following osd-specific defaults:

leveldb_log = “”

• CephFS support for the legacy anchor table has finally been removed.Users with file systems created before firefly should ensure that inodeswith multiple hard links are modified prior to the upgrade to ensure thatthe backtraces are written properly. For example:

sudo find /mnt/cephfs -type f -links +1 -exec touch \{\} \;

• We disallow nonsensical ‘tier cache-mode’ transitions. From this pointonward, ‘writeback’ can only transition to ‘forward’ and ‘forward’can transition to 1) ‘writeback’ if there are dirty objects, or 2) any ifthere are no dirty objects.

Notable changes since v0.93

• build: a few cmake fixes (Matt Benjamin)
• build: fix build on RHEL/CentOS 5.9 (Rohan Mars)
• build: reorganize Makefile to allow modular builds (Boris Ranto)
• ceph-fuse: be more forgiving on remount (#10982 Greg Farnum)
• ceph: improve CLI parsing (#11093 David Zafman)
• common: fix cluster logging to default channel (#11177 Sage Weil)
• crush: fix parsing of straw2 buckets (#11015 Sage Weil)
• doc: update man pages (David Zafman)
• librados: fix leak in C_TwoContexts (Xiong Yiliang)
• librados: fix leak in watch/notify path (Sage Weil)
• librbd: fix and improve AIO cache invalidation (#10958 Jason Dillaman)
• librbd: fix memory leak (Jason Dillaman)
• librbd: fix ordering/queueing of resize operations (Jason Dillaman)
• librbd: validate image is r/w on resize/flatten (Jason Dillaman)
• librbd: various internal locking fixes (Jason Dillaman)
• lttng: tracing is disabled until we streamline dependencies (Josh Durgin)
• mon: add bootstrap-rgw profile (Sage Weil)
• mon: do not pollute mon dir with CSV files from CRUSH check (Loic Dachary)
• mon: fix clock drift time check interval (#10546 Joao Eduardo Luis)
• mon: fix units in store stats (Joao Eduardo Luis)
• mon: improve error handling on erasure code profile set (#10488, #11144 Loic Dachary)
• mon: set {read,write}_tier on ‘osd tier add-cache …’ (Jianpeng Ma)
• ms: xio: fix misc bugs (Matt Benjamin, Vu Pham)
• osd: DBObjectMap: fix locking to prevent rare crash (#9891 Samuel Just)
• osd: fix and document last_epoch_started semantics (Samuel Just)
• osd: fix divergent entry handling on PG split (Samuel Just)
• osd: fix leak on shutdown (Kefu Chai)
• osd: fix recording of digest on scrub (Samuel Just)
• osd: fix whiteout handling (Sage Weil)
• rbd: allow v2 striping parameters for clones and imports (Jason Dillaman)
• rbd: fix formatted output of image features (Jason Dillaman)
• rbd: updat eman page (Ilya Dryomov)
• rgw: don’t overwrite bucket/object owner when setting ACLs (#10978 Yehuda Sadeh)
• rgw: enable IPv6 for civetweb (#10965 Yehuda Sadeh)
• rgw: fix sysvinit script when rgw_socket_path is not defined (#11159 Yehuda Sadeh, Dan Mick)
• rgw: pass civetweb configurables through (#10907 Yehuda Sadeh)
• rgw: use new watch/notify API (Yehuda Sadeh, Sage Weil)
• osd: reverted degraded writes feature due to 11155

Notable changes since v0.87.x Giant

• add experimental features option (Sage Weil)
• arch: fix NEON feaeture detection (#10185 Loic Dachary)
• asyncmsgr: misc fixes (Haomai Wang)
• buffer: add ‘shareable’ construct (Matt Benjamin)
• buffer: add list::get_contiguous (Sage Weil)
• buffer: avoid rebuild if buffer already contiguous (Jianpeng Ma)
• build: CMake support (Ali Maredia, Casey Bodley, Adam Emerson, Marcus Watts, Matt Benjamin)
• build: a few cmake fixes (Matt Benjamin)
• build: aarch64 build fixes (Noah Watkins, Haomai Wang)
• build: adjust build deps for yasm, virtualenv (Jianpeng Ma)
• build: fix ‘make check’ races (#10384 Loic Dachary)
• build: fix build on RHEL/CentOS 5.9 (Rohan Mars)
• build: fix pkg names when libkeyutils is missing (Pankag Garg, Ken Dreyer)
• build: improve build dependency tooling (Loic Dachary)
• build: reorganize Makefile to allow modular builds (Boris Ranto)
• build: support for jemalloc (Shishir Gowda)
• ceph-disk: Scientific Linux support (Dan van der Ster)
• ceph-disk: allow journal partition re-use (#10146 Loic Dachary, Dav van der Ster)
• ceph-disk: call partx/partprobe consistency (#9721 Loic Dachary)
• ceph-disk: do not re-use partition if encryption is required (Loic Dachary)
• ceph-disk: fix dmcrypt key permissions (Loic Dachary)
• ceph-disk: fix umount race condition (#10096 Blaine Gardner)
• ceph-disk: improved systemd support (Owen Synge)
• ceph-disk: init=none option (Loic Dachary)
• ceph-disk: misc fixes (Christos Stavrakakis)
• ceph-disk: respect –statedir for keyring (Loic Dachary)
• ceph-disk: set guid if reusing journal partition (Dan van der Ster)
• ceph-disk: support LUKS for encrypted partitions (Andrew Bartlett, Loic Dachary)
• ceph-fuse, libcephfs: POSIX file lock support (Yan, Zheng)
• ceph-fuse, libcephfs: allow xattr caps in inject_release_failure (#9800 John Spray)
• ceph-fuse, libcephfs: fix I_COMPLETE_ORDERED checks (#9894 Yan, Zheng)
• ceph-fuse, libcephfs: fix cap flush overflow (Greg Farnum, Yan, Zheng)
• ceph-fuse, libcephfs: fix root inode xattrs (Yan, Zheng)
• ceph-fuse, libcephfs: preserve dir ordering (#9178 Yan, Zheng)
• ceph-fuse, libcephfs: trim inodes before reconnecting to MDS (Yan, Zheng)
• ceph-fuse,libcephfs: add support for O_NOFOLLOW and O_PATH (Greg Farnum)
• ceph-fuse,libcephfs: resend requests before completing cap reconnect (#10912 Yan, Zheng)
• ceph-fuse: be more forgiving on remount (#10982 Greg Farnum)
• ceph-fuse: fix dentry invalidation on 3.18+ kernels (#9997 Yan, Zheng)
• ceph-fuse: fix kernel cache trimming (#10277 Yan, Zheng)
• ceph-fuse: select kernel cache invalidation mechanism based on kernel version (Greg Farnum)
• ceph-monstore-tool: fix shutdown (#10093 Loic Dachary)
• ceph-monstore-tool: fix/improve CLI (Joao Eduardo Luis)
• ceph-objectstore-tool: fix import (#10090 David Zafman)
• ceph-objectstore-tool: improved import (David Zafman)
• ceph-objectstore-tool: many improvements and tests (David Zafman)
• ceph-objectstore-tool: many many improvements (David Zafman)
• ceph-objectstore-tool: misc improvements, fixes (#9870 #9871 David Zafman)
• ceph.spec: package rbd-replay-prep (Ken Dreyer)
• ceph: add ‘ceph osd df [tree]’ command (#10452 Mykola Golub)
• ceph: do not parse injectargs twice (Loic Dachary)
• ceph: fix ‘ceph tell …’ command validation (#10439 Joao Eduardo Luis)
• ceph: improve ‘ceph osd tree’ output (Mykola Golub)
• ceph: improve CLI parsing (#11093 David Zafman)
• ceph: make ‘ceph -s’ output more readable (Sage Weil)
• ceph: make ‘ceph -s’ show PG state counts in sorted order (Sage Weil)
• ceph: make ‘ceph tell mon.* version’ work (Mykola Golub)
• ceph: new ‘ceph tell mds.$name_or_rank_or_gid’ (John Spray)
• ceph: show primary-affinity in ‘ceph osd tree’ (Mykola Golub)
• ceph: test robustness (Joao Eduardo Luis)
• ceph_objectstore_tool: behave with sharded flag (#9661 David Zafman)
• cephfs-journal-tool: add recover_dentries function (#9883 John Spray)
• cephfs-journal-tool: fix journal import (#10025 John Spray)
• cephfs-journal-tool: skip up to expire_pos (#9977 John Spray)
• cleanup rados.h definitions with macros (Ilya Dryomov)
• common: add ‘perf reset …’ admin command (Jianpeng Ma)
• common: add TableFormatter (Andreas Peters)
• common: add newline to flushed json output (Sage Weil)
• common: check syncfs() return code (Jianpeng Ma)
• common: do not unlock rwlock on destruction (Federico Simoncelli)
• common: filtering for ‘perf dump’ (John Spray)
• common: fix Formatter factory breakage (#10547 Loic Dachary)
• common: fix block device discard check (#10296 Sage Weil)
• common: make json-pretty output prettier (Sage Weil)
• common: remove broken CEPH_LOCKDEP optoin (Kefu Chai)
• common: shared_cache unit tests (Cheng Cheng)
• common: support new gperftools header locations (Key Dreyer)
• config: add $cctid meta variable (Adam Crume)
• crush: fix buffer overrun for poorly formed rules (#9492 Johnu George)
• crush: fix detach_bucket (#10095 Sage Weil)
• crush: fix parsing of straw2 buckets (#11015 Sage Weil)
• crush: fix several bugs in adjust_item_weight (Rongze Zhu)
• crush: fix tree bucket behavior (Rongze Zhu)
• crush: improve constness (Loic Dachary)
• crush: new and improved straw2 bucket type (Sage Weil, Christina Anderson, Xiaoxi Chen)
• crush: straw bucket weight calculation fixes (#9998 Sage Weil)
• crush: update tries stats for indep rules (#10349 Loic Dachary)
• crush: use larger choose_tries value for erasure code rulesets (#10353 Loic Dachary)
• crushtool: add –location command (Sage Weil, Loic Dachary)
• debian,rpm: move RBD udev rules to ceph-common (#10864 Ken Dreyer)
• debian: split python-ceph into python-{rbd,rados,cephfs} (Boris Ranto)
• default to libnss instead of crypto++ (Federico Gimenez)
• doc: CephFS disaster recovery guidance (John Spray)
• doc: CephFS for early adopters (John Spray)
• doc: add build-doc guidlines for Fedora and CentOS/RHEL (Nilamdyuti Goswami)
• doc: add dumpling to firefly upgrade section (#7679 John Wilkins)
• doc: ceph osd reweight vs crush weight (Laurent Guerby)
• doc: do not suggest dangerous XFS nobarrier option (Dan van der Ster)
• doc: document erasure coded pool operations (#9970 Loic Dachary)
• doc: document the LRC per-layer plugin configuration (Yuan Zhou)
• doc: enable rbd cache on openstack deployments (Sebastien Han)
• doc: erasure code doc updates (Loic Dachary)
• doc: file system osd config settings (Kevin Dalley)
• doc: fix OpenStack Glance docs (#10478 Sebastien Han)
• doc: improved installation nots on CentOS/RHEL installs (John Wilkins)
• doc: key/value store config reference (John Wilkins)
• doc: misc cleanups (Adam Spiers, Sebastien Han, Nilamdyuti Goswami, Ken Dreyer, John Wilkins)
• doc: misc improvements (Nilamdyuti Goswami, John Wilkins, Chris Holcombe)
• doc: misc updates (#9793 #9922 #10204 #10203 Travis Rhoden, Hazem, Ayari, Florian Coste, Andy Allan, Frank Yu, Baptiste Veuillez-Mainard, Yuan Zhou, Armando Segnini, Robert Jansen, Tyler Brekke, Viktor Suprun)
• doc: misc updates (Alfredo Deza, VRan Liu)
• doc: misc updates (Nilamdyuti Goswami, John Wilkins)
• doc: new man pages (Nilamdyuti Goswami)
• doc: preflight doc fixes (John Wilkins)
• doc: replace cloudfiles with swiftclient Python Swift example (Tim Freund)
• doc: update PG count guide (Gerben Meijer, Laurent Guerby, Loic Dachary)
• doc: update man pages (David Zafman)
• doc: update openstack docs for Juno (Sebastien Han)
• doc: update release descriptions (Ken Dreyer)
• doc: update sepia hardware inventory (Sandon Van Ness)
• erasure-code: add mSHEC erasure code support (Takeshi Miyamae)
• erasure-code: improved docs (#10340 Loic Dachary)
• erasure-code: set max_size to 20 (#10363 Loic Dachary)
• fix cluster logging from non-mon daemons (Sage Weil)
• init-ceph: check for systemd-run before using it (Boris Ranto)
• install-deps.sh: do not require sudo when root (Loic Dachary)
• keyvaluestore: misc fixes (Haomai Wang)
• keyvaluestore: performance improvements (Haomai Wang)
• libcephfs,ceph-fuse: add ‘status’ asok (John Spray)
• libcephfs,ceph-fuse: fix getting zero-length xattr (#10552 Yan, Zheng)
• libcephfs: fix dirfrag trimming (#10387 Yan, Zheng)
• libcephfs: fix mount timeout (#10041 Yan, Zheng)
• libcephfs: fix test (#10415 Yan, Zheng)
• libcephfs: fix use-afer-free on umount (#10412 Yan, Zheng)
• libcephfs: include ceph and git version in client metadata (Sage Weil)
• librados, osd: new watch/notify implementation (Sage Weil)
• librados: add blacklist_add convenience method (Jason Dillaman)
• librados: add rados_pool_get_base_tier() call (Adam Crume)
• librados: add watch_flush() operation (Sage Weil, Haomai Wang)
• librados: avoid memcpy on getxattr, read (Jianpeng Ma)
• librados: cap buffer length (Loic Dachary)
• librados: create ioctx by pool id (Jason Dillaman)
• librados: do notify completion in fast-dispatch (Sage Weil)
• librados: drop ‘category’ feature (Sage Weil)
• librados: expose rados_{read|write}_op_assert_version in C API (Kim Vandry)
• librados: fix infinite loop with skipped map epochs (#9986 Ding Dinghua)
• librados: fix iterator operator= bugs (#10082 David Zafman, Yehuda Sadeh)
• librados: fix leak in C_TwoContexts (Xiong Yiliang)
• librados: fix leak in watch/notify path (Sage Weil)
• librados: fix null deref when pool DNE (#9944 Sage Weil)
• librados: fix objecter races (#9617 Josh Durgin)
• librados: fix pool deletion handling (#10372 Sage Weil)
• librados: fix pool name caching (#10458 Radoslaw Zarzynski)
• librados: fix resource leak, misc bugs (#10425 Radoslaw Zarzynski)
• librados: fix some watch/notify locking (Jason Dillaman, Josh Durgin)
• librados: fix timer race from recent refactor (Sage Weil)
• librados: new fadvise API (Ma Jianpeng)
• librados: only export public API symbols (Jason Dillaman)
• librados: remove shadowed variable (Kefu Chain)
• librados: translate op flags from C APIs (Matthew Richards)
• libradosstriper: fix remove() (Dongmao Zhang)
• libradosstriper: fix shutdown hang (Dongmao Zhang)
• libradosstriper: fix stat strtoll (Dongmao Zhang)
• libradosstriper: fix trunc method (#10129 Sebastien Ponce)
• libradosstriper: fix write_full when ENOENT (#10758 Sebastien Ponce)
• libradosstriper: misc fixes (Sebastien Ponce)
• librbd: CRC protection for RBD image map (Jason Dillaman)
• librbd: add missing python docstrings (Jason Dillaman)
• librbd: add per-image object map for improved performance (Jason Dillaman)
• librbd: add readahead (Adam Crume)
• librbd: add support for an “object map” indicating which objects exist (Jason Dillaman)
• librbd: adjust internal locking (Josh Durgin, Jason Dillaman)
• librbd: better handling of watch errors (Jason Dillaman)
• librbd: complete pending ops before closing image (#10299 Josh Durgin)
• librbd: coordinate maint operations through lock owner (Jason Dillaman)
• librbd: copy-on-read (Min Chen, Li Wang, Yunchuan Wen, Cheng Cheng, Jason Dillaman)
• librbd: differentiate between R/O vs R/W features (Jason Dillaman)
• librbd: don’t close a closed parent in failure path (#10030 Jason Dillaman)
• librbd: enforce write ordering with a snapshot (Jason Dillaman)
• librbd: exclusive image locking (Jason Dillaman)
• librbd: fadvise API (Ma Jianpeng)
• librbd: fadvise-style hints; add misc hints for certain operations (Jianpeng Ma)
• librbd: fix and improve AIO cache invalidation (#10958 Jason Dillaman)
• librbd: fix cache tiers in list_children and snap_unprotect (Adam Crume)
• librbd: fix coverity false-positives (Jason Dillaman)
• librbd: fix diff test (#10002 Josh Durgin)
• librbd: fix list_children from invalid pool ioctxs (#10123 Jason Dillaman)
• librbd: fix locking for readahead (#10045 Jason Dillaman)
• librbd: fix memory leak (Jason Dillaman)
• librbd: fix ordering/queueing of resize operations (Jason Dillaman)
• librbd: fix performance regression in ObjectCacher (#9513 Adam Crume)
• librbd: fix snap create races (Jason Dillaman)
• librbd: fix write vs import race (#10590 Jason Dillaman)
• librbd: flush AIO operations asynchronously (#10714 Jason Dillaman)
• librbd: gracefully handle deleted/renamed pools (#10270 Jason Dillaman)
• librbd: lttng tracepoints (Adam Crume)
• librbd: make async versions of long-running maint operations (Jason Dillaman)
• librbd: misc fixes (Xinxin Shu, Jason Dillaman)
• librbd: mock tests (Jason Dillaman)
• librbd: only export public API symbols (Jason Dillaman)
• librbd: optionally blacklist clients before breaking locks (#10761 Jason Dillaman)
• librbd: prevent copyup during shrink (Jason Dillaman)
• librbd: refactor unit tests to use fixtures (Jason Dillaman)
• librbd: validate image is r/w on resize/flatten (Jason Dillaman)
• librbd: various internal locking fixes (Jason Dillaman)
• many coverity fixes (Danny Al-Gaaf)
• many many coverity cleanups (Danny Al-Gaaf)
• mds: ‘flush journal’ admin command (John Spray)
• mds: ENOSPC and OSDMap epoch barriers (#7317 John Spray)
• mds: a whole bunch of initial scrub infrastructure (Greg Farnum)
• mds: add cephfs-table-tool (John Spray)
• mds: asok command for fetching subtree map (John Spray)
• mds: avoid sending traceless replies in most cases (Yan, Zheng)
• mds: constify MDSCacheObjects (John Spray)
• mds: dirfrag buf fix (Yan, Zheng)
• mds: disallow most commands on inactive MDS’s (Greg Farnum)
• mds: drop dentries, leases on deleted directories (#10164 Yan, Zheng)
• mds: export dir asok command (John Spray)
• mds: fix MDLog IO callback deadlock (John Spray)
• mds: fix compat_version for MClientSession (#9945 John Spray)
• mds: fix deadlock during journal probe vs purge (#10229 Yan, Zheng)
• mds: fix race trimming log segments (Yan, Zheng)
• mds: fix reply snapbl (Yan, Zheng)
• mds: fix sessionmap lifecycle bugs (Yan, Zheng)
• mds: fix stray/purge perfcounters (#10388 John Spray)
• mds: handle heartbeat_reset during shutdown (#10382 John Spray)
• mds: handle zero-size xattr (#10335 Yan, Zheng)
• mds: initialize root inode xattr version (Yan, Zheng)
• mds: introduce auth caps (John Spray)
• mds: many many snapshot-related fixes (Yan, Zheng)
• mds: misc bugs (Greg Farnum, John Spray, Yan, Zheng, Henry Change)
• mds: refactor, improve Session storage (John Spray)
• mds: store backtrace for stray dir (Yan, Zheng)
• mds: subtree quota support (Yunchuan Wen)
• mds: verify backtrace when fetching dirfrag (#9557 Yan, Zheng)
• memstore: free space tracking (John Spray)
• misc cleanup (Danny Al-Gaaf, David Anderson)
• misc coverity fixes (Danny Al-Gaaf)
• misc coverity fixes (Danny Al-Gaaf)
• misc: various valgrind fixes and cleanups (Danny Al-Gaaf)
• mon: ‘osd crush reweight-all’ command (Sage Weil)
• mon: add ‘ceph osd rename-bucket …’ command (Loic Dachary)
• mon: add bootstrap-rgw profile (Sage Weil)
• mon: add max pgs per osd warning (Sage Weil)
• mon: add noforward flag for some mon commands (Mykola Golub)
• mon: allow adding tiers to fs pools (#10135 John Spray)
• mon: allow full flag to be manually cleared (#9323 Sage Weil)
• mon: clean up auth list output (Loic Dachary)
• mon: delay failure injection (Joao Eduardo Luis)
• mon: disallow empty pool names (#10555 Wido den Hollander)
• mon: do not deactivate last mds (#10862 John Spray)
• mon: do not pollute mon dir with CSV files from CRUSH check (Loic Dachary)
• mon: drop old ceph_mon_store_converter (Sage Weil)
• mon: fix ‘ceph pg dump_stuck degraded’ (Xinxin Shu)
• mon: fix ‘mds fail’ for standby MDSs (John Spray)
• mon: fix ‘osd crush link’ id resolution (John Spray)
• mon: fix ‘profile osd’ use of config-key function on mon (#10844 Joao Eduardo Luis)
• mon: fix _ratio units and types (Sage Weil)
• mon: fix JSON dumps to dump floats as flots and not strings (Sage Weil)
• mon: fix MDS health status from peons (#10151 John Spray)
• mon: fix caching for min_last_epoch_clean (#9987 Sage Weil)
• mon: fix clock drift time check interval (#10546 Joao Eduardo Luis)
• mon: fix compatset initalization during mkfs (Joao Eduardo Luis)
• mon: fix error output for add_data_pool (#9852 Joao Eduardo Luis)
• mon: fix feature tracking during elections (Joao Eduardo Luis)
• mon: fix formatter ‘pg stat’ command output (Sage Weil)
• mon: fix mds gid/rank/state parsing (John Spray)
• mon: fix misc error paths (Joao Eduardo Luis)
• mon: fix paxos off-by-one corner case (#9301 Sage Weil)
• mon: fix paxos timeouts (#10220 Joao Eduardo Luis)
• mon: fix stashed monmap encoding (#5203 Xie Rui)
• mon: fix units in store stats (Joao Eduardo Luis)
• mon: get canonical OSDMap from leader (#10422 Sage Weil)
• mon: ignore failure reports from before up_from (#10762 Dan van der Ster, Sage Weil)
• mon: implement ‘fs reset’ command (John Spray)
• mon: improve error handling on erasure code profile set (#10488, #11144 Loic Dachary)
• mon: improved corrupt CRUSH map detection (Joao Eduardo Luis)
• mon: include entity name in audit log for forwarded requests (#9913 Joao Eduardo Luis)
• mon: include pg_temp count in osdmap summary (Sage Weil)
• mon: log health summary to cluster log (#9440 Joao Eduardo Luis)
• mon: make ‘mds fail’ idempotent (John Spray)
• mon: make pg dump {sum,pgs,pgs_brief} work for format=plain (#5963 #6759 Mykola Golub)
• mon: new ‘ceph pool ls [detail]’ command (Sage Weil)
• mon: new pool safety flags nodelete, nopgchange, nosizechange (#9792 Mykola Golub)
• mon: new, friendly ‘ceph pg ls …’ command (Xinxin Shu)
• mon: paxos: allow reads while proposing (#9321 #9322 Joao Eduardo Luis)
• mon: prevent MDS transition from STOPPING (#10791 Greg Farnum)
• mon: propose all pending work in one transaction (Sage Weil)
• mon: remove pg_temps for nonexistent pools (Joao Eduardo Luis)
• mon: require mon_allow_pool_delete option to remove pools (Sage Weil)
• mon: respect down flag when promoting standbys (John Spray)
• mon: set globalid prealloc to larger value (Sage Weil)
• mon: set {read,write}_tier on ‘osd tier add-cache …’ (Jianpeng Ma)
• mon: skip zeroed osd stats in get_rule_avail (#10257 Joao Eduardo Luis)
• mon: validate min_size range (Jianpeng Ma)
• mon: wait for writeable before cross-proposing (#9794 Joao Eduardo Luis)
• mount.ceph: fix suprious error message (#10351 Yan, Zheng)
• ms: xio: fix misc bugs (Matt Benjamin, Vu Pham)
• msgr: async: bind threads to CPU cores, improved poll (Haomai Wang)
• msgr: async: many fixes, unit tests (Haomai Wang)
• msgr: async: several fixes (Haomai Wang)
• msgr: asyncmessenger: add kqueue support (#9926 Haomai Wang)
• msgr: avoid useless new/delete (Haomai Wang)
• msgr: fix RESETSESSION bug (#10080 Greg Farnum)
• msgr: fix crc configuration (Mykola Golub)
• msgr: fix delay injection bug (#9910 Sage Weil, Greg Farnum)
• msgr: misc unit tests (Haomai Wang)
• msgr: new AsymcMessenger alternative implementation (Haomai Wang)
• msgr: prefetch data when doing recv (Yehuda Sadeh)
• msgr: simple: fix rare deadlock (Greg Farnum)
• msgr: simple: retry binding to port on failure (#10029 Wido den Hollander)
• msgr: xio: XioMessenger RDMA support (Casey Bodley, Vu Pham, Matt Benjamin)
• objectstore: deprecate collection attrs (Sage Weil)
• osd, librados: fadvise-style librados hints (Jianpeng Ma)
• osd, librados: fix xattr_cmp_u64 (Dongmao Zhang)
• osd, librados: revamp PG listing API to handle namespaces (#9031 #9262 #9438 David Zafman)
• osd, mds: ‘ops’ as shorthand for ‘dump_ops_in_flight’ on asok (Sage Weil)
• osd, mon: add checksums to all OSDMaps (Sage Weil)
• osd, mon: send intiial pg create time from mon to osd (#9887 David Zafman)
• osd,mon: add ‘norebalance’ flag (Kefu Chai)
• osd,mon: specify OSD features explicitly in MOSDBoot (#10911 Sage Weil)
• osd: DBObjectMap: fix locking to prevent rare crash (#9891 Samuel Just)
• osd: EIO on whole-object reads when checksum is wrong (Sage Weil)
• osd: add erasure code corpus (Loic Dachary)
• osd: add fadvise flags to ObjectStore API (Jianpeng Ma)
• osd: add get_latest_osdmap asok command (#9483 #9484 Mykola Golub)
• osd: add misc tests (Loic Dachary, Danny Al-Gaaf)
• osd: add option to prioritize heartbeat network traffic (Jian Wen)
• osd: add support for the SHEC erasure-code algorithm (Takeshi Miyamae, Loic Dachary)
• osd: allow deletion of objects with watcher (#2339 Sage Weil)
• osd: allow recovery while below min_size (Samuel Just)
• osd: allow recovery with fewer than min_size OSDs (Samuel Just)
• osd: allow sparse read for Push/Pull (Haomai Wang)
• osd: allow whiteout deletion in cache pool (Sage Weil)
• osd: allow writes to degraded objects (Samuel Just)
• osd: allow writes to degraded objects (Samuel Just)
• osd: avoid publishing unchanged PG stats (Sage Weil)
• osd: batch pg log trim (Xinze Chi)
• osd: cache pool: ignore min flush age when cache is full (Xinze Chi)
• osd: cache recent ObjectContexts (Dong Yuan)
• osd: cache reverse_nibbles hash value (Dong Yuan)
• osd: clean up internal ObjectStore interface (Sage Weil)
• osd: cleanup boost optionals (William Kennington)
• osd: clear cache on interval change (Samuel Just)
• osd: do no proxy reads unless target OSDs are new (#10788 Sage Weil)
• osd: do not abort deep scrub on missing hinfo (#10018 Loic Dachary)
• osd: do not update digest on inconsistent object (#10524 Samuel Just)
• osd: don’t record digests for snapdirs (#10536 Samuel Just)
• osd: drop upgrade support for pre-dumpling (Sage Weil)
• osd: enable and use posix_fadvise (Sage Weil)
• osd: erasure coding: allow bench.sh to test ISA backend (Yuan Zhou)
• osd: erasure-code: encoding regression tests, corpus (#9420 Loic Dachary)
• osd: erasure-code: enforce chunk size alignment (#10211 Loic Dachary)
• osd: erasure-code: jerasure support for NEON (Loic Dachary)
• osd: erasure-code: relax cauchy w restrictions (#10325 David Zhang, Loic Dachary)
• osd: erasure-code: update gf-complete to latest upstream (Loic Dachary)
• osd: expose non-journal backends via ceph-osd CLI (Hoamai Wang)
• osd: filejournal: don’t cache journal when not using direct IO (Jianpeng Ma)
• osd: fix JSON output for stray OSDs (Loic Dachary)
• osd: fix OSDCap parser on old (el6) boost::spirit (#10757 Kefu Chai)
• osd: fix OSDCap parsing on el6 (#10757 Kefu Chai)
• osd: fix ObjectStore::Transaction encoding version (#10734 Samuel Just)
• osd: fix WBTHrottle perf counters (Haomai Wang)
• osd: fix and document last_epoch_started semantics (Samuel Just)
• osd: fix auth object selection during repair (#10524 Samuel Just)
• osd: fix backfill bug (#10150 Samuel Just)
• osd: fix bug in pending digest updates (#10840 Samuel Just)
• osd: fix cancel_proxy_read_ops (Sage Weil)
• osd: fix cleanup of interrupted pg deletion (#10617 Sage Weil)
• osd: fix divergent entry handling on PG split (Samuel Just)
• osd: fix ghobject_t formatted output to include shard (#10063 Loic Dachary)
• osd: fix ioprio option (Mykola Golub)
• osd: fix ioprio options (Loic Dachary)
• osd: fix journal shutdown race (Sage Weil)
• osd: fix journal wrapping bug (#10883 David Zafman)
• osd: fix leak in SnapTrimWQ (#10421 Kefu Chai)
• osd: fix leak on shutdown (Kefu Chai)
• osd: fix memstore free space calculation (Xiaoxi Chen)
• osd: fix mixed-version peering issues (Samuel Just)
• osd: fix object age eviction (Zhiqiang Wang)
• osd: fix object atime calculation (Xinze Chi)
• osd: fix object digest update bug (#10840 Samuel Just)
• osd: fix occasional peering stalls (#10431 Sage Weil)
• osd: fix ordering issue with new transaction encoding (#10534 Dong Yuan)
• osd: fix osd peer check on scrub messages (#9555 Sage Weil)
• osd: fix past_interval display bug (#9752 Loic Dachary)
• osd: fix past_interval generation (#10427 #10430 David Zafman)
• osd: fix pgls filter ops (#9439 David Zafman)
• osd: fix recording of digest on scrub (Samuel Just)
• osd: fix scrub delay bug (#10693 Samuel Just)
• osd: fix scrub vs try-flush bug (#8011 Samuel Just)
• osd: fix short read handling on push (#8121 David Zafman)
• osd: fix stderr with -f or -d (Dan Mick)
• osd: fix transaction accounting (Jianpeng Ma)
• osd: fix watch reconnect race (#10441 Sage Weil)
• osd: fix watch timeout cache state update (#10784 David Zafman)
• osd: fix whiteout handling (Sage Weil)
• osd: flush snapshots from cache tier immediately (Sage Weil)
• osd: force promotion of watch/notify ops (Zhiqiang Wang)
• osd: handle no-op write with snapshot (#10262 Sage Weil)
• osd: improve idempotency detection across cache promotion/demotion (#8935 Sage Weil, Samuel Just)
• osd: include activating peers in blocked_by (#10477 Sage Weil)
• osd: jerasure and gf-complete updates from upstream (#10216 Loic Dachary)
• osd: journal: check fsync/fdatasync result (Jianpeng Ma)
• osd: journal: fix alignment checks, avoid useless memmove (Jianpeng Ma)
• osd: journal: fix hang on shutdown (#10474 David Zafman)
• osd: journal: fix header.committed_up_to (Xinze Chi)
• osd: journal: fix journal zeroing when direct IO is enabled (Xie Rui)
• osd: journal: initialize throttle (Ning Yao)
• osd: journal: misc bug fixes (#6003 David Zafman, Samuel Just)
• osd: journal: update committed_thru after replay (#6756 Samuel Just)
• osd: keyvaluestore: cleanup dead code (Ning Yao)
• osd: keyvaluestore: fix getattr semantics (Haomai Wang)
• osd: keyvaluestore: fix key ordering (#10119 Haomai Wang)
• osd: keyvaluestore_dev: optimization (Chendi Xue)
• osd: limit in-flight read requests (Jason Dillaman)
• osd: log when scrub or repair starts (Loic Dachary)
• osd: make misdirected op checks robust for EC pools (#9835 Sage Weil)
• osd: memstore: fix size limit (Xiaoxi Chen)
• osd: misc FIEMAP fixes (Ma Jianpeng)
• osd: misc cleanup (Xinze Chi, Yongyue Sun)
• osd: misc optimizations (Xinxin Shu, Zhiqiang Wang, Xinze Chi)
• osd: misc scrub fixes (#10017 Loic Dachary)
• osd: new ‘activating’ state between peering and active (Sage Weil)
• osd: new optimized encoding for ObjectStore::Transaction (Dong Yuan)
• osd: optimize Finisher (Xinze Chi)
• osd: optimize WBThrottle map with unordered_map (Ning Yao)
• osd: optimize filter_snapc (Ning Yao)
• osd: preserve reqids for idempotency checks for promote/demote (Sage Weil, Zhiqiang Wang, Samuel Just)
• osd: proxy read support (Zhiqiang Wang)
• osd: proxy reads during cache promote (Zhiqiang Wang)
• osd: remove dead locking code (Xinxin Shu)
• osd: remove legacy classic scrub code (Sage Weil)
• osd: remove unused fields in MOSDSubOp (Xiaoxi Chen)
• osd: removed some dead code (Xinze Chi)
• osd: replace MOSDSubOp messages with simpler, optimized MOSDRepOp (Xiaoxi Chen)
• osd: restrict scrub to certain times of day (Xinze Chi)
• osd: rocksdb: fix shutdown (Hoamai Wang)
• osd: store PG metadata in per-collection objects for better concurrency (Sage Weil)
• osd: store whole-object checksums on scrub, write_full (Sage Weil)
• osd: support for discard for journal trim (Jianpeng Ma)
• osd: use FIEMAP_FLAGS_SYNC instead of fsync (Jianpeng Ma)
• osd: verify kernel is new enough before using XFS extsize ioctl, enable by default (#9956 Sage Weil)
• pybind: fix memory leak in librados bindings (Billy Olsen)
• pyrados: add object lock support (#6114 Mehdi Abaakouk)
• pyrados: fix misnamed wait_* routings (#10104 Dan Mick)
• pyrados: misc cleanups (Kefu Chai)
• qa: add large auth ticket tests (Ilya Dryomov)
• qa: fix mds tests (#10539 John Spray)
• qa: fix osd create dup tests (#10083 Loic Dachary)
• qa: ignore duplicates in rados ls (Josh Durgin)
• qa: improve hadoop tests (Noah Watkins)
• qa: many ‘make check’ improvements (Loic Dachary)
• qa: misc tests (Loic Dachary, Yan, Zheng)
• qa: parallelize make check (Loic Dachary)
• qa: reorg fs quota tests (Greg Farnum)
• qa: tolerate nearly-full disk for make check (Loic Dachary)
• rados: fix put of /dev/null (Loic Dachary)
• rados: fix usage (Jianpeng Ma)
• rados: parse command-line arguments more strictly (#8983 Adam Crume)
• rados: use copy-from operation for copy, cppool (Sage Weil)
• radosgw-admin: add replicalog update command (Yehuda Sadeh)
• rbd-fuse: clean up on shutdown (Josh Durgin)
• rbd-fuse: fix memory leak (Adam Crume)
• rbd-replay-many (Adam Crume)
• rbd-replay: –anonymize flag to rbd-replay-prep (Adam Crume)
• rbd: add ‘merge-diff’ function (MingXin Liu, Yunchuan Wen, Li Wang)
• rbd: allow v2 striping parameters for clones and imports (Jason Dillaman)
• rbd: fix ‘rbd diff’ for non-existent objects (Adam Crume)
• rbd: fix buffer handling on image import (#10590 Jason Dillaman)
• rbd: fix error when striping with format 1 (Sebastien Han)
• rbd: fix export for image sizes over 2GB (Vicente Cheng)
• rbd: fix formatted output of image features (Jason Dillaman)
• rbd: leave exclusive lockin goff by default (Jason Dillaman)
• rbd: updat eman page (Ilya Dryomov)
• rbd: update init-rbdmap to fix dup mount point (Karel Striegel)
• rbd: use IO hints for import, export, and bench operations (#10462 Jason Dillaman)
• rbd: use rolling average for rbd bench-write throughput (Jason Dillaman)
• rbd_recover_tool: RBD image recovery tool (Min Chen)
• rgw: S3-style object versioning support (Yehuda Sadeh)
• rgw: add location header when object is in another region (VRan Liu)
• rgw: change multipart upload id magic (#10271 Yehuda Sadeh)
• rgw: check keystone auth for S3 POST requests (#10062 Abhishek Lekshmanan)
• rgw: check timestamp on s3 keystone auth (#10062 Abhishek Lekshmanan)
• rgw: conditional PUT on ETag (#8562 Ray Lv)
• rgw: create subuser if needed when creating user (#10103 Yehuda Sadeh)
• rgw: decode http query params correction (#10271 Yehuda Sadeh)
• rgw: don’t overwrite bucket/object owner when setting ACLs (#10978 Yehuda Sadeh)
• rgw: enable IPv6 for civetweb (#10965 Yehuda Sadeh)
• rgw: extend replica log API (purge-all) (Yehuda Sadeh)
• rgw: fail S3 POST if keystone not configured (#10688 Valery Tschopp, Yehuda Sadeh)
• rgw: fix If-Modified-Since (VRan Liu)
• rgw: fix XML header on get ACL request (#10106 Yehuda Sadeh)
• rgw: fix bucket removal with data purge (Yehuda Sadeh)
• rgw: fix content length check (#10701 Axel Dunkel, Yehuda Sadeh)
• rgw: fix content-length update (#9576 Yehuda Sadeh)
• rgw: fix disabling of max_size quota (#9907 Dong Lei)
• rgw: fix error codes (#10334 #10329 Yehuda Sadeh)
• rgw: fix incorrect len when len is 0 (#9877 Yehuda Sadeh)
• rgw: fix object copy content type (#9478 Yehuda Sadeh)
• rgw: fix partial GET in swift (#10553 Yehuda Sadeh)
• rgw: fix replica log indexing (#8251 Yehuda Sadeh)
• rgw: fix shutdown (#10472 Yehuda Sadeh)
• rgw: fix swift metadata header name (Dmytro Iurchenko)
• rgw: fix sysvinit script when rgw_socket_path is not defined (#11159 Yehuda Sadeh, Dan Mick)
• rgw: fix user stags in get-user-info API (#9359 Ray Lv)
• rgw: include XML ns on get ACL request (#10106 Yehuda Sadeh)
• rgw: index swift keys appropriately (#10471 Yehuda Sadeh)
• rgw: make sysvinit script set ulimit -n properly (Sage Weil)
• rgw: misc fixes (#10307 Yehuda Sadeh)
• rgw: only track cleanup for objects we write (#10311 Yehuda Sadeh)
• rgw: pass civetweb configurables through (#10907 Yehuda Sadeh)
• rgw: prevent illegal bucket policy that doesn’t match placement rule (Yehuda Sadeh)
• rgw: remove multipart entries from bucket index on abort (#10719 Yehuda Sadeh)
• rgw: remove swift user manifest (DLO) hash calculation (#9973 Yehuda Sadeh)
• rgw: respond with 204 to POST on containers (#10667 Yuan Zhou)
• rgw: return timestamp on GET/HEAD (#8911 Yehuda Sadeh)
• rgw: reuse fcgx connection struct (#10194 Yehuda Sadeh)
• rgw: run radosgw as apache with systemd (#10125 Loic Dachary)
• rgw: send explicit HTTP status string (Yehuda Sadeh)
• rgw: set ETag on object copy (#9479 Yehuda Sadeh)
• rgw: set length for keystone token validation request (#7796 Yehuda Sadeh, Mark Kirkwood)
• rgw: support X-Storage-Policy header for Swift storage policy compat (Yehuda Sadeh)
• rgw: support multiple host names (#7467 Yehuda Sadeh)
• rgw: swift: dump container’s custom metadata (#10665 Ahmad Faheem, Dmytro Iurchenko)
• rgw: swift: support Accept header for response format (#10746 Dmytro Iurchenko)
• rgw: swift: support for X-Remove-Container-Meta-{key} (#10475 Dmytro Iurchenko)
• rgw: tweak error codes (#10329 #10334 Yehuda Sadeh)
• rgw: update bucket index on attr changes, for multi-site sync (#5595 Yehuda Sadeh)
• rgw: use rn for http headers (#9254 Yehuda Sadeh)
• rgw: use gc for multipart abort (#10445 Aaron Bassett, Yehuda Sadeh)
• rgw: use new watch/notify API (Yehuda Sadeh, Sage Weil)
• rpm: misc fixes (Key Dreyer)
• rpm: move rgw logrotate to radosgw subpackage (Ken Dreyer)
• systemd: better systemd unit files (Owen Synge)
• sysvinit: fix race in ‘stop’ (#10389 Loic Dachary)
• test: fix bufferlist tests (Jianpeng Ma)
• tests: ability to run unit tests under docker (Loic Dachary)
• tests: centos-6 dockerfile (#10755 Loic Dachary)
• tests: improve docker-based tests (Loic Dachary)
• tests: unit tests for shared_cache (Dong Yuan)
• udev: fix rules for CentOS7/RHEL7 (Loic Dachary)
• use clock_gettime instead of gettimeofday (Jianpeng Ma)
• vstart.sh: set up environment for s3-tests (Luis Pabon)
• vstart.sh: work with cmake (Yehuda Sadeh)

v0.93

This is the first release candidate for Hammer, and includes all ofthe features that will be present in the final release. We welcomeand encourage any and all testing in non-production clusters to identifyany problems with functionality, stability, or performance before thefinal Hammer release.

We suggest some caution in one area: librbd. There is a lot of newfunctionality around object maps and locking that is disabled bydefault but may still affect stability for existing images. We arecontinuing to shake out those bugs so that the final Hammer release(probably v0.94) will be rock solid.

Major features since Giant include:

• cephfs: journal scavenger repair tool (John Spray)
• crush: new and improved straw2 bucket type (Sage Weil, Christina Anderson, Xiaoxi Chen)
• doc: improved guidance for CephFS early adopters (John Spray)
• librbd: add per-image object map for improved performance (Jason Dillaman)
• librbd: copy-on-read (Min Chen, Li Wang, Yunchuan Wen, Cheng Cheng)
• librados: fadvise-style IO hints (Jianpeng Ma)
• mds: many many snapshot-related fixes (Yan, Zheng)
• mon: new ‘ceph osd df’ command (Mykola Golub)
• mon: new ‘ceph pg ls …’ command (Xinxin Shu)
• osd: improved performance for high-performance backends
• osd: improved recovery behavior (Samuel Just)
• osd: improved cache tier behavior with reads (Zhiqiang Wang)
• rgw: S3-compatible bucket versioning support (Yehuda Sadeh)
• rgw: large bucket index sharding (Guang Yang, Yehuda Sadeh)
• RDMA “xio” messenger support (Matt Benjamin, Vu Pham)

Upgrading

• If you are upgrading from v0.92, you must stop all OSD daemons and flush theirjournals (ceph-osd-iNNN—flush-journal) before upgrading. There wasa transaction encoding bug in v0.92 that broke compatibility. Upgrading fromv0.91 or anything earlier is safe.
• No special restrictions when upgrading from firefly or giant.

Notable Changes

• build: CMake support (Ali Maredia, Casey Bodley, Adam Emerson, Marcus Watts, Matt Benjamin)
• ceph-disk: do not re-use partition if encryption is required (Loic Dachary)
• ceph-disk: support LUKS for encrypted partitions (Andrew Bartlett, Loic Dachary)
• ceph-fuse,libcephfs: add support for O_NOFOLLOW and O_PATH (Greg Farnum)
• ceph-fuse,libcephfs: resend requests before completing cap reconnect (#10912 Yan, Zheng)
• ceph-fuse: select kernel cache invalidation mechanism based on kernel version (Greg Farnum)
• ceph-objectstore-tool: improved import (David Zafman)
• ceph-objectstore-tool: misc improvements, fixes (#9870 #9871 David Zafman)
• ceph: add ‘ceph osd df [tree]’ command (#10452 Mykola Golub)
• ceph: fix ‘ceph tell …’ command validation (#10439 Joao Eduardo Luis)
• ceph: improve ‘ceph osd tree’ output (Mykola Golub)
• cephfs-journal-tool: add recover_dentries function (#9883 John Spray)
• common: add newline to flushed json output (Sage Weil)
• common: filtering for ‘perf dump’ (John Spray)
• common: fix Formatter factory breakage (#10547 Loic Dachary)
• common: make json-pretty output prettier (Sage Weil)
• crush: new and improved straw2 bucket type (Sage Weil, Christina Anderson, Xiaoxi Chen)
• crush: update tries stats for indep rules (#10349 Loic Dachary)
• crush: use larger choose_tries value for erasure code rulesets (#10353 Loic Dachary)
• debian,rpm: move RBD udev rules to ceph-common (#10864 Ken Dreyer)
• debian: split python-ceph into python-{rbd,rados,cephfs} (Boris Ranto)
• doc: CephFS disaster recovery guidance (John Spray)
• doc: CephFS for early adopters (John Spray)
• doc: fix OpenStack Glance docs (#10478 Sebastien Han)
• doc: misc updates (#9793 #9922 #10204 #10203 Travis Rhoden, Hazem, Ayari, Florian Coste, Andy Allan, Frank Yu, Baptiste Veuillez-Mainard, Yuan Zhou, Armando Segnini, Robert Jansen, Tyler Brekke, Viktor Suprun)
• doc: replace cloudfiles with swiftclient Python Swift example (Tim Freund)
• erasure-code: add mSHEC erasure code support (Takeshi Miyamae)
• erasure-code: improved docs (#10340 Loic Dachary)
• erasure-code: set max_size to 20 (#10363 Loic Dachary)
• libcephfs,ceph-fuse: fix getting zero-length xattr (#10552 Yan, Zheng)
• librados: add blacklist_add convenience method (Jason Dillaman)
• librados: expose rados_{read|write}_op_assert_version in C API (Kim Vandry)
• librados: fix pool name caching (#10458 Radoslaw Zarzynski)
• librados: fix resource leak, misc bugs (#10425 Radoslaw Zarzynski)
• librados: fix some watch/notify locking (Jason Dillaman, Josh Durgin)
• libradosstriper: fix write_full when ENOENT (#10758 Sebastien Ponce)
• librbd: CRC protection for RBD image map (Jason Dillaman)
• librbd: add per-image object map for improved performance (Jason Dillaman)
• librbd: add support for an “object map” indicating which objects exist (Jason Dillaman)
• librbd: adjust internal locking (Josh Durgin, Jason Dillaman)
• librbd: better handling of watch errors (Jason Dillaman)
• librbd: coordinate maint operations through lock owner (Jason Dillaman)
• librbd: copy-on-read (Min Chen, Li Wang, Yunchuan Wen, Cheng Cheng, Jason Dillaman)
• librbd: enforce write ordering with a snapshot (Jason Dillaman)
• librbd: fadvise-style hints; add misc hints for certain operations (Jianpeng Ma)
• librbd: fix coverity false-positives (Jason Dillaman)
• librbd: fix snap create races (Jason Dillaman)
• librbd: flush AIO operations asynchronously (#10714 Jason Dillaman)
• librbd: make async versions of long-running maint operations (Jason Dillaman)
• librbd: mock tests (Jason Dillaman)
• librbd: optionally blacklist clients before breaking locks (#10761 Jason Dillaman)
• librbd: prevent copyup during shrink (Jason Dillaman)
• mds: add cephfs-table-tool (John Spray)
• mds: avoid sending traceless replies in most cases (Yan, Zheng)
• mds: export dir asok command (John Spray)
• mds: fix stray/purge perfcounters (#10388 John Spray)
• mds: handle heartbeat_reset during shutdown (#10382 John Spray)
• mds: many many snapshot-related fixes (Yan, Zheng)
• mds: refactor, improve Session storage (John Spray)
• misc coverity fixes (Danny Al-Gaaf)
• mon: add noforward flag for some mon commands (Mykola Golub)
• mon: disallow empty pool names (#10555 Wido den Hollander)
• mon: do not deactivate last mds (#10862 John Spray)
• mon: drop old ceph_mon_store_converter (Sage Weil)
• mon: fix ‘ceph pg dump_stuck degraded’ (Xinxin Shu)
• mon: fix ‘profile osd’ use of config-key function on mon (#10844 Joao Eduardo Luis)
• mon: fix compatset initalization during mkfs (Joao Eduardo Luis)
• mon: fix feature tracking during elections (Joao Eduardo Luis)
• mon: fix mds gid/rank/state parsing (John Spray)
• mon: ignore failure reports from before up_from (#10762 Dan van der Ster, Sage Weil)
• mon: improved corrupt CRUSH map detection (Joao Eduardo Luis)
• mon: include pg_temp count in osdmap summary (Sage Weil)
• mon: log health summary to cluster log (#9440 Joao Eduardo Luis)
• mon: make ‘mds fail’ idempotent (John Spray)
• mon: make pg dump {sum,pgs,pgs_brief} work for format=plain (#5963 #6759 Mykola Golub)
• mon: new pool safety flags nodelete, nopgchange, nosizechange (#9792 Mykola Golub)
• mon: new, friendly ‘ceph pg ls …’ command (Xinxin Shu)
• mon: prevent MDS transition from STOPPING (#10791 Greg Farnum)
• mon: propose all pending work in one transaction (Sage Weil)
• mon: remove pg_temps for nonexistent pools (Joao Eduardo Luis)
• mon: require mon_allow_pool_delete option to remove pools (Sage Weil)
• mon: set globalid prealloc to larger value (Sage Weil)
• mon: skip zeroed osd stats in get_rule_avail (#10257 Joao Eduardo Luis)
• mon: validate min_size range (Jianpeng Ma)
• msgr: async: bind threads to CPU cores, improved poll (Haomai Wang)
• msgr: fix crc configuration (Mykola Golub)
• msgr: misc unit tests (Haomai Wang)
• msgr: xio: XioMessenger RDMA support (Casey Bodley, Vu Pham, Matt Benjamin)
• osd, librados: fadvise-style librados hints (Jianpeng Ma)
• osd, librados: fix xattr_cmp_u64 (Dongmao Zhang)
• osd,mon: add ‘norebalance’ flag (Kefu Chai)
• osd,mon: specify OSD features explicitly in MOSDBoot (#10911 Sage Weil)
• osd: add option to prioritize heartbeat network traffic (Jian Wen)
• osd: add support for the SHEC erasure-code algorithm (Takeshi Miyamae, Loic Dachary)
• osd: allow recovery while below min_size (Samuel Just)
• osd: allow recovery with fewer than min_size OSDs (Samuel Just)
• osd: allow writes to degraded objects (Samuel Just)
• osd: allow writes to degraded objects (Samuel Just)
• osd: avoid publishing unchanged PG stats (Sage Weil)
• osd: cache recent ObjectContexts (Dong Yuan)
• osd: clear cache on interval change (Samuel Just)
• osd: do no proxy reads unless target OSDs are new (#10788 Sage Weil)
• osd: do not update digest on inconsistent object (#10524 Samuel Just)
• osd: don’t record digests for snapdirs (#10536 Samuel Just)
• osd: fix OSDCap parser on old (el6) boost::spirit (#10757 Kefu Chai)
• osd: fix OSDCap parsing on el6 (#10757 Kefu Chai)
• osd: fix ObjectStore::Transaction encoding version (#10734 Samuel Just)
• osd: fix auth object selection during repair (#10524 Samuel Just)
• osd: fix bug in pending digest updates (#10840 Samuel Just)
• osd: fix cancel_proxy_read_ops (Sage Weil)
• osd: fix cleanup of interrupted pg deletion (#10617 Sage Weil)
• osd: fix journal wrapping bug (#10883 David Zafman)
• osd: fix leak in SnapTrimWQ (#10421 Kefu Chai)
• osd: fix memstore free space calculation (Xiaoxi Chen)
• osd: fix mixed-version peering issues (Samuel Just)
• osd: fix object digest update bug (#10840 Samuel Just)
• osd: fix ordering issue with new transaction encoding (#10534 Dong Yuan)
• osd: fix past_interval generation (#10427 #10430 David Zafman)
• osd: fix short read handling on push (#8121 David Zafman)
• osd: fix watch timeout cache state update (#10784 David Zafman)
• osd: force promotion of watch/notify ops (Zhiqiang Wang)
• osd: improve idempotency detection across cache promotion/demotion (#8935 Sage Weil, Samuel Just)
• osd: include activating peers in blocked_by (#10477 Sage Weil)
• osd: jerasure and gf-complete updates from upstream (#10216 Loic Dachary)
• osd: journal: check fsync/fdatasync result (Jianpeng Ma)
• osd: journal: fix hang on shutdown (#10474 David Zafman)
• osd: journal: fix header.committed_up_to (Xinze Chi)
• osd: journal: initialize throttle (Ning Yao)
• osd: journal: misc bug fixes (#6003 David Zafman, Samuel Just)
• osd: misc cleanup (Xinze Chi, Yongyue Sun)
• osd: new ‘activating’ state between peering and active (Sage Weil)
• osd: preserve reqids for idempotency checks for promote/demote (Sage Weil, Zhiqiang Wang, Samuel Just)
• osd: remove dead locking code (Xinxin Shu)
• osd: restrict scrub to certain times of day (Xinze Chi)
• osd: rocksdb: fix shutdown (Hoamai Wang)
• pybind: fix memory leak in librados bindings (Billy Olsen)
• qa: fix mds tests (#10539 John Spray)
• qa: ignore duplicates in rados ls (Josh Durgin)
• qa: improve hadoop tests (Noah Watkins)
• qa: reorg fs quota tests (Greg Farnum)
• rados: fix usage (Jianpeng Ma)
• radosgw-admin: add replicalog update command (Yehuda Sadeh)
• rbd-fuse: clean up on shutdown (Josh Durgin)
• rbd: add ‘merge-diff’ function (MingXin Liu, Yunchuan Wen, Li Wang)
• rbd: fix buffer handling on image import (#10590 Jason Dillaman)
• rbd: leave exclusive lockin goff by default (Jason Dillaman)
• rbd: update init-rbdmap to fix dup mount point (Karel Striegel)
• rbd: use IO hints for import, export, and bench operations (#10462 Jason Dillaman)
• rbd_recover_tool: RBD image recovery tool (Min Chen)
• rgw: S3-style object versioning support (Yehuda Sadeh)
• rgw: check keystone auth for S3 POST requests (#10062 Abhishek Lekshmanan)
• rgw: extend replica log API (purge-all) (Yehuda Sadeh)
• rgw: fail S3 POST if keystone not configured (#10688 Valery Tschopp, Yehuda Sadeh)
• rgw: fix XML header on get ACL request (#10106 Yehuda Sadeh)
• rgw: fix bucket removal with data purge (Yehuda Sadeh)
• rgw: fix replica log indexing (#8251 Yehuda Sadeh)
• rgw: fix swift metadata header name (Dmytro Iurchenko)
• rgw: remove multipart entries from bucket index on abort (#10719 Yehuda Sadeh)
• rgw: respond with 204 to POST on containers (#10667 Yuan Zhou)
• rgw: reuse fcgx connection struct (#10194 Yehuda Sadeh)
• rgw: support multiple host names (#7467 Yehuda Sadeh)
• rgw: swift: dump container’s custom metadata (#10665 Ahmad Faheem, Dmytro Iurchenko)
• rgw: swift: support Accept header for response format (#10746 Dmytro Iurchenko)
• rgw: swift: support for X-Remove-Container-Meta-{key} (#10475 Dmytro Iurchenko)
• rpm: move rgw logrotate to radosgw subpackage (Ken Dreyer)
• tests: centos-6 dockerfile (#10755 Loic Dachary)
• tests: unit tests for shared_cache (Dong Yuan)
• vstart.sh: work with cmake (Yehuda Sadeh)

v0.92

This is the second-to-last chunk of new stuff before Hammer. Big itemsinclude additional checksums on OSD objects, proxied reads in thecache tier, image locking in RBD, optimized OSD Transaction andreplication messages, and a big pile of RGW and MDS bug fixes.

Upgrading

• The experimental ‘keyvaluestore-dev’ OSD backend has been renamed‘keyvaluestore’ (for simplicity) and marked as experimental. Toenable this untested feature and acknowledge that you understandthat it is untested and may destroy data, you need to add thefollowing to your ceph.conf:

enable experimental unrecoverable data corrupting featuers = keyvaluestore

• The following librados C API function calls take a ‘flags’ argument whose valueis now correctly interpreted:

rados_write_op_operate()rados_aio_write_op_operate()rados_read_op_operate()rados_aio_read_op_operate()

The flags were not correctly being translated from the librados constants to theinternal values. Now they are. Any code that is passing flags to these methodsshould be audited to ensure that they are using the correct LIBRADOSOP_FLAG*constants.

• The ‘rados’ CLI ‘copy’ and ‘cppool’ commands now use the copy-from operation,which means the latest CLI cannot run these commands against pre-firefly OSDs.

• The librados watch/notify API now includes a watch_flush() operation to flushthe async queue of notify operations. This should be called by any watch/notifyuser prior to rados_shutdown().

Notable Changes

• add experimental features option (Sage Weil)
• build: fix ‘make check’ races (#10384 Loic Dachary)
• build: fix pkg names when libkeyutils is missing (Pankag Garg, Ken Dreyer)
• ceph: make ‘ceph -s’ show PG state counts in sorted order (Sage Weil)
• ceph: make ‘ceph tell mon.* version’ work (Mykola Golub)
• ceph-monstore-tool: fix/improve CLI (Joao Eduardo Luis)
• ceph: show primary-affinity in ‘ceph osd tree’ (Mykola Golub)
• common: add TableFormatter (Andreas Peters)
• common: check syncfs() return code (Jianpeng Ma)
• doc: do not suggest dangerous XFS nobarrier option (Dan van der Ster)
• doc: misc updates (Nilamdyuti Goswami, John Wilkins)
• install-deps.sh: do not require sudo when root (Loic Dachary)
• libcephfs: fix dirfrag trimming (#10387 Yan, Zheng)
• libcephfs: fix mount timeout (#10041 Yan, Zheng)
• libcephfs: fix test (#10415 Yan, Zheng)
• libcephfs: fix use-afer-free on umount (#10412 Yan, Zheng)
• libcephfs: include ceph and git version in client metadata (Sage Weil)
• librados: add watch_flush() operation (Sage Weil, Haomai Wang)
• librados: avoid memcpy on getxattr, read (Jianpeng Ma)
• librados: create ioctx by pool id (Jason Dillaman)
• librados: do notify completion in fast-dispatch (Sage Weil)
• librados: remove shadowed variable (Kefu Chain)
• librados: translate op flags from C APIs (Matthew Richards)
• librbd: differentiate between R/O vs R/W features (Jason Dillaman)
• librbd: exclusive image locking (Jason Dillaman)
• librbd: fix write vs import race (#10590 Jason Dillaman)
• librbd: gracefully handle deleted/renamed pools (#10270 Jason Dillaman)
• mds: asok command for fetching subtree map (John Spray)
• mds: constify MDSCacheObjects (John Spray)
• misc: various valgrind fixes and cleanups (Danny Al-Gaaf)
• mon: fix ‘mds fail’ for standby MDSs (John Spray)
• mon: fix stashed monmap encoding (#5203 Xie Rui)
• mon: implement ‘fs reset’ command (John Spray)
• mon: respect down flag when promoting standbys (John Spray)
• mount.ceph: fix suprious error message (#10351 Yan, Zheng)
• msgr: async: many fixes, unit tests (Haomai Wang)
• msgr: simple: retry binding to port on failure (#10029 Wido den Hollander)
• osd: add fadvise flags to ObjectStore API (Jianpeng Ma)
• osd: add get_latest_osdmap asok command (#9483 #9484 Mykola Golub)
• osd: EIO on whole-object reads when checksum is wrong (Sage Weil)
• osd: filejournal: don’t cache journal when not using direct IO (Jianpeng Ma)
• osd: fix ioprio option (Mykola Golub)
• osd: fix scrub delay bug (#10693 Samuel Just)
• osd: fix watch reconnect race (#10441 Sage Weil)
• osd: handle no-op write with snapshot (#10262 Sage Weil)
• osd: journal: fix journal zeroing when direct IO is enabled (Xie Rui)
• osd: keyvaluestore: cleanup dead code (Ning Yao)
• osd, mds: ‘ops’ as shorthand for ‘dump_ops_in_flight’ on asok (Sage Weil)
• osd: memstore: fix size limit (Xiaoxi Chen)
• osd: misc scrub fixes (#10017 Loic Dachary)
• osd: new optimized encoding for ObjectStore::Transaction (Dong Yuan)
• osd: optimize filter_snapc (Ning Yao)
• osd: optimize WBThrottle map with unordered_map (Ning Yao)
• osd: proxy reads during cache promote (Zhiqiang Wang)
• osd: proxy read support (Zhiqiang Wang)
• osd: remove legacy classic scrub code (Sage Weil)
• osd: remove unused fields in MOSDSubOp (Xiaoxi Chen)
• osd: replace MOSDSubOp messages with simpler, optimized MOSDRepOp (Xiaoxi Chen)
• osd: store whole-object checksums on scrub, write_full (Sage Weil)
• osd: verify kernel is new enough before using XFS extsize ioctl, enable by default (#9956 Sage Weil)
• rados: use copy-from operation for copy, cppool (Sage Weil)
• rgw: change multipart upload id magic (#10271 Yehuda Sadeh)
• rgw: decode http query params correction (#10271 Yehuda Sadeh)
• rgw: fix content length check (#10701 Axel Dunkel, Yehuda Sadeh)
• rgw: fix partial GET in swift (#10553 Yehuda Sadeh)
• rgw: fix shutdown (#10472 Yehuda Sadeh)
• rgw: include XML ns on get ACL request (#10106 Yehuda Sadeh)
• rgw: misc fixes (#10307 Yehuda Sadeh)
• rgw: only track cleanup for objects we write (#10311 Yehuda Sadeh)
• rgw: tweak error codes (#10329 #10334 Yehuda Sadeh)
• rgw: use gc for multipart abort (#10445 Aaron Bassett, Yehuda Sadeh)
• sysvinit: fix race in ‘stop’ (#10389 Loic Dachary)
• test: fix bufferlist tests (Jianpeng Ma)
• tests: improve docker-based tests (Loic Dachary)

v0.91

We are quickly approaching the Hammer feature freeze but have a fewmore dev releases to go before we get there. The headline items aresubtree-based quota support in CephFS (ceph-fuse/libcephfs clientsupport only for now), a rewrite of the watch/notify librados API usedby RBD and RGW, OSDMap checksums to ensure that maps are alwaysconsistent inside the cluster, new API calls in librados and librbdfor IO hinting modeled after posix_fadvise, and improved storage ofper-PG state.

We expect two more releases before the Hammer feature freeze (v0.93).

Upgrading

• The ‘category’ field for objects has been removed. This was originally addedto track PG stat summations over different categories of objects for use byradosgw. It is no longer has any known users and is prone to abuse because itcan lead to a pg_stat_t structure that is unbounded. The librados API callsthat accept this field now ignore it, and the OSD no longers tracks theper-category summations.
• The output for ‘rados df’ has changed. The ‘category’ level has beeneliminated, so there is now a single stat object per pool. The structure ofthe JSON output is different, and the plaintext output has one less column.
• The ‘rados create [category]’ optional category argument is nolonger supported or recognized.
• rados.py’s Rados class no longer has a del method; it was causingproblems on interpreter shutdown and use of threads. If your code hasRados objects with limited lifetimes and you’re concerned about lockedresources, call Rados.shutdown() explicitly.
• There is a new version of the librados watch/notify API with vastlyimproved semantics. Any applications using this interface areencouraged to migrate to the new API. The old API calls are markedas deprecated and will eventually be removed.
• The librados rados_unwatch() call used to be safe to call on aninvalid handle. The new version has undefined behavior when passeda bogus value (for example, when rados_watch() returns an error andhandle is not defined).
• The structure of the formatted ‘pg stat’ command is changed for theportion that counts states by name to avoid using the ‘+’ character(which appears in state names) as part of the XML token (it is notlegal).

Notable Changes

• asyncmsgr: misc fixes (Haomai Wang)
• buffer: add ‘shareable’ construct (Matt Benjamin)
• build: aarch64 build fixes (Noah Watkins, Haomai Wang)
• build: support for jemalloc (Shishir Gowda)
• ceph-disk: allow journal partition re-use (#10146 Loic Dachary, Dav van der Ster)
• ceph-disk: misc fixes (Christos Stavrakakis)
• ceph-fuse: fix kernel cache trimming (#10277 Yan, Zheng)
• ceph-objectstore-tool: many many improvements (David Zafman)
• common: support new gperftools header locations (Key Dreyer)
• crush: straw bucket weight calculation fixes (#9998 Sage Weil)
• doc: misc improvements (Nilamdyuti Goswami, John Wilkins, Chris Holcombe)
• libcephfs,ceph-fuse: add ‘status’ asok (John Spray)
• librados, osd: new watch/notify implementation (Sage Weil)
• librados: drop ‘category’ feature (Sage Weil)
• librados: fix pool deletion handling (#10372 Sage Weil)
• librados: new fadvise API (Ma Jianpeng)
• libradosstriper: fix remove() (Dongmao Zhang)
• librbd: complete pending ops before closing image (#10299 Josh Durgin)
• librbd: fadvise API (Ma Jianpeng)
• mds: ENOSPC and OSDMap epoch barriers (#7317 John Spray)
• mds: dirfrag buf fix (Yan, Zheng)
• mds: disallow most commands on inactive MDS’s (Greg Farnum)
• mds: drop dentries, leases on deleted directories (#10164 Yan, Zheng)
• mds: handle zero-size xattr (#10335 Yan, Zheng)
• mds: subtree quota support (Yunchuan Wen)
• memstore: free space tracking (John Spray)
• misc cleanup (Danny Al-Gaaf, David Anderson)
• mon: ‘osd crush reweight-all’ command (Sage Weil)
• mon: allow full flag to be manually cleared (#9323 Sage Weil)
• mon: delay failure injection (Joao Eduardo Luis)
• mon: fix paxos timeouts (#10220 Joao Eduardo Luis)
• mon: get canonical OSDMap from leader (#10422 Sage Weil)
• msgr: fix RESETSESSION bug (#10080 Greg Farnum)
• objectstore: deprecate collection attrs (Sage Weil)
• osd, mon: add checksums to all OSDMaps (Sage Weil)
• osd: allow deletion of objects with watcher (#2339 Sage Weil)
• osd: allow sparse read for Push/Pull (Haomai Wang)
• osd: cache reverse_nibbles hash value (Dong Yuan)
• osd: drop upgrade support for pre-dumpling (Sage Weil)
• osd: enable and use posix_fadvise (Sage Weil)
• osd: erasure-code: enforce chunk size alignment (#10211 Loic Dachary)
• osd: erasure-code: jerasure support for NEON (Loic Dachary)
• osd: erasure-code: relax cauchy w restrictions (#10325 David Zhang, Loic Dachary)
• osd: erasure-code: update gf-complete to latest upstream (Loic Dachary)
• osd: fix WBTHrottle perf counters (Haomai Wang)
• osd: fix backfill bug (#10150 Samuel Just)
• osd: fix occasional peering stalls (#10431 Sage Weil)
• osd: fix scrub vs try-flush bug (#8011 Samuel Just)
• osd: fix stderr with -f or -d (Dan Mick)
• osd: misc FIEMAP fixes (Ma Jianpeng)
• osd: optimize Finisher (Xinze Chi)
• osd: store PG metadata in per-collection objects for better concurrency (Sage Weil)
• pyrados: add object lock support (#6114 Mehdi Abaakouk)
• pyrados: fix misnamed wait_* routings (#10104 Dan Mick)
• pyrados: misc cleanups (Kefu Chai)
• qa: add large auth ticket tests (Ilya Dryomov)
• qa: many ‘make check’ improvements (Loic Dachary)
• qa: misc tests (Loic Dachary, Yan, Zheng)
• rgw: conditional PUT on ETag (#8562 Ray Lv)
• rgw: fix error codes (#10334 #10329 Yehuda Sadeh)
• rgw: index swift keys appropriately (#10471 Yehuda Sadeh)
• rgw: prevent illegal bucket policy that doesn’t match placement rule (Yehuda Sadeh)
• rgw: run radosgw as apache with systemd (#10125 Loic Dachary)
• rgw: support X-Storage-Policy header for Swift storage policy compat (Yehuda Sadeh)
• rgw: use rn for http headers (#9254 Yehuda Sadeh)
• rpm: misc fixes (Key Dreyer)

v0.90

This is the last development release before Christmas. There are someAPI cleanups for librados and librbd, and lots of bug fixes across theboard for the OSD, MDS, RGW, and CRUSH. The OSD also gets support fordiscard (potentially helpful on SSDs, although it is off by default), and thereare several improvements to ceph-disk.

The next two development releases will be getting a slew of newfunctionality for hammer. Stay tuned!

Upgrading

• Previously, the formatted output of ‘ceph pg stat -f …’ was a fullpg dump that included all metadata about all PGs in the system. Itis now a concise summary of high-level PG stats, just like theunformatted ‘ceph pg stat’ command.
• All JSON dumps of floating point values were incorrecting surrounding thevalue with quotes. These quotes have been removed. Any consumer of structuredJSON output that was consuming the floating point values was previously havingto interpret the quoted string and will most likely need to be fixed to takethe unquoted number.

Notable Changes

• arch: fix NEON feaeture detection (#10185 Loic Dachary)
• build: adjust build deps for yasm, virtualenv (Jianpeng Ma)
• build: improve build dependency tooling (Loic Dachary)
• ceph-disk: call partx/partprobe consistency (#9721 Loic Dachary)
• ceph-disk: fix dmcrypt key permissions (Loic Dachary)
• ceph-disk: fix umount race condition (#10096 Blaine Gardner)
• ceph-disk: init=none option (Loic Dachary)
• ceph-monstore-tool: fix shutdown (#10093 Loic Dachary)
• ceph-objectstore-tool: fix import (#10090 David Zafman)
• ceph-objectstore-tool: many improvements and tests (David Zafman)
• ceph.spec: package rbd-replay-prep (Ken Dreyer)
• common: add ‘perf reset …’ admin command (Jianpeng Ma)
• common: do not unlock rwlock on destruction (Federico Simoncelli)
• common: fix block device discard check (#10296 Sage Weil)
• common: remove broken CEPH_LOCKDEP optoin (Kefu Chai)
• crush: fix tree bucket behavior (Rongze Zhu)
• doc: add build-doc guidlines for Fedora and CentOS/RHEL (Nilamdyuti Goswami)
• doc: enable rbd cache on openstack deployments (Sebastien Han)
• doc: improved installation nots on CentOS/RHEL installs (John Wilkins)
• doc: misc cleanups (Adam Spiers, Sebastien Han, Nilamdyuti Goswami, Ken Dreyer, John Wilkins)
• doc: new man pages (Nilamdyuti Goswami)
• doc: update release descriptions (Ken Dreyer)
• doc: update sepia hardware inventory (Sandon Van Ness)
• librados: only export public API symbols (Jason Dillaman)
• libradosstriper: fix stat strtoll (Dongmao Zhang)
• libradosstriper: fix trunc method (#10129 Sebastien Ponce)
• librbd: fix list_children from invalid pool ioctxs (#10123 Jason Dillaman)
• librbd: only export public API symbols (Jason Dillaman)
• many coverity fixes (Danny Al-Gaaf)
• mds: ‘flush journal’ admin command (John Spray)
• mds: fix MDLog IO callback deadlock (John Spray)
• mds: fix deadlock during journal probe vs purge (#10229 Yan, Zheng)
• mds: fix race trimming log segments (Yan, Zheng)
• mds: store backtrace for stray dir (Yan, Zheng)
• mds: verify backtrace when fetching dirfrag (#9557 Yan, Zheng)
• mon: add max pgs per osd warning (Sage Weil)
• mon: fix _ratio units and types (Sage Weil)
• mon: fix JSON dumps to dump floats as flots and not strings (Sage Weil)
• mon: fix formatter ‘pg stat’ command output (Sage Weil)
• msgr: async: several fixes (Haomai Wang)
• msgr: simple: fix rare deadlock (Greg Farnum)
• osd: batch pg log trim (Xinze Chi)
• osd: clean up internal ObjectStore interface (Sage Weil)
• osd: do not abort deep scrub on missing hinfo (#10018 Loic Dachary)
• osd: fix ghobject_t formatted output to include shard (#10063 Loic Dachary)
• osd: fix osd peer check on scrub messages (#9555 Sage Weil)
• osd: fix pgls filter ops (#9439 David Zafman)
• osd: flush snapshots from cache tier immediately (Sage Weil)
• osd: keyvaluestore: fix getattr semantics (Haomai Wang)
• osd: keyvaluestore: fix key ordering (#10119 Haomai Wang)
• osd: limit in-flight read requests (Jason Dillaman)
• osd: log when scrub or repair starts (Loic Dachary)
• osd: support for discard for journal trim (Jianpeng Ma)
• qa: fix osd create dup tests (#10083 Loic Dachary)
• rgw: add location header when object is in another region (VRan Liu)
• rgw: check timestamp on s3 keystone auth (#10062 Abhishek Lekshmanan)
• rgw: make sysvinit script set ulimit -n properly (Sage Weil)
• systemd: better systemd unit files (Owen Synge)
• tests: ability to run unit tests under docker (Loic Dachary)

v0.89

This is the second development release since Giant. The big itemsinclude the first batch of scrub patchs from Greg for CephFS, a reworkin the librados object listing API to properly handle namespaces, anda pile of bug fixes for RGW. There are also several smaller issuesfixed up in the performance area with buffer alignment and memorycopies, osd cache tiering agent, and various CephFS fixes.

Upgrading

• New ability to list all objects from all namespaces can fail orreturn incomplete results when not all OSDs have been upgraded.Features rados –all ls, rados cppool, rados export, radoscache-flush-evict-all and rados cache-try-flush-evict-all can alsofail or return incomplete results.

Notable Changes

• buffer: add list::get_contiguous (Sage Weil)
• buffer: avoid rebuild if buffer already contiguous (Jianpeng Ma)
• ceph-disk: improved systemd support (Owen Synge)
• ceph-disk: set guid if reusing journal partition (Dan van der Ster)
• ceph-fuse, libcephfs: allow xattr caps in inject_release_failure (#9800 John Spray)
• ceph-fuse, libcephfs: fix I_COMPLETE_ORDERED checks (#9894 Yan, Zheng)
• ceph-fuse: fix dentry invalidation on 3.18+ kernels (#9997 Yan, Zheng)
• crush: fix detach_bucket (#10095 Sage Weil)
• crush: fix several bugs in adjust_item_weight (Rongze Zhu)
• doc: add dumpling to firefly upgrade section (#7679 John Wilkins)
• doc: document erasure coded pool operations (#9970 Loic Dachary)
• doc: file system osd config settings (Kevin Dalley)
• doc: key/value store config reference (John Wilkins)
• doc: update openstack docs for Juno (Sebastien Han)
• fix cluster logging from non-mon daemons (Sage Weil)
• init-ceph: check for systemd-run before using it (Boris Ranto)
• librados: fix infinite loop with skipped map epochs (#9986 Ding Dinghua)
• librados: fix iterator operator= bugs (#10082 David Zafman, Yehuda Sadeh)
• librados: fix null deref when pool DNE (#9944 Sage Weil)
• librados: fix timer race from recent refactor (Sage Weil)
• libradosstriper: fix shutdown hang (Dongmao Zhang)
• librbd: don’t close a closed parent in failure path (#10030 Jason Dillaman)
• librbd: fix diff test (#10002 Josh Durgin)
• librbd: fix locking for readahead (#10045 Jason Dillaman)
• librbd: refactor unit tests to use fixtures (Jason Dillaman)
• many many coverity cleanups (Danny Al-Gaaf)
• mds: a whole bunch of initial scrub infrastructure (Greg Farnum)
• mds: fix compat_version for MClientSession (#9945 John Spray)
• mds: fix reply snapbl (Yan, Zheng)
• mon: allow adding tiers to fs pools (#10135 John Spray)
• mon: fix MDS health status from peons (#10151 John Spray)
• mon: fix caching for min_last_epoch_clean (#9987 Sage Weil)
• mon: fix error output for add_data_pool (#9852 Joao Eduardo Luis)
• mon: include entity name in audit log for forwarded requests (#9913 Joao Eduardo Luis)
• mon: paxos: allow reads while proposing (#9321 #9322 Joao Eduardo Luis)
• msgr: asyncmessenger: add kqueue support (#9926 Haomai Wang)
• osd, librados: revamp PG listing API to handle namespaces (#9031 #9262 #9438 David Zafman)
• osd, mon: send intiial pg create time from mon to osd (#9887 David Zafman)
• osd: allow whiteout deletion in cache pool (Sage Weil)
• osd: cache pool: ignore min flush age when cache is full (Xinze Chi)
• osd: erasure coding: allow bench.sh to test ISA backend (Yuan Zhou)
• osd: erasure-code: encoding regression tests, corpus (#9420 Loic Dachary)
• osd: fix journal shutdown race (Sage Weil)
• osd: fix object age eviction (Zhiqiang Wang)
• osd: fix object atime calculation (Xinze Chi)
• osd: fix past_interval display bug (#9752 Loic Dachary)
• osd: journal: fix alignment checks, avoid useless memmove (Jianpeng Ma)
• osd: journal: update committed_thru after replay (#6756 Samuel Just)
• osd: keyvaluestore_dev: optimization (Chendi Xue)
• osd: make misdirected op checks robust for EC pools (#9835 Sage Weil)
• osd: removed some dead code (Xinze Chi)
• qa: parallelize make check (Loic Dachary)
• qa: tolerate nearly-full disk for make check (Loic Dachary)
• rgw: create subuser if needed when creating user (#10103 Yehuda Sadeh)
• rgw: fix If-Modified-Since (VRan Liu)
• rgw: fix content-length update (#9576 Yehuda Sadeh)
• rgw: fix disabling of max_size quota (#9907 Dong Lei)
• rgw: fix incorrect len when len is 0 (#9877 Yehuda Sadeh)
• rgw: fix object copy content type (#9478 Yehuda Sadeh)
• rgw: fix user stags in get-user-info API (#9359 Ray Lv)
• rgw: remove swift user manifest (DLO) hash calculation (#9973 Yehuda Sadeh)
• rgw: return timestamp on GET/HEAD (#8911 Yehuda Sadeh)
• rgw: set ETag on object copy (#9479 Yehuda Sadeh)
• rgw: update bucket index on attr changes, for multi-site sync (#5595 Yehuda Sadeh)

v0.88

This is the first development release after Giant. The two mainfeatures merged this round are the new AsyncMessenger (an alternativeimplementation of the network layer) from Haomai Wang at UnitedStack,and support for POSIX file locks in ceph-fuse and libcephfs from Yan,Zheng. There is also a big pile of smaller items that re merged whilewe were stabilizing Giant, including a range of smaller performanceand bug fixes and some new tracepoints for LTTNG.

Notable Changes

• ceph-disk: Scientific Linux support (Dan van der Ster)
• ceph-disk: respect –statedir for keyring (Loic Dachary)
• ceph-fuse, libcephfs: POSIX file lock support (Yan, Zheng)
• ceph-fuse, libcephfs: fix cap flush overflow (Greg Farnum, Yan, Zheng)
• ceph-fuse, libcephfs: fix root inode xattrs (Yan, Zheng)
• ceph-fuse, libcephfs: preserve dir ordering (#9178 Yan, Zheng)
• ceph-fuse, libcephfs: trim inodes before reconnecting to MDS (Yan, Zheng)
• ceph: do not parse injectargs twice (Loic Dachary)
• ceph: make ‘ceph -s’ output more readable (Sage Weil)
• ceph: new ‘ceph tell mds.$name_or_rank_or_gid’ (John Spray)
• ceph: test robustness (Joao Eduardo Luis)
• ceph_objectstore_tool: behave with sharded flag (#9661 David Zafman)
• cephfs-journal-tool: fix journal import (#10025 John Spray)
• cephfs-journal-tool: skip up to expire_pos (#9977 John Spray)
• cleanup rados.h definitions with macros (Ilya Dryomov)
• common: shared_cache unit tests (Cheng Cheng)
• config: add $cctid meta variable (Adam Crume)
• crush: fix buffer overrun for poorly formed rules (#9492 Johnu George)
• crush: improve constness (Loic Dachary)
• crushtool: add –location command (Sage Weil, Loic Dachary)
• default to libnss instead of crypto++ (Federico Gimenez)
• doc: ceph osd reweight vs crush weight (Laurent Guerby)
• doc: document the LRC per-layer plugin configuration (Yuan Zhou)
• doc: erasure code doc updates (Loic Dachary)
• doc: misc updates (Alfredo Deza, VRan Liu)
• doc: preflight doc fixes (John Wilkins)
• doc: update PG count guide (Gerben Meijer, Laurent Guerby, Loic Dachary)
• keyvaluestore: misc fixes (Haomai Wang)
• keyvaluestore: performance improvements (Haomai Wang)
• librados: add rados_pool_get_base_tier() call (Adam Crume)
• librados: cap buffer length (Loic Dachary)
• librados: fix objecter races (#9617 Josh Durgin)
• libradosstriper: misc fixes (Sebastien Ponce)
• librbd: add missing python docstrings (Jason Dillaman)
• librbd: add readahead (Adam Crume)
• librbd: fix cache tiers in list_children and snap_unprotect (Adam Crume)
• librbd: fix performance regression in ObjectCacher (#9513 Adam Crume)
• librbd: lttng tracepoints (Adam Crume)
• librbd: misc fixes (Xinxin Shu, Jason Dillaman)
• mds: fix sessionmap lifecycle bugs (Yan, Zheng)
• mds: initialize root inode xattr version (Yan, Zheng)
• mds: introduce auth caps (John Spray)
• mds: misc bugs (Greg Farnum, John Spray, Yan, Zheng, Henry Change)
• misc coverity fixes (Danny Al-Gaaf)
• mon: add ‘ceph osd rename-bucket …’ command (Loic Dachary)
• mon: clean up auth list output (Loic Dachary)
• mon: fix ‘osd crush link’ id resolution (John Spray)
• mon: fix misc error paths (Joao Eduardo Luis)
• mon: fix paxos off-by-one corner case (#9301 Sage Weil)
• mon: new ‘ceph pool ls [detail]’ command (Sage Weil)
• mon: wait for writeable before cross-proposing (#9794 Joao Eduardo Luis)
• msgr: avoid useless new/delete (Haomai Wang)
• msgr: fix delay injection bug (#9910 Sage Weil, Greg Farnum)
• msgr: new AsymcMessenger alternative implementation (Haomai Wang)
• msgr: prefetch data when doing recv (Yehuda Sadeh)
• osd: add erasure code corpus (Loic Dachary)
• osd: add misc tests (Loic Dachary, Danny Al-Gaaf)
• osd: cleanup boost optionals (William Kennington)
• osd: expose non-journal backends via ceph-osd CLI (Hoamai Wang)
• osd: fix JSON output for stray OSDs (Loic Dachary)
• osd: fix ioprio options (Loic Dachary)
• osd: fix transaction accounting (Jianpeng Ma)
• osd: misc optimizations (Xinxin Shu, Zhiqiang Wang, Xinze Chi)
• osd: use FIEMAP_FLAGS_SYNC instead of fsync (Jianpeng Ma)
• rados: fix put of /dev/null (Loic Dachary)
• rados: parse command-line arguments more strictly (#8983 Adam Crume)
• rbd-fuse: fix memory leak (Adam Crume)
• rbd-replay-many (Adam Crume)
• rbd-replay: –anonymize flag to rbd-replay-prep (Adam Crume)
• rbd: fix ‘rbd diff’ for non-existent objects (Adam Crume)
• rbd: fix error when striping with format 1 (Sebastien Han)
• rbd: fix export for image sizes over 2GB (Vicente Cheng)
• rbd: use rolling average for rbd bench-write throughput (Jason Dillaman)
• rgw: send explicit HTTP status string (Yehuda Sadeh)
• rgw: set length for keystone token validation request (#7796 Yehuda Sadeh, Mark Kirkwood)
• udev: fix rules for CentOS7/RHEL7 (Loic Dachary)
• use clock_gettime instead of gettimeofday (Jianpeng Ma)
• vstart.sh: set up environment for s3-tests (Luis Pabon)

v0.87.2 Giant

This is the second (and possibly final) point release for Giant.

We recommend all v0.87.x Giant users upgrade to this release.

Notable Changes

• ceph-objectstore-tool: only output unsupported features when incompatible (#11176 David Zafman)
• common: do not implicitly unlock rwlock on destruction (Federico Simoncelli)
• common: make wait timeout on empty queue configurable (#10818 Samuel Just)
• crush: pick ruleset id that matches and rule id (Xiaoxi Chen)
• crush: set_choose_tries = 100 for new erasure code rulesets (#10353 Loic Dachary)
• librados: check initialized atomic safely (#9617 Josh Durgin)
• librados: fix failed tick_event assert (#11183 Zhiqiang Wang)
• librados: fix looping on skipped maps (#9986 Ding Dinghua)
• librados: fix op submit with timeout (#10340 Samuel Just)
• librados: pybind: fix memory leak (#10723 Billy Olsen)
• librados: pybind: keep reference to callbacks (#10775 Josh Durgin)
• librados: translate operation flags from C APIs (Matthew Richards)
• libradosstriper: fix write_full on ENOENT (#10758 Sebastien Ponce)
• libradosstriper: use strtoll instead of strtol (Dongmao Zhang)
• mds: fix assertion caused by system time moving backwards (#11053 Yan, Zheng)
• mon: allow injection of random delays on writes (Joao Eduardo Luis)
• mon: do not trust small osd epoch cache values (#10787 Sage Weil)
• mon: fail non-blocking flush if object is being scrubbed (#8011 Samuel Just)
• mon: fix division by zero in stats dump (Joao Eduardo Luis)
• mon: fix get_rule_avail when no osds (#10257 Joao Eduardo Luis)
• mon: fix timeout rounds period (#10546 Joao Eduardo Luis)
• mon: ignore osd failures before up_from (#10762 Dan van der Ster, Sage Weil)
• mon: paxos: reset accept timeout before writing to store (#10220 Joao Eduardo Luis)
• mon: return if fs exists on ‘fs new’ (Joao Eduardo Luis)
• mon: use EntityName when expanding profiles (#10844 Joao Eduardo Luis)
• mon: verify cross-service proposal preconditions (#10643 Joao Eduardo Luis)
• mon: wait for osdmon to be writeable when requesting proposal (#9794 Joao Eduardo Luis)
• mount.ceph: avoid spurious error message about /etc/mtab (#10351 Yan, Zheng)
• msg/simple: allow RESETSESSION when we forget an endpoint (#10080 Greg Farnum)
• msg/simple: discard delay queue before incoming queue (#9910 Sage Weil)
• osd: clear_primary_state when leaving Primary (#10059 Samuel Just)
• osd: do not ignore deleted pgs on startup (#10617 Sage Weil)
• osd: fix FileJournal wrap to get header out first (#10883 David Zafman)
• osd: fix PG leak in SnapTrimWQ (#10421 Kefu Chai)
• osd: fix journalq population in do_read_entry (#6003 Samuel Just)
• osd: fix operator== for op_queue_age_hit and fs_perf_stat (#10259 Samuel Just)
• osd: fix rare assert after split (#10430 David Zafman)
• osd: get pgid ancestor from last_map when building past intervals (#10430 David Zafman)
• osd: include rollback_info_trimmed_to in {read,write}_log (#10157 Samuel Just)
• osd: lock header_lock in DBObjectMap::sync (#9891 Samuel Just)
• osd: requeue blocked op before flush it was blocked on (#10512 Sage Weil)
• osd: tolerate missing object between list and attr get on backfill (#10150 Samuel Just)
• osd: use correct atime for eviction decision (Xinze Chi)
• rgw: flush XML header on get ACL request (#10106 Yehuda Sadeh)
• rgw: index swift keys appropriately (#10471 Hemant Bruman, Yehuda Sadeh)
• rgw: send cancel for bucket index pending ops (#10770 Baijiaruo, Yehuda Sadeh)
• rgw: swift: support X_Remove_Container-Meta-{key} (#01475 Dmytro Iurchenko)

For more detailed information, see thecompletechangelog.

v0.87.1 Giant

This is the first (and possibly final) point release for Giant. Our focuson stability fixes will be directed towards Hammer and Firefly.

We recommend that all v0.87 Giant users upgrade to this release.

Upgrading

• Due to a change in the Linux kernel version 3.18 and the limits of the FUSEinterface, ceph-fuse needs be mounted as root on at least some systems. Seeissues #9997, #10277, and #10542 for details.

Notable Changes

• build: disable stack-execute bit on assembler objects (#10114 Dan Mick)
• build: support boost 1.57.0 (#10688 Ken Dreyer)
• ceph-disk: fix dmcrypt file permissions (#9785 Loic Dachary)
• ceph-disk: run partprobe after zap, behave with partx or partprobe (#9665 #9721 Loic Dachary)
• cephfs-journal-tool: fix import for aged journals (#9977 John Spray)
• cephfs-journal-tool: fix journal import (#10025 John Spray)
• ceph-fuse: use remount to trim kernel dcache (#10277 Yan, Zheng)
• common: add cctid meta variable (#6228 Adam Crume)
• common: fix dump of shard for ghobject_t (#10063 Loic Dachary)
• crush: fix bucket weight underflow (#9998 Pawel Sadowski)
• erasure-code: enforce chunk size alignment (#10211 Loic Dachary)
• erasure-code: regression test suite (#9420 Loic Dachary)
• erasure-code: relax caucy w restrictions (#10325 Loic Dachary)
• libcephfs,ceph-fuse: allow xattr caps on inject_release_failure (#9800 John Spray)
• libcephfs,ceph-fuse: fix cap flush tid comparison (#9869 Greg Farnum)
• libcephfs,ceph-fuse: new flag to indicated sorted dcache (#9178 Yan, Zheng)
• libcephfs,ceph-fuse: prune cache before reconnecting to MDS (Yan, Zheng)
• librados: limit number of in-flight read requests (#9854 Jason Dillaman)
• libradospy: fix thread shutdown (#8797 Dan Mick)
• libradosstriper: fix locking issue in truncate (#10129 Sebastien Ponce)
• librbd: complete pending ops before closing mage (#10299 Jason Dillaman)
• librbd: fix error path on image open failure (#10030 Jason Dillaman)
• librbd: gracefully handle deleted/renamed pools (#10270 Jason Dillaman)
• librbd: handle errors when creating ioctx while listing children (#10123 Jason Dillaman)
• mds: fix compat version in MClientSession (#9945 John Spray)
• mds: fix journaler write error handling (#10011 John Spray)
• mds: fix locking for file size recovery (#10229 Yan, Zheng)
• mds: handle heartbeat_reset during shutdown (#10382 John Spray)
• mds: store backtrace for straydir (Yan, Zheng)
• mon: allow tiers for FS pools (#10135 John Spray)
• mon: fix caching of last_epoch_clean, osdmap trimming (#9987 Sage Weil)
• mon: fix ‘fs ls’ on peons (#10288 John Spray)
• mon: fix MDS health status from peons (#10151 John Spray)
• mon: fix paxos off-by-one (#9301 Sage Weil)
• msgr: simple: do not block on takeover while holding global lock (#9921 Greg Farnum)
• osd: deep scrub must not abort if hinfo is missing (#10018 Loic Dachary)
• osd: fix misdirected op detection (#9835 Sage Weil)
• osd: fix past_interval display for acting (#9752 Loic Dachary)
• osd: fix PG peering backoff when behind on osdmaps (#10431 Sage Weil)
• osd: handle no-op write with snapshot case (#10262 Ssage Weil)
• osd: use fast-dispatch (Sage Weil, Greg Farnum)
• rados: fix write to /dev/null (Loic Dachary)
• radosgw-admin: create subuser when needed (#10103 Yehuda Sadeh)
• rbd: avoid invalidating aio_write buffer during image import (#10590 Jason Dillaman)
• rbd: fix export with images > 2GB (Vicente Cheng)
• rgw: change multipart upload id magic (#10271 Georgios Dimitrakakis, Yehuda Sadeh)
• rgw: check keystone auth for S3 POST (#10062 Abhishek Lekshmanan)
• rgw: check timestamp for S3 keystone auth (#10062 Abhishek Lekshmanan)
• rgw: fix partial GET with swift (#10553 Yehuda Sadeh)
• rgw: fix quota disable (#9907 Dong Lei)
• rgw: fix rare corruption of object metadata on put (#9576 Yehuda Sadeh)
• rgw: fix S3 object copy content-type (#9478 Yehuda Sadeh)
• rgw: headers end with rn (#9254 Benedikt Fraunhofer, Yehuda Sadeh)
• rgw: remove swift user manifest DLO hash calculation (#9973 Yehuda Sadeh)
• rgw: return correct len when len is 0 (#9877 Yehuda Sadeh)
• rgw: return X-Timestamp field (#8911 Yehuda Sadeh)
• rgw: run radosgw as apache with systemd (#10125)
• rgw: sent ETag on S3 object copy (#9479 Yehuda Sadeh)
• rgw: sent HTTP status reason explicitly in fastcgi (Yehuda Sadeh)
• rgw: set length for keystone token validation (#7796 Mark Kirkwood, Yehuda Sadeh)
• rgw: set ulimit -n on sysvinit before starting daemon (#9587 Sage Weil)
• rgw: update bucket index on set_attrs (#5595 Yehuda Sadeh)
• rgw: update swift subuser permission masks when authenticating (#9918 Yehuda Sadeh)
• rgw: URL decode HTTP query params correction (#10271 Georgios Dimitrakakis, Yehuda Sadeh)
• rgw: use cached attrs while reading object attrs (#10307 Yehuda Sadeh)
• rgw: use strict_strtoll for content length (#10701 Axel Dunkel, Yehuda Sadeh)

For more detailed information, see thecompletechangelog.

v0.87 Giant

This release will form the basis for the stable release Giant,v0.87.x. Highlights for Giant include:

• RADOS Performance: a range of improvements have been made in theOSD and client-side librados code that improve the throughput onflash backends and improve parallelism and scaling on fast machines.
• CephFS: we have fixed a raft of bugs in CephFS and built somebasic journal recovery and diagnostic tools. Stability andperformance of single-MDS systems is vastly improved in Giant.Although we do not yet recommend CephFS for production deployments,we do encourage testing for non-critical workloads so that we canbetter guage the feature, usability, performance, and stabilitygaps.
• Local Recovery Codes: the OSDs now support an erasure-coding schemethat stores some additional data blocks to reduce the IO required torecover from single OSD failures.
• Degraded vs misplaced: the Ceph health reports from ‘ceph -s’ andrelated commands now make a distinction between data that isdegraded (there are fewer than the desired number of copies) anddata that is misplaced (stored in the wrong location in thecluster). The distinction is important because the latter does notcompromise data safety.
• Tiering improvements: we have made several improvements to thecache tiering implementation that improve performance. Mostnotably, objects are not promoted into the cache tier by a singleread; they must be found to be sufficiently hot before that happens.
• Monitor performance: the monitors now perform writes to the localdata store asynchronously, improving overall responsiveness.
• Recovery tools: the ceph_objectstore_tool is greatly expanded toallow manipulation of an individual OSDs data store for debuggingand repair purposes. This is most heavily used by our QAinfrastructure to exercise recovery code.

Upgrade Sequencing

• If your existing cluster is running a version older than v0.80.xFirefly, please first upgrade to the latest Firefly release beforemoving on to Giant. We have not tested upgrades directly fromEmperor, Dumpling, or older releases.

We have tested:

• Firefly to Giant
• Dumpling to Firefly to Giant

• Please upgrade daemons in the following order:

1. Monitors
2. OSDs
3. MDSs and/or radosgw

Note that the relative ordering of OSDs and monitors should not matter, butwe primarily tested upgrading monitors first.

Upgrading from v0.80x Firefly

• The client-side caching for librbd is now enabled by default (rbdcache = true). A safety option (rbd cache writethrough until flush= true) is also enabled so that writeback caching is not used untilthe library observes a ‘flush’ command, indicating that the librbdusers is passing that operation through from the guest VM. Thisavoids potential data loss when used with older versions of qemuthat do not support flush.

leveldb_write_buffer_size = 810241024 = 33554432 // 8MBleveldb_cache_size = 51210241204 = 536870912 // 512MBleveldb_block_size = 64*1024 = 65536 // 64KBleveldb_compression = falseleveldb_log = “”

OSDs will still maintain the following osd-specific defaults:

leveldb_log = “”

• The ‘rados getxattr …’ command used to add a gratuitous newline to the attrvalue; it now does not.

• The _kbperf counters on the monitor have been removed. These arereplaced with a new set of _bytes counters (e.g., cluster_osd_kb isreplaced by cluster_osd_bytes).

• The rd_kb and wr_kb fields in the JSON dumps for pool stats (accessedvia the cephdfdetail-fjson-pretty and related commands) have beenreplaced with corresponding *_bytes fields. Similarly, thetotal_space, total_used, and total_avail fields are replaced withtotal_bytes, total_used_bytes, and total_avail_bytes fields.

• The radosdf—format=json output read_bytes and write_bytesfields were incorrectly reporting ops; this is now fixed.

• The radosdf—format=json output previously included read_kb andwrite_kb fields; these have been removed. Please use read_bytes andwrite_bytes instead (and divide by 1024 if appropriate).

• The experimental keyvaluestore-dev OSD backend had an on-disk formatchange that prevents existing OSD data from being upgraded. Thisaffects developers and testers only.

• mon-specific and osd-specific leveldb options have been removed.From this point onward users should use the leveldb*_ genericoptions and add the options in the appropriate sections of theirconfiguration files. Monitors will still maintain the followingmonitor-specific defaults:

leveldb_write_buffer_size = 810241024 = 33554432 // 8MBleveldb_cache_size = 51210241204 = 536870912 // 512MBleveldb_block_size = 64*1024 = 65536 // 64KBleveldb_compression = falseleveldb_log = “”

OSDs will still maintain the following osd-specific defaults:

leveldb_log = “”

• CephFS support for the legacy anchor table has finally been removed.Users with file systems created before firefly should ensure that inodeswith multiple hard links are modified prior to the upgrade to ensure thatthe backtraces are written properly. For example:

sudo find /mnt/cephfs -type f -links +1 -exec touch \{\} \;

• We disallow nonsensical ‘tier cache-mode’ transitions. From this pointonward, ‘writeback’ can only transition to ‘forward’ and ‘forward’can transition to 1) ‘writeback’ if there are dirty objects, or 2) any ifthere are no dirty objects.

Notable Changes since v0.86

• ceph-disk: use new udev rules for centos7/rhel7 (#9747 Loic Dachary)
• libcephfs-java: fix fstat mode (Noah Watkins)
• librados: fix deadlock when listing PG contents (Guang Yang)
• librados: misc fixes to the new threading model (#9582 #9706 #9845 #9873 Sage Weil)
• mds: fix inotable initialization (Henry C Chang)
• mds: gracefully handle unknown lock type in flock requests (Yan, Zheng)
• mon: add read-only, read-write, and role-definer rols (Joao Eduardo Luis)
• mon: fix mon cap checks (Joao Eduardo Luis)
• mon: misc fixes for new paxos async writes (#9635 Sage Weil)
• mon: set scrub timestamps on PG creation (#9496 Joao Eduardo Luis)
• osd: erasure code: fix buffer alignment (Janne Grunau, Loic Dachary)
• osd: fix alloc hint induced crashes on mixed clusters (#9419 David Zafman)
• osd: fix backfill reservation release on rejection (#9626, Samuel Just)
• osd: fix ioprio option parsing (#9676 #9677 Loic Dachary)
• osd: fix memory leak during snap trimming (#9113 Samuel Just)
• osd: misc peering and recovery fixes (#9614 #9696 #9731 #9718 #9821 #9875 Samuel Just, Guang Yang)

Notable Changes since v0.80.x Firefly

• bash completion improvements (Wido den Hollander)
• brag: fixes, improvements (Loic Dachary)
• buffer: improve rebuild_page_aligned (Ma Jianpeng)
• build: fix build on alpha (Michael Cree, Dmitry Smirnov)
• build: fix CentOS 5 (Gerben Meijer)
• build: fix yasm check for x32 (Daniel Schepler, Sage Weil)
• ceph-brag: add tox tests (Alfredo Deza)
• ceph-conf: flush log on exit (Sage Weil)
• ceph.conf: update sample (Sebastien Han)
• ceph-dencoder: refactor build a bit to limit dependencies (Sage Weil, Dan Mick)
• ceph-disk: add Scientific Linux support (Dan van der Ster)
• ceph-disk: do not inadvertantly create directories (Owne Synge)
• ceph-disk: fix dmcrypt support (Sage Weil)
• ceph-disk: fix dmcrypt support (Stephen Taylor)
• ceph-disk: handle corrupt volumes (Stuart Longlang)
• ceph-disk: linter cleanup, logging improvements (Alfredo Deza)
• ceph-disk: partprobe as needed (Eric Eastman)
• ceph-disk: show information about dmcrypt in ‘ceph-disk list’ output (Sage Weil)
• ceph-disk: use partition type UUIDs and blkid (Sage Weil)
• ceph: fix for non-default cluster names (#8944, Dan Mick)
• ceph-fuse, libcephfs: asok hooks for handling session resets, timeouts (Yan, Zheng)
• ceph-fuse, libcephfs: fix crash in trim_caps (John Spray)
• ceph-fuse, libcephfs: improve cap trimming (John Spray)
• ceph-fuse, libcephfs: improve traceless reply handling (Sage Weil)
• ceph-fuse, libcephfs: virtual xattrs for rstat (Yan, Zheng)
• ceph_objectstore_tool: vastly improved and extended tool for working offline with OSD data stores (David Zafman)
• ceph.spec: many fixes (Erik Logtenberg, Boris Ranto, Dan Mick, Sandon Van Ness)
• ceph.spec: split out ceph-common package, other fixes (Sandon Van Ness)
• ceph_test_librbd_fsx: fix RNG, make deterministic (Ilya Dryomov)
• cephtool: fix help (Yilong Zhao)
• cephtool: refactor and improve CLI tests (Joao Eduardo Luis)
• cephtool: test cleanup (Joao Eduardo Luis)
• clang build fixes (John Spray, Danny Al-Gaaf)
• client: improved MDS session dumps (John Spray)
• common: add config diff admin socket command (Joao Eduardo Luis)
• common: add rwlock assertion checks (Yehuda Sadeh)
• common: fix dup log messages (#9080, Sage Weil)
• common: perfcounters now use atomics and go faster (Sage Weil)
• config: support G, M, K, etc. suffixes (Joao Eduardo Luis)
• coverity cleanups (Danny Al-Gaaf)
• crush: clean up CrushWrapper interface (Xioaxi Chen)
• crush: include new tunables in dump (Sage Weil)
• crush: make ruleset ids unique (Xiaoxi Chen, Loic Dachary)
• crush: only require rule features if the rule is used (#8963, Sage Weil)
• crushtool: send output to stdout, not stderr (Wido den Hollander)
• doc: cache tiering (John Wilkins)
• doc: CRUSH updates (John Wilkins)
• doc: document new upstream wireshark dissector (Kevin Cox)
• doc: improve manual install docs (Francois Lafont)
• doc: keystone integration docs (John Wilkins)
• doc: librados example fixes (Kevin Dalley)
• doc: many doc updates (John Wilkins)
• doc: many install doc updates (John Wilkins)
• doc: misc updates (John Wilkins, Loic Dachary, David Moreau Simard, Wido den Hollander. Volker Voigt, Alfredo Deza, Stephen Jahl, Dan van der Ster)
• doc: osd primary affinity (John Wilkins)
• doc: pool quotas (John Wilkins)
• doc: pre-flight doc improvements (Kevin Dalley)
• doc: switch to an unencumbered font (Ross Turk)
• doc: updated simple configuration guides (John Wilkins)
• doc: update erasure docs (Loic Dachary, Venky Shankar)
• doc: update openstack docs (Josh Durgin)
• filestore: disable use of XFS hint (buggy on old kernels) (Samuel Just)
• filestore: fix xattr spillout (Greg Farnum, Haomai Wang)
• fix hppa arch build (Dmitry Smirnov)
• fix i386 builds (Sage Weil)
• fix struct vs class inconsistencies (Thorsten Behrens)
• global: write pid file even when running in foreground (Alexandre Oliva)
• hadoop: improve tests (Huamin Chen, Greg Farnum, John Spray)
• hadoop: update hadoop tests for Hadoop 2.0 (Haumin Chen)
• init-ceph: continue starting other daemons on crush or mount failure (#8343, Sage Weil)
• journaler: fix locking (Zheng, Yan)
• keyvaluestore: fix hint crash (#8381, Haomai Wang)
• keyvaluestore: header cache (Haomai Wang)
• libcephfs-java: build against older JNI headers (Greg Farnum)
• libcephfs-java: fix gcj-jdk build (Dmitry Smirnov)
• librados: fix crash on read op timeout (#9362 Matthias Kiefer, Sage Weil)
• librados: fix lock leaks in error paths (#9022, Paval Rallabhandi)
• librados: fix pool existence check (#8835, Pavan Rallabhandi)
• librados: fix rados_pool_list bounds checks (Sage Weil)
• librados: fix shutdown race (#9130 Sage Weil)
• librados: fix watch/notify test (#7934 David Zafman)
• librados: fix watch reregistration on acting set change (#9220 Samuel Just)
• librados: give Objecter fine-grained locks (Yehuda Sadeh, Sage Weil, John Spray)
• librados: lttng tracepoitns (Adam Crume)
• librados, osd: return ETIMEDOUT on failed notify (Sage Weil)
• librados: pybind: fix reads when 0 is present (#9547 Mohammad Salehe)
• librados_striper: striping library for librados (Sebastien Ponce)
• librbd, ceph-fuse: reduce cache flush overhead (Haomai Wang)
• librbd: check error code on cache invalidate (Josh Durgin)
• librbd: enable caching by default (Sage Weil)
• librbd: enforce cache size on read requests (Jason Dillaman)
• librbd: fix crash using clone of flattened image (#8845, Josh Durgin)
• librbd: fix error path when opening image (#8912, Josh Durgin)
• librbd: handle blacklisting during shutdown (#9105 John Spray)
• librbd: lttng tracepoints (Adam Crume)
• librbd: new libkrbd library for kernel map/unmap/showmapped (Ilya Dryomov)
• librbd: store and retrieve snapshot metadata based on id (Josh Durgin)
• libs3: update to latest (Danny Al-Gaaf)
• log: fix derr level (Joao Eduardo Luis)
• logrotate: fix osd log rotation on ubuntu (Sage Weil)
• lttng: tracing infrastructure (Noah Watkins, Adam Crume)
• mailmap: many updates (Loic Dachary)
• mailmap: updates (Loic Dachary, Abhishek Lekshmanan, M Ranga Swami Reddy)
• Makefile: fix out of source builds (Stefan Eilemann)
• many many coverity fixes, cleanups (Danny Al-Gaaf)
• mds: adapt to new Objecter locking, give types to all Contexts (John Spray)
• mds: add file system name, enabled flag (John Spray)
• mds: add internal health checks (John Spray)
• mds: add min/max UID for snapshot creation/deletion (#9029, Wido den Hollander)
• mds: avoid tight mon reconnect loop (#9428 Sage Weil)
• mds: boot refactor, cleanup (John Spray)
• mds: cephfs-journal-tool (John Spray)
• mds: fix crash killing sessions (#9173 John Spray)
• mds: fix ctime updates (#9514 Greg Farnum)
• mds: fix journal conversion with standby-replay (John Spray)
• mds: fix replay locking (Yan, Zheng)
• mds: fix standby-replay cache trimming (#8648 Zheng, Yan)
• mds: fix xattr bug triggered by ACLs (Yan, Zheng)
• mds: give perfcounters meaningful names (Sage Weil)
• mds: improve health reporting to monitor (John Spray)
• mds: improve Journaler on-disk format (John Spray)
• mds: improve journal locking (Zheng, Yan)
• mds, libcephfs: use client timestamp for mtime/ctime (Sage Weil)
• mds: make max file recoveries tunable (Sage Weil)
• mds: misc encoding improvements (John Spray)
• mds: misc fixes for multi-mds (Yan, Zheng)
• mds: multi-mds fixes (Yan, Zheng)
• mds: OPTracker integration, dump_ops_in_flight (Greg Farnum)
• mds: prioritize file recovery when appropriate (Sage Weil)
• mds: refactor beacon, improve reliability (John Spray)
• mds: remove legacy anchor table (Yan, Zheng)
• mds: remove legacy discover ino (Yan, Zheng)
• mds: restart on EBLACKLISTED (John Spray)
• mds: separate inode recovery queue (John Spray)
• mds: session ls, evict commands (John Spray)
• mds: submit log events in async thread (Yan, Zheng)
• mds: track RECALL progress, report failure (#9284 John Spray)
• mds: update segment references during journal write (John Spray, Greg Farnum)
• mds: use client-provided timestamp for user-visible file metadata (Yan, Zheng)
• mds: use meaningful names for clients (John Spray)
• mds: validate journal header on load and save (John Spray)
• mds: warn clients which aren’t revoking caps (Zheng, Yan, John Spray)
• misc build errors/warnings for Fedora 20 (Boris Ranto)
• misc build fixes for OS X (John Spray)
• misc cleanup (Christophe Courtaut)
• misc integer size cleanups (Kevin Cox)
• misc memory leaks, cleanups, fixes (Danny Al-Gaaf, Sahid Ferdjaoui)
• misc suse fixes (Danny Al-Gaaf)
• misc word size fixes (Kevin Cox)
• mon: add audit log for all admin commands (Joao Eduardo Luis)
• mon: add cluster fingerprint (Sage Weil)
• mon: add get-quota commands (Joao Eduardo Luis)
• mon: add ‘osd blocked-by’ command to easily see which OSDs are blocking peering progress (Sage Weil)
• mon: add ‘osd reweight-by-pg’ command (Sage Weil, Guang Yang)
• mon: add perfcounters for paxos operations (Sage Weil)
• mon: avoid creating unnecessary rule on pool create (#9304 Loic Dachary)
• monclient: fix hang (Sage Weil)
• mon: create default EC profile if needed (Loic Dachary)
• mon: do not create file system by default (John Spray)
• mon: do not spam log (Aanchal Agrawal, Sage Weil)
• mon: drop mon- and osd- specific leveldb options (Joao Eduardo Luis)
• mon: ec pool profile fixes (Loic Dachary)
• mon: fix bug when no auth keys are present (#8851, Joao Eduardo Luis)
• mon: fix ‘ceph df’ output for available space (Xiaoxi Chen)
• mon: fix compat version for MForward (Joao Eduardo Luis)
• mon: fix crash on loopback messages and paxos timeouts (#9062, Sage Weil)
• mon: fix default replication pool ruleset choice (#8373, John Spray)
• mon: fix divide by zero when pg_num is adjusted before OSDs are added (#9101, Sage Weil)
• mon: fix double-free of old MOSDBoot (Sage Weil)
• mon: fix health down messages (Sage Weil)
• mon: fix occasional memory leak after session reset (#9176, Sage Weil)
• mon: fix op write latency perfcounter (#9217 Xinxin Shu)
• mon: fix ‘osd perf’ reported latency (#9269 Samuel Just)
• mon: fix quorum feature check (#8738, Greg Farnum)
• mon: fix ruleset/ruleid bugs (#9044, Loic Dachary)
• mon: fix set cache_target_full_ratio (#8440, Geoffrey Hartz)
• mon: fix store check on startup (Joao Eduardo Luis)
• mon: include per-pool ‘max avail’ in df output (Sage Weil)
• mon: make paxos transaction commits asynchronous (Sage Weil)
• mon: make usage dumps in terms of bytes, not kB (Sage Weil)
• mon: ‘osd crush reweight-subtree …’ (Sage Weil)
• mon, osd: relax client EC support requirements (Sage Weil)
• mon: preload erasure plugins (#9153 Loic Dachary)
• mon: prevent cache pools from being used directly by CephFS (#9435 John Spray)
• mon: prevent EC pools from being used with cephfs (Joao Eduardo Luis)
• mon: prevent implicit destruction of OSDs with ‘osd setmaxosd …’ (#8865, Anand Bhat)
• mon: prevent nonsensical cache-mode transitions (Joao Eduardo Luis)
• mon: restore original weight when auto-marked out OSDs restart (Sage Weil)
• mon: restrict some pool properties to tiered pools (Joao Eduardo Luis)
• mon: some instrumentation (Sage Weil)
• mon: use msg header tid for MMonGetVersionReply (Ilya Dryomov)
• mon: use user-provided ruleset for replicated pool (Xiaoxi Chen)
• mon: verify all quorum members are contiguous at end of Paxos round (#9053, Sage Weil)
• mon: verify available disk space on startup (#9502 Joao Eduardo Luis)
• mon: verify erasure plugin version on load (Loic Dachary)
• msgr: avoid big lock when sending (most) messages (Greg Farnum)
• msgr: fix logged address (Yongyue Sun)
• msgr: misc locking fixes for fast dispatch (#8891, Sage Weil)
• msgr: refactor to cleanly separate SimpleMessenger implemenetation, move toward Connection-based calls (Matt Benjamin, Sage Wei)
• objecter: flag operations that are redirected by caching (Sage Weil)
• objectstore: clean up KeyValueDB interface for key/value backends (Sage Weil)
• osd: account for hit_set_archive bytes (Sage Weil)
• osd: add ability to prehash filestore directories (Guang Yang)
• osd: add ‘dump_reservations’ admin socket command (Sage Weil)
• osd: add feature bit for erasure plugins (Loic Dachary)
• osd: add header cache for KeyValueStore (Haomai Wang)
• osd: add ISA erasure plugin table cache (Andreas-Joachim Peters)
• osd: add local_mtime for use by cache agent (Zhiqiang Wang)
• osd: add local recovery code (LRC) erasure plugin (Loic Dachary)
• osd: add prototype KineticStore based on Seagate Kinetic (Josh Durgin)
• osd: add READFORWARD caching mode (Luis Pabon)
• osd: add superblock for KeyValueStore backend (Haomai Wang)
• osd: add support for Intel ISA-L erasure code library (Andreas-Joachim Peters)
• osd: allow map cache size to be adjusted at runtime (Sage Weil)
• osd: avoid refcounting overhead by passing a few things by ref (Somnath Roy)
• osd: avoid sharing PG info that is not durable (Samuel Just)
• osd: bound osdmap epoch skew between PGs (Sage Weil)
• osd: cache tier flushing fixes for snapped objects (Samuel Just)
• osd: cap hit_set size (#9339 Samuel Just)
• osd: clean up shard_id_t, shard_t (Loic Dachary)
• osd: clear FDCache on unlink (#8914 Loic Dachary)
• osd: clear slow request latency info on osd up/down (Sage Weil)
• osd: do not evict blocked objects (#9285 Zhiqiang Wang)
• osd: do not skip promote for write-ordered reads (#9064, Samuel Just)
• osd: fix agent early finish looping (David Zafman)
• osd: fix ambigous encoding order for blacklisted clients (#9211, Sage Weil)
• osd: fix bogus assert during OSD shutdown (Sage Weil)
• osd: fix bug with long object names and rename (#8701, Sage Weil)
• osd: fix cache flush corner case for snapshotted objects (#9054, Samuel Just)
• osd: fix cache full -> not full requeueing (#8931, Sage Weil)
• osd: fix clone deletion case (#8334, Sam Just)
• osd: fix clone vs cache_evict bug (#8629 Sage Weil)
• osd: fix connection reconnect race (Greg Farnum)
• osd: fix crash from duplicate backfill reservation (#8863 Sage Weil)
• osd: fix dead peer connection checks (#9295 Greg Farnum, Sage Weil)
• osd: fix discard of old/obsolete subop replies (#9259, Samuel Just)
• osd: fix discard of peer messages from previous intervals (Greg Farnum)
• osd: fix dump of open fds on EMFILE (Sage Weil)
• osd: fix dumps (Joao Eduardo Luis)
• osd: fix erasure-code lib initialization (Loic Dachary)
• osd: fix extent normalization (Adam Crume)
• osd: fix filestore removal corner case (#8332, Sam Just)
• osd: fix flush vs OpContext (Samuel Just)
• osd: fix gating of messages from old OSD instances (Greg Farnum)
• osd: fix hang waiting for osdmap (#8338, Greg Farnum)
• osd: fix interval check corner case during peering (#8104, Sam Just)
• osd: fix ISA erasure alignment (Loic Dachary, Andreas-Joachim Peters)
• osd: fix journal dump (Ma Jianpeng)
• osd: fix journal-less operation (Sage Weil)
• osd: fix keyvaluestore scrub (#8589 Haomai Wang)
• osd: fix keyvaluestore upgrade (Haomai Wang)
• osd: fix loopback msgr issue (Ma Jianpeng)
• osd: fix LSB release parsing (Danny Al-Gaaf)
• osd: fix MarkMeDown and other shutdown races (Sage Weil)
• osd: fix memstore bugs with collection_move_rename, lock ordering (Sage Weil)
• osd: fix min_read_recency_for_promote default on upgrade (Zhiqiang Wang)
• osd: fix mon feature bit requirements bug and resulting log spam (Sage Weil)
• osd: fix mount/remount sync race (#9144 Sage Weil)
• osd: fix PG object listing/ordering bug (Guang Yang)
• osd: fix PG stat errors with tiering (#9082, Sage Weil)
• osd: fix purged_snap initialization on backfill (Sage Weil, Samuel Just, Dan van der Ster, Florian Haas)
• osd: fix race condition on object deletion (#9480 Somnath Roy)
• osd: fix recovery chunk size usage during EC recovery (Ma Jianpeng)
• osd: fix recovery reservation deadlock for EC pools (Samuel Just)
• osd: fix removal of old xattrs when overwriting chained xattrs (Ma Jianpeng)
• osd: fix requesting queueing on PG split (Samuel Just)
• osd: fix scrub vs cache bugs (Samuel Just)
• osd: fix snap object writeback from cache tier (#9054 Samuel Just)
• osd: fix trim of hitsets (Sage Weil)
• osd: force new xattrs into leveldb if fs returns E2BIG (#7779, Sage Weil)
• osd: implement alignment on chunk sizes (Loic Dachary)
• osd: improved backfill priorities (Sage Weil)
• osd: improve journal shutdown (Ma Jianpeng, Mark Kirkwood)
• osd: improve locking for KeyValueStore (Haomai Wang)
• osd: improve locking in OpTracker (Pavan Rallabhandi, Somnath Roy)
• osd: improve prioritization of recovery of degraded over misplaced objects (Sage Weil)
• osd: improve tiering agent arithmetic (Zhiqiang Wang, Sage Weil, Samuel Just)
• osd: include backend information in metadata reported to mon (Sage Weil)
• osd: locking, sharding, caching improvements in FileStore’s FDCache (Somnath Roy, Greg Farnum)
• osd: lttng tracepoints for filestore (Noah Watkins)
• osd: make blacklist encoding deterministic (#9211 Sage Weil)
• osd: make tiering behave if hit_sets aren’t enabled (Sage Weil)
• osd: many important bug fixes (Samuel Just)
• osd: many many core fixes (Samuel Just)
• osd: many many important fixes (#8231 #8315 #9113 #9179 #9293 #9294 #9326 #9453 #9481 #9482 #9497 #9574 Samuel Just)
• osd: mark pools with incomplete clones (Sage Weil)
• osd: misc erasure code plugin fixes (Loic Dachary)
• osd: misc locking fixes for fast dispatch (Samuel Just, Ma Jianpeng)
• osd, mon: add rocksdb support (Xinxin Shu, Sage Weil)
• osd, mon: config sanity checks on start (Sage Weil, Joao Eduardo Luis)
• osd, mon: distinguish between “misplaced” and “degraded” objects in cluster health and PG state reporting (Sage Weil)
• osd, msgr: fast-dispatch of OSD ops (Greg Farnum, Samuel Just)
• osd, objecter: resend ops on last_force_op_resend barrier; fix cache overlay op ordering (Sage Weil)
• osd: preload erasure plugins (#9153 Loic Dachary)
• osd: prevent old rados clients from using tiered pools (#8714, Sage Weil)
• osd: reduce OpTracker overhead (Somnath Roy)
• osd: refactor some ErasureCode functionality into command parent class (Loic Dachary)
• osd: remove obsolete classic scrub code (David Zafman)
• osd: scrub PGs with invalid stats (Sage Weil)
• osd: set configurable hard limits on object and xattr names (Sage Weil, Haomai Wang)
• osd: set rollback_info_completed on create (#8625, Samuel Just)
• osd: sharded threadpool to improve parallelism (Somnath Roy)
• osd: shard OpTracker to improve performance (Somnath Roy)
• osd: simple io prioritization for scrub (Sage Weil)
• osd: simple scrub throttling (Sage Weil)
• osd: simple snap trimmer throttle (Sage Weil)
• osd: tests for bench command (Loic Dachary)
• osd: trim old EC objects quickly; verify on scrub (Samuel Just)
• osd: use FIEMAP to inform copy_range (Haomai Wang)
• osd: use local time for tiering decisions (Zhiqiang Wang)
• osd: use xfs hint less frequently (Ilya Dryomov)
• osd: verify erasure plugin version on load (Loic Dachary)
• osd: work around GCC 4.8 bug in journal code (Matt Benjamin)
• pybind/rados: fix small timeouts (John Spray)
• qa: xfstests updates (Ilya Dryomov)
• rados: allow setxattr value to be read from stdin (Sage Weil)
• rados bench: fix arg order (Kevin Dalley)
• rados: drop gratuitous n from getxattr command (Sage Weil)
• rados: fix bench write arithmetic (Jiangheng)
• rados: fix {read,write}_ops values for df output (Sage Weil)
• rbd: add rbdmap pre- and post post- hooks, fix misc bugs (Dmitry Smirnov)
• rbd-fuse: allow exposing single image (Stephen Taylor)
• rbd-fuse: fix unlink (Josh Durgin)
• rbd: improve option default behavior (Josh Durgin)
• rbd: parallelize rbd import, export (Jason Dillaman)
• rbd: rbd-replay utility to replay captured rbd workload traces (Adam Crume)
• rbd: use write-back (not write-through) when caching is enabled (Jason Dillaman)
• removed mkcephfs (deprecated since dumpling)
• rest-api: fix help (Ailing Zhang)
• rgw: add civetweb as default frontent on port 7490 (#9013 Yehuda Sadeh)
• rgw: add –min-rewrite-stripe-size for object restriper (Yehuda Sadeh)
• rgw: add powerdns hook for dynamic DNS for global clusters (Wido den Hollander)
• rgw: add S3 bucket get location operation (Abhishek Lekshmanan)
• rgw: allow : in S3 access key (Roman Haritonov)
• rgw: automatically align writes to EC pool (#8442, Yehuda Sadeh)
• rgw: bucket link uses instance id (Yehuda Sadeh)
• rgw: cache bucket info (Yehuda Sadeh)
• rgw: cache decoded user info (Yehuda Sadeh)
• rgw: check entity permission for put_metadata (#8428, Yehuda Sadeh)
• rgw: copy object data is target bucket is in a different pool (#9039, Yehuda Sadeh)
• rgw: do not try to authenticate CORS preflight requests (#8718, Robert Hubbard, Yehuda Sadeh)
• rgw: fix admin create user op (#8583 Ray Lv)
• rgw: fix civetweb URL decoding (#8621, Yehuda Sadeh)
• rgw: fix crash on swift CORS preflight request (#8586, Yehuda Sadeh)
• rgw: fix log filename suffix (#9353 Alexandre Marangone)
• rgw: fix memory leak following chunk read error (Yehuda Sadeh)
• rgw: fix memory leaks (Andrey Kuznetsov)
• rgw: fix multipart object attr regression (#8452, Yehuda Sadeh)
• rgw: fix multipart upload (#8846, Silvain Munaut, Yehuda Sadeh)
• rgw: fix radosgw-admin ‘show log’ command (#8553, Yehuda Sadeh)
• rgw: fix removal of objects during object creation (Patrycja Szablowska, Yehuda Sadeh)
• rgw: fix striping for copied objects (#9089, Yehuda Sadeh)
• rgw: fix test for identify whether an object has a tail (#9226, Yehuda Sadeh)
• rgw: fix URL decoding (#8702, Brian Rak)
• rgw: fix URL escaping (Yehuda Sadeh)
• rgw: fix usage (Abhishek Lekshmanan)
• rgw: fix user manifest (Yehuda Sadeh)
• rgw: fix when stripe size is not a multiple of chunk size (#8937, Yehuda Sadeh)
• rgw: handle empty extra pool name (Yehuda Sadeh)
• rgw: improve civetweb logging (Yehuda Sadeh)
• rgw: improve delimited listing of bucket, misc fixes (Yehuda Sadeh)
• rgw: improve -h (Abhishek Lekshmanan)
• rgw: many fixes for civetweb (Yehuda Sadeh)
• rgw: misc civetweb fixes (Yehuda Sadeh)
• rgw: misc civetweb frontend fixes (Yehuda Sadeh)
• rgw: object and bucket rewrite functions to allow restriping old objects (Yehuda Sadeh)
• rgw: powerdns backend for global namespaces (Wido den Hollander)
• rgw: prevent multiobject PUT race (Yehuda Sadeh)
• rgw: send user manifest header (Yehuda Sadeh)
• rgw: subuser creation fixes (#8587 Yehuda Sadeh)
• rgw: use systemd-run from sysvinit script (JuanJose Galvez)
• rpm: do not restart daemons on upgrade (Alfredo Deza)
• rpm: misc packaging fixes for rhel7 (Sandon Van Ness)
• rpm: split ceph-common from ceph (Sandon Van Ness)
• systemd: initial systemd config files (Federico Simoncelli)
• systemd: wrap started daemons in new systemd environment (Sage Weil, Dan Mick)
• sysvinit: add support for non-default cluster names (Alfredo Deza)
• sysvinit: less sensitive to failures (Sage Weil)
• test_librbd_fsx: test krbd as well as librbd (Ilya Dryomov)
• unit test improvements (Loic Dachary)
• upstart: increase max open files limit (Sage Weil)
• vstart.sh: fix/improve rgw support (Luis Pabon, Abhishek Lekshmanan)

v0.86

This is a release candidate for Giant, which will hopefully be outin another week or two. We did a feature freeze about a month agoand since then have been doing only stabilization and bug fixing (anda handful on low-risk enhancements). A fair bit of new functionalitywent into the final sprint, but it’s baked for quite a while now andwe’re feeling pretty good about it.

Major items include:

• librados locking refactor to improve scaling and client performance
• local recovery code (LRC) erasure code plugin to trade someadditional storage overhead for improved recovery performance
• LTTNG tracing framework, with initial tracepoints in librados,librbd, and the OSD FileStore backend
• separate monitor audit log for all administrative commands
• asynchronos monitor transaction commits to reduce the impact onmonitor read requests while processing updates
• low-level tool for working with individual OSD data stores fordebugging, recovery, and testing
• many MDS improvements (bug fixes, health reporting)

There are still a handful of known bugs in this release, but nothingsevere enough to prevent a release. By and large we are prettypleased with the stability and expect the final Giant release to bequite reliable.

Please try this out on your non-production clusters for a preview

Notable Changes

• buffer: improve rebuild_page_aligned (Ma Jianpeng)
• build: fix CentOS 5 (Gerben Meijer)
• build: fix build on alpha (Michael Cree, Dmitry Smirnov)
• build: fix yasm check for x32 (Daniel Schepler, Sage Weil)
• ceph-disk: add Scientific Linux support (Dan van der Ster)
• ceph-fuse, libcephfs: fix crash in trim_caps (John Spray)
• ceph-fuse, libcephfs: improve cap trimming (John Spray)
• ceph-fuse, libcephfs: virtual xattrs for rstat (Yan, Zheng)
• ceph.conf: update sample (Sebastien Han)
• ceph.spec: many fixes (Erik Logtenberg, Boris Ranto, Dan Mick, Sandon Van Ness)
• ceph_objectstore_tool: vastly improved and extended tool for working offline with OSD data stores (David Zafman)
• common: add config diff admin socket command (Joao Eduardo Luis)
• common: add rwlock assertion checks (Yehuda Sadeh)
• crush: clean up CrushWrapper interface (Xioaxi Chen)
• crush: make ruleset ids unique (Xiaoxi Chen, Loic Dachary)
• doc: improve manual install docs (Francois Lafont)
• doc: misc updates (John Wilkins, Loic Dachary, David Moreau Simard, Wido den Hollander. Volker Voigt, Alfredo Deza, Stephen Jahl, Dan van der Ster)
• global: write pid file even when running in foreground (Alexandre Oliva)
• hadoop: improve tests (Huamin Chen, Greg Farnum, John Spray)
• journaler: fix locking (Zheng, Yan)
• librados, osd: return ETIMEDOUT on failed notify (Sage Weil)
• librados: fix crash on read op timeout (#9362 Matthias Kiefer, Sage Weil)
• librados: fix shutdown race (#9130 Sage Weil)
• librados: fix watch reregistration on acting set change (#9220 Samuel Just)
• librados: fix watch/notify test (#7934 David Zafman)
• librados: give Objecter fine-grained locks (Yehuda Sadeh, Sage Weil, John Spray)
• librados: lttng tracepoitns (Adam Crume)
• librados: pybind: fix reads when 0 is present (#9547 Mohammad Salehe)
• librbd: enforce cache size on read requests (Jason Dillaman)
• librbd: handle blacklisting during shutdown (#9105 John Spray)
• librbd: lttng tracepoints (Adam Crume)
• lttng: tracing infrastructure (Noah Watkins, Adam Crume)
• mailmap: updates (Loic Dachary, Abhishek Lekshmanan, M Ranga Swami Reddy)
• many many coverity fixes, cleanups (Danny Al-Gaaf)
• mds: adapt to new Objecter locking, give types to all Contexts (John Spray)
• mds: add internal health checks (John Spray)
• mds: avoid tight mon reconnect loop (#9428 Sage Weil)
• mds: fix crash killing sessions (#9173 John Spray)
• mds: fix ctime updates (#9514 Greg Farnum)
• mds: fix replay locking (Yan, Zheng)
• mds: fix standby-replay cache trimming (#8648 Zheng, Yan)
• mds: give perfcounters meaningful names (Sage Weil)
• mds: improve health reporting to monitor (John Spray)
• mds: improve journal locking (Zheng, Yan)
• mds: make max file recoveries tunable (Sage Weil)
• mds: prioritize file recovery when appropriate (Sage Weil)
• mds: refactor beacon, improve reliability (John Spray)
• mds: restart on EBLACKLISTED (John Spray)
• mds: track RECALL progress, report failure (#9284 John Spray)
• mds: update segment references during journal write (John Spray, Greg Farnum)
• mds: use meaningful names for clients (John Spray)
• mds: warn clients which aren’t revoking caps (Zheng, Yan, John Spray)
• mon: add ‘osd reweight-by-pg’ command (Sage Weil, Guang Yang)
• mon: add audit log for all admin commands (Joao Eduardo Luis)
• mon: add cluster fingerprint (Sage Weil)
• mon: avoid creating unnecessary rule on pool create (#9304 Loic Dachary)
• mon: do not spam log (Aanchal Agrawal, Sage Weil)
• mon: fix ‘osd perf’ reported latency (#9269 Samuel Just)
• mon: fix double-free of old MOSDBoot (Sage Weil)
• mon: fix op write latency perfcounter (#9217 Xinxin Shu)
• mon: fix store check on startup (Joao Eduardo Luis)
• mon: make paxos transaction commits asynchronous (Sage Weil)
• mon: preload erasure plugins (#9153 Loic Dachary)
• mon: prevent cache pools from being used directly by CephFS (#9435 John Spray)
• mon: use user-provided ruleset for replicated pool (Xiaoxi Chen)
• mon: verify available disk space on startup (#9502 Joao Eduardo Luis)
• mon: verify erasure plugin version on load (Loic Dachary)
• msgr: fix logged address (Yongyue Sun)
• osd: account for hit_set_archive bytes (Sage Weil)
• osd: add ISA erasure plugin table cache (Andreas-Joachim Peters)
• osd: add ability to prehash filestore directories (Guang Yang)
• osd: add feature bit for erasure plugins (Loic Dachary)
• osd: add local recovery code (LRC) erasure plugin (Loic Dachary)
• osd: cap hit_set size (#9339 Samuel Just)
• osd: clear FDCache on unlink (#8914 Loic Dachary)
• osd: do not evict blocked objects (#9285 Zhiqiang Wang)
• osd: fix ISA erasure alignment (Loic Dachary, Andreas-Joachim Peters)
• osd: fix clone vs cache_evict bug (#8629 Sage Weil)
• osd: fix crash from duplicate backfill reservation (#8863 Sage Weil)
• osd: fix dead peer connection checks (#9295 Greg Farnum, Sage Weil)
• osd: fix keyvaluestore scrub (#8589 Haomai Wang)
• osd: fix keyvaluestore upgrade (Haomai Wang)
• osd: fix min_read_recency_for_promote default on upgrade (Zhiqiang Wang)
• osd: fix mount/remount sync race (#9144 Sage Weil)
• osd: fix purged_snap initialization on backfill (Sage Weil, Samuel Just, Dan van der Ster, Florian Haas)
• osd: fix race condition on object deletion (#9480 Somnath Roy)
• osd: fix snap object writeback from cache tier (#9054 Samuel Just)
• osd: improve journal shutdown (Ma Jianpeng, Mark Kirkwood)
• osd: improve locking in OpTracker (Pavan Rallabhandi, Somnath Roy)
• osd: improve tiering agent arithmetic (Zhiqiang Wang, Sage Weil, Samuel Just)
• osd: lttng tracepoints for filestore (Noah Watkins)
• osd: make blacklist encoding deterministic (#9211 Sage Weil)
• osd: many many important fixes (#8231 #8315 #9113 #9179 #9293 #9294 #9326 #9453 #9481 #9482 #9497 #9574 Samuel Just)
• osd: misc erasure code plugin fixes (Loic Dachary)
• osd: preload erasure plugins (#9153 Loic Dachary)
• osd: shard OpTracker to improve performance (Somnath Roy)
• osd: use local time for tiering decisions (Zhiqiang Wang)
• osd: verify erasure plugin version on load (Loic Dachary)
• rados: fix bench write arithmetic (Jiangheng)
• rbd: parallelize rbd import, export (Jason Dillaman)
• rbd: rbd-replay utility to replay captured rbd workload traces (Adam Crume)
• rbd: use write-back (not write-through) when caching is enabled (Jason Dillaman)
• rgw: add S3 bucket get location operation (Abhishek Lekshmanan)
• rgw: add civetweb as default frontent on port 7490 (#9013 Yehuda Sadeh)
• rgw: allow : in S3 access key (Roman Haritonov)
• rgw: fix admin create user op (#8583 Ray Lv)
• rgw: fix log filename suffix (#9353 Alexandre Marangone)
• rgw: fix usage (Abhishek Lekshmanan)
• rgw: many fixes for civetweb (Yehuda Sadeh)
• rgw: subuser creation fixes (#8587 Yehuda Sadeh)
• rgw: use systemd-run from sysvinit script (JuanJose Galvez)
• unit test improvements (Loic Dachary)
• vstart.sh: fix/improve rgw support (Luis Pabon, Abhishek Lekshmanan)

v0.85

This is the second-to-last development release before Giant thatcontains new functionality. The big items to land during this cycleare the messenger refactoring from Matt Benjmain that lays somegroundwork for RDMA support, a performance improvement series fromSanDisk that improves performance on SSDs, lots of improvements to ournew standalone civetweb-based RGW frontend, and a new ‘osd blocked-by’mon command that allows admins to easily identify which OSDs areblocking peering progress. The other big change is that the OSDs andMonitors now distinguish between “misplaced” and “degraded” objects:the latter means there are fewer copies than we’d like, while theformer simply means the are not stored in the locations where we wantthem to be.

Also of note is a change to librbd that enables client-side caching bydefault. This is coupled with another option that makes the cachewrite-through until a “flush” operations is observed: this impliesthat the librbd user (usually a VM guest OS) supports barriers andflush and that it is safe for the cache to switch into writeback modewithout compromising data safety or integrity. It has long beenrecommended practice that these options be enabled (e.g., in OpenStackenvironments) but until now it has not been the default.

We have frozen the tree for the looming Giant release, and the nextdevelopment release will be a release candidate with a final batch ofnew functionality.

Upgrading

• The client-side caching for librbd is now enabled by default (rbdcache = true). A safety option (rbd cache writethrough until flush= true) is also enabled so that writeback caching is not used untilthe library observes a ‘flush’ command, indicating that the librbdusers is passing that operation through from the guest VM. Thisavoids potential data loss when used with older versions of qemuthat do not support flush.

leveldb_write_buffer_size = 3210241024 = 33554432 // 32MBleveldb_cache_size = 51210241204 = 536870912 // 512MBleveldb_block_size = 64*1024 = 65536 // 64KBleveldb_compression = falseleveldb_log = “”

OSDs will still maintain the following osd-specific defaults:

leveldb_log = “”

• The ‘rados getxattr …’ command used to add a gratuitous newline to the attrvalue; it now does not.

Notable Changes

• ceph-disk: do not inadvertantly create directories (Owne Synge)
• ceph-disk: fix dmcrypt support (Sage Weil)
• ceph-disk: linter cleanup, logging improvements (Alfredo Deza)
• ceph-disk: show information about dmcrypt in ‘ceph-disk list’ output (Sage Weil)
• ceph-disk: use partition type UUIDs and blkid (Sage Weil)
• ceph: fix for non-default cluster names (#8944, Dan Mick)
• doc: document new upstream wireshark dissector (Kevin Cox)
• doc: many install doc updates (John Wilkins)
• librados: fix lock leaks in error paths (#9022, Paval Rallabhandi)
• librados: fix pool existence check (#8835, Pavan Rallabhandi)
• librbd: enable caching by default (Sage Weil)
• librbd: fix crash using clone of flattened image (#8845, Josh Durgin)
• librbd: store and retrieve snapshot metadata based on id (Josh Durgin)
• mailmap: many updates (Loic Dachary)
• mds: add min/max UID for snapshot creation/deletion (#9029, Wido den Hollander)
• misc build errors/warnings for Fedora 20 (Boris Ranto)
• mon: add ‘osd blocked-by’ command to easily see which OSDs are blocking peering progress (Sage Weil)
• mon: add perfcounters for paxos operations (Sage Weil)
• mon: create default EC profile if needed (Loic Dachary)
• mon: fix crash on loopback messages and paxos timeouts (#9062, Sage Weil)
• mon: fix divide by zero when pg_num is adjusted before OSDs are added (#9101, Sage Weil)
• mon: fix occasional memory leak after session reset (#9176, Sage Weil)
• mon: fix ruleset/ruleid bugs (#9044, Loic Dachary)
• mon: make usage dumps in terms of bytes, not kB (Sage Weil)
• mon: prevent implicit destruction of OSDs with ‘osd setmaxosd …’ (#8865, Anand Bhat)
• mon: verify all quorum members are contiguous at end of Paxos round (#9053, Sage Weil)
• msgr: refactor to cleanly separate SimpleMessenger implemenetation, move toward Connection-based calls (Matt Benjamin, Sage Wei)
• objectstore: clean up KeyValueDB interface for key/value backends (Sage Weil)
• osd: add local_mtime for use by cache agent (Zhiqiang Wang)
• osd: add superblock for KeyValueStore backend (Haomai Wang)
• osd: add support for Intel ISA-L erasure code library (Andreas-Joachim Peters)
• osd: do not skip promote for write-ordered reads (#9064, Samuel Just)
• osd: fix ambigous encoding order for blacklisted clients (#9211, Sage Weil)
• osd: fix cache flush corner case for snapshotted objects (#9054, Samuel Just)
• osd: fix discard of old/obsolete subop replies (#9259, Samuel Just)
• osd: fix discard of peer messages from previous intervals (Greg Farnum)
• osd: fix dump of open fds on EMFILE (Sage Weil)
• osd: fix journal dump (Ma Jianpeng)
• osd: fix mon feature bit requirements bug and resulting log spam (Sage Weil)
• osd: fix recovery chunk size usage during EC recovery (Ma Jianpeng)
• osd: fix recovery reservation deadlock for EC pools (Samuel Just)
• osd: fix removal of old xattrs when overwriting chained xattrs (Ma Jianpeng)
• osd: fix requesting queueing on PG split (Samuel Just)
• osd: force new xattrs into leveldb if fs returns E2BIG (#7779, Sage Weil)
• osd: implement alignment on chunk sizes (Loic Dachary)
• osd: improve prioritization of recovery of degraded over misplaced objects (Sage Weil)
• osd: locking, sharding, caching improvements in FileStore’s FDCache (Somnath Roy, Greg Farnum)
• osd: many important bug fixes (Samuel Just)
• osd, mon: add rocksdb support (Xinxin Shu, Sage Weil)
• osd, mon: distinguish between “misplaced” and “degraded” objects in cluster health and PG state reporting (Sage Weil)
• osd: refactor some ErasureCode functionality into command parent class (Loic Dachary)
• osd: set rollback_info_completed on create (#8625, Samuel Just)
• rados: allow setxattr value to be read from stdin (Sage Weil)
• rados: drop gratuitous n from getxattr command (Sage Weil)
• rgw: add –min-rewrite-stripe-size for object restriper (Yehuda Sadeh)
• rgw: add powerdns hook for dynamic DNS for global clusters (Wido den Hollander)
• rgw: copy object data is target bucket is in a different pool (#9039, Yehuda Sadeh)
• rgw: do not try to authenticate CORS preflight requests (#8718, Robert Hubbard, Yehuda Sadeh)
• rgw: fix civetweb URL decoding (#8621, Yehuda Sadeh)
• rgw: fix removal of objects during object creation (Patrycja Szablowska, Yehuda Sadeh)
• rgw: fix striping for copied objects (#9089, Yehuda Sadeh)
• rgw: fix test for identify whether an object has a tail (#9226, Yehuda Sadeh)
• rgw: fix when stripe size is not a multiple of chunk size (#8937, Yehuda Sadeh)
• rgw: improve civetweb logging (Yehuda Sadeh)
• rgw: misc civetweb frontend fixes (Yehuda Sadeh)
• sysvinit: add support for non-default cluster names (Alfredo Deza)

v0.84

The next Ceph development release is here! This release containsseveral meaty items, including some MDS improvements for journaling,the ability to remove the CephFS file system (and name it), severalmon cleanups with tiered pools, several OSD performance branches, anew “read forward” RADOS caching mode, a prototype Kinetic OSDbackend, and various radosgw improvements (especially with the newstandalone civetweb frontend). And there are a zillion OSD bugfixes. Things are looking pretty good for the Giant release that iscoming up in the next month.

Upgrading

• The _kbperf counters on the monitor have been removed. These arereplaced with a new set of _bytes counters (e.g., cluster_osd_kb isreplaced by cluster_osd_bytes).
• The rd_kb and wr_kb fields in the JSON dumps for pool stats (accessedvia the cephdfdetail-fjson-pretty and related commands) have beenreplaced with corresponding *_bytes fields. Similarly, thetotal_space, total_used, and total_avail fields are replaced withtotal_bytes, total_used_bytes, and total_avail_bytes fields.
• The radosdf—format=json output read_bytes and write_bytesfields were incorrectly reporting ops; this is now fixed.
• The radosdf—format=json output previously included read_kb andwrite_kb fields; these have been removed. Please use read_bytes andwrite_bytes instead (and divide by 1024 if appropriate).

Notable Changes

• ceph-conf: flush log on exit (Sage Weil)
• ceph-dencoder: refactor build a bit to limit dependencies (Sage Weil, Dan Mick)
• ceph.spec: split out ceph-common package, other fixes (Sandon Van Ness)
• ceph_test_librbd_fsx: fix RNG, make deterministic (Ilya Dryomov)
• cephtool: refactor and improve CLI tests (Joao Eduardo Luis)
• client: improved MDS session dumps (John Spray)
• common: fix dup log messages (#9080, Sage Weil)
• crush: include new tunables in dump (Sage Weil)
• crush: only require rule features if the rule is used (#8963, Sage Weil)
• crushtool: send output to stdout, not stderr (Wido den Hollander)
• fix i386 builds (Sage Weil)
• fix struct vs class inconsistencies (Thorsten Behrens)
• hadoop: update hadoop tests for Hadoop 2.0 (Haumin Chen)
• librbd, ceph-fuse: reduce cache flush overhead (Haomai Wang)
• librbd: fix error path when opening image (#8912, Josh Durgin)
• mds: add file system name, enabled flag (John Spray)
• mds: boot refactor, cleanup (John Spray)
• mds: fix journal conversion with standby-replay (John Spray)
• mds: separate inode recovery queue (John Spray)
• mds: session ls, evict commands (John Spray)
• mds: submit log events in async thread (Yan, Zheng)
• mds: use client-provided timestamp for user-visible file metadata (Yan, Zheng)
• mds: validate journal header on load and save (John Spray)
• misc build fixes for OS X (John Spray)
• misc integer size cleanups (Kevin Cox)
• mon: add get-quota commands (Joao Eduardo Luis)
• mon: do not create file system by default (John Spray)
• mon: fix ‘ceph df’ output for available space (Xiaoxi Chen)
• mon: fix bug when no auth keys are present (#8851, Joao Eduardo Luis)
• mon: fix compat version for MForward (Joao Eduardo Luis)
• mon: restrict some pool properties to tiered pools (Joao Eduardo Luis)
• msgr: misc locking fixes for fast dispatch (#8891, Sage Weil)
• osd: add ‘dump_reservations’ admin socket command (Sage Weil)
• osd: add READFORWARD caching mode (Luis Pabon)
• osd: add header cache for KeyValueStore (Haomai Wang)
• osd: add prototype KineticStore based on Seagate Kinetic (Josh Durgin)
• osd: allow map cache size to be adjusted at runtime (Sage Weil)
• osd: avoid refcounting overhead by passing a few things by ref (Somnath Roy)
• osd: avoid sharing PG info that is not durable (Samuel Just)
• osd: clear slow request latency info on osd up/down (Sage Weil)
• osd: fix PG object listing/ordering bug (Guang Yang)
• osd: fix PG stat errors with tiering (#9082, Sage Weil)
• osd: fix bug with long object names and rename (#8701, Sage Weil)
• osd: fix cache full -> not full requeueing (#8931, Sage Weil)
• osd: fix gating of messages from old OSD instances (Greg Farnum)
• osd: fix memstore bugs with collection_move_rename, lock ordering (Sage Weil)
• osd: improve locking for KeyValueStore (Haomai Wang)
• osd: make tiering behave if hit_sets aren’t enabled (Sage Weil)
• osd: mark pools with incomplete clones (Sage Weil)
• osd: misc locking fixes for fast dispatch (Samuel Just, Ma Jianpeng)
• osd: prevent old rados clients from using tiered pools (#8714, Sage Weil)
• osd: reduce OpTracker overhead (Somnath Roy)
• osd: set configurable hard limits on object and xattr names (Sage Weil, Haomai Wang)
• osd: trim old EC objects quickly; verify on scrub (Samuel Just)
• osd: work around GCC 4.8 bug in journal code (Matt Benjamin)
• rados bench: fix arg order (Kevin Dalley)
• rados: fix {read,write}_ops values for df output (Sage Weil)
• rbd: add rbdmap pre- and post post- hooks, fix misc bugs (Dmitry Smirnov)
• rbd: improve option default behavior (Josh Durgin)
• rgw: automatically align writes to EC pool (#8442, Yehuda Sadeh)
• rgw: fix crash on swift CORS preflight request (#8586, Yehuda Sadeh)
• rgw: fix memory leaks (Andrey Kuznetsov)
• rgw: fix multipart upload (#8846, Silvain Munaut, Yehuda Sadeh)
• rgw: improve -h (Abhishek Lekshmanan)
• rgw: improve delimited listing of bucket, misc fixes (Yehuda Sadeh)
• rgw: misc civetweb fixes (Yehuda Sadeh)
• rgw: powerdns backend for global namespaces (Wido den Hollander)
• systemd: initial systemd config files (Federico Simoncelli)

v0.83

Another Ceph development release! This has been a longer cycle, sothere has been quite a bit of bug fixing and stabilization in thisround. There is also a bunch of packaging fixes for RPM distros(RHEL/CentOS, Fedora, and SUSE) and for systemd. We’ve also added a newlibrados-striper library from Sebastien Ponce that provides a genericstriping API for applications to code to.

Upgrading

• The experimental keyvaluestore-dev OSD backend had an on-disk formatchange that prevents existing OSD data from being upgraded. Thisaffects developers and testers only.

• mon-specific and osd-specific leveldb options have been removed.From this point onward users should use the leveldb*_ genericoptions and add the options in the appropriate sections of theirconfiguration files. Monitors will still maintain the followingmonitor-specific defaults:

leveldb_write_buffer_size = 3210241024 = 33554432 // 32MBleveldb_cache_size = 51210241204 = 536870912 // 512MBleveldb_block_size = 64*1024 = 65536 // 64KBleveldb_compression = falseleveldb_log = “”

OSDs will still maintain the following osd-specific defaults:

leveldb_log = “”

Notable Changes

• ceph-disk: fix dmcrypt support (Stephen Taylor)
• cephtool: fix help (Yilong Zhao)
• cephtool: test cleanup (Joao Eduardo Luis)
• doc: librados example fixes (Kevin Dalley)
• doc: many doc updates (John Wilkins)
• doc: update erasure docs (Loic Dachary, Venky Shankar)
• filestore: disable use of XFS hint (buggy on old kernels) (Samuel Just)
• filestore: fix xattr spillout (Greg Farnum, Haomai Wang)
• keyvaluestore: header cache (Haomai Wang)
• librados_striper: striping library for librados (Sebastien Ponce)
• libs3: update to latest (Danny Al-Gaaf)
• log: fix derr level (Joao Eduardo Luis)
• logrotate: fix osd log rotation on ubuntu (Sage Weil)
• mds: fix xattr bug triggered by ACLs (Yan, Zheng)
• misc memory leaks, cleanups, fixes (Danny Al-Gaaf, Sahid Ferdjaoui)
• misc suse fixes (Danny Al-Gaaf)
• misc word size fixes (Kevin Cox)
• mon: drop mon- and osd- specific leveldb options (Joao Eduardo Luis)
• mon: ec pool profile fixes (Loic Dachary)
• mon: fix health down messages (Sage Weil)
• mon: fix quorum feature check (#8738, Greg Farnum)
• mon: ‘osd crush reweight-subtree …’ (Sage Weil)
• mon, osd: relax client EC support requirements (Sage Weil)
• mon: some instrumentation (Sage Weil)
• objecter: flag operations that are redirected by caching (Sage Weil)
• osd: clean up shard_id_t, shard_t (Loic Dachary)
• osd: fix connection reconnect race (Greg Farnum)
• osd: fix dumps (Joao Eduardo Luis)
• osd: fix erasure-code lib initialization (Loic Dachary)
• osd: fix extent normalization (Adam Crume)
• osd: fix loopback msgr issue (Ma Jianpeng)
• osd: fix LSB release parsing (Danny Al-Gaaf)
• osd: improved backfill priorities (Sage Weil)
• osd: many many core fixes (Samuel Just)
• osd, mon: config sanity checks on start (Sage Weil, Joao Eduardo Luis)
• osd: sharded threadpool to improve parallelism (Somnath Roy)
• osd: simple io prioritization for scrub (Sage Weil)
• osd: simple scrub throttling (Sage Weil)
• osd: tests for bench command (Loic Dachary)
• osd: use xfs hint less frequently (Ilya Dryomov)
• pybind/rados: fix small timeouts (John Spray)
• qa: xfstests updates (Ilya Dryomov)
• rgw: cache bucket info (Yehuda Sadeh)
• rgw: cache decoded user info (Yehuda Sadeh)
• rgw: fix multipart object attr regression (#8452, Yehuda Sadeh)
• rgw: fix radosgw-admin ‘show log’ command (#8553, Yehuda Sadeh)
• rgw: fix URL decoding (#8702, Brian Rak)
• rgw: handle empty extra pool name (Yehuda Sadeh)
• rpm: do not restart daemons on upgrade (Alfredo Deza)
• rpm: misc packaging fixes for rhel7 (Sandon Van Ness)
• rpm: split ceph-common from ceph (Sandon Van Ness)
• systemd: wrap started daemons in new systemd environment (Sage Weil, Dan Mick)
• sysvinit: less sensitive to failures (Sage Weil)
• upstart: increase max open files limit (Sage Weil)

v0.82

This is the second post-firefly development release. It includes a rangeof bug fixes and some usability improvements. There are some MDS debuggingand diagnostic tools, an improved ‘ceph df’, and some OSD backend refactoringand cleanup.

Notable Changes

• ceph-brag: add tox tests (Alfredo Deza)
• common: perfcounters now use atomics and go faster (Sage Weil)
• doc: CRUSH updates (John Wilkins)
• doc: osd primary affinity (John Wilkins)
• doc: pool quotas (John Wilkins)
• doc: pre-flight doc improvements (Kevin Dalley)
• doc: switch to an unencumbered font (Ross Turk)
• doc: update openstack docs (Josh Durgin)
• fix hppa arch build (Dmitry Smirnov)
• init-ceph: continue starting other daemons on crush or mount failure (#8343, Sage Weil)
• keyvaluestore: fix hint crash (#8381, Haomai Wang)
• libcephfs-java: build against older JNI headers (Greg Farnum)
• librados: fix rados_pool_list bounds checks (Sage Weil)
• mds: cephfs-journal-tool (John Spray)
• mds: improve Journaler on-disk format (John Spray)
• mds, libcephfs: use client timestamp for mtime/ctime (Sage Weil)
• mds: misc encoding improvements (John Spray)
• mds: misc fixes for multi-mds (Yan, Zheng)
• mds: OPTracker integration, dump_ops_in_flight (Greg Farnum)
• misc cleanup (Christophe Courtaut)
• mon: fix default replication pool ruleset choice (#8373, John Spray)
• mon: fix set cache_target_full_ratio (#8440, Geoffrey Hartz)
• mon: include per-pool ‘max avail’ in df output (Sage Weil)
• mon: prevent EC pools from being used with cephfs (Joao Eduardo Luis)
• mon: restore original weight when auto-marked out OSDs restart (Sage Weil)
• mon: use msg header tid for MMonGetVersionReply (Ilya Dryomov)
• osd: fix bogus assert during OSD shutdown (Sage Weil)
• osd: fix clone deletion case (#8334, Sam Just)
• osd: fix filestore removal corner case (#8332, Sam Just)
• osd: fix hang waiting for osdmap (#8338, Greg Farnum)
• osd: fix interval check corner case during peering (#8104, Sam Just)
• osd: fix journal-less operation (Sage Weil)
• osd: include backend information in metadata reported to mon (Sage Weil)
• rest-api: fix help (Ailing Zhang)
• rgw: check entity permission for put_metadata (#8428, Yehuda Sadeh)

v0.81

This is the first development release since Firefly. It includes alot of work that we delayed merging while stabilizing things. Lots ofnew functionality, as well as several fixes that are baking a bit beforegetting backported.

Upgrading

• CephFS support for the legacy anchor table has finally been removed.Users with file systems created before firefly should ensure that inodeswith multiple hard links are modified prior to the upgrade to ensure thatthe backtraces are written properly. For example:

sudo find /mnt/cephfs -type f -links +1 -exec touch \{\} \;

• Disallow nonsensical ‘tier cache-mode’ transitions. From this pointonward, ‘writeback’ can only transition to ‘forward’ and ‘forward’can transition to 1) ‘writeback’ if there are dirty objects, or 2) any ifthere are no dirty objects.

Notable Changes

• bash completion improvements (Wido den Hollander)
• brag: fixes, improvements (Loic Dachary)
• ceph-disk: handle corrupt volumes (Stuart Longlang)
• ceph-disk: partprobe as needed (Eric Eastman)
• ceph-fuse, libcephfs: asok hooks for handling session resets, timeouts (Yan, Zheng)
• ceph-fuse, libcephfs: improve traceless reply handling (Sage Weil)
• clang build fixes (John Spray, Danny Al-Gaaf)
• config: support G, M, K, etc. suffixes (Joao Eduardo Luis)
• coverity cleanups (Danny Al-Gaaf)
• doc: cache tiering (John Wilkins)
• doc: keystone integration docs (John Wilkins)
• doc: updated simple configuration guides (John Wilkins)
• libcephfs-java: fix gcj-jdk build (Dmitry Smirnov)
• librbd: check error code on cache invalidate (Josh Durgin)
• librbd: new libkrbd library for kernel map/unmap/showmapped (Ilya Dryomov)
• Makefile: fix out of source builds (Stefan Eilemann)
• mds: multi-mds fixes (Yan, Zheng)
• mds: remove legacy anchor table (Yan, Zheng)
• mds: remove legacy discover ino (Yan, Zheng)
• monclient: fix hang (Sage Weil)
• mon: prevent nonsensical cache-mode transitions (Joao Eduardo Luis)
• msgr: avoid big lock when sending (most) messages (Greg Farnum)
• osd: bound osdmap epoch skew between PGs (Sage Weil)
• osd: cache tier flushing fixes for snapped objects (Samuel Just)
• osd: fix agent early finish looping (David Zafman)
• osd: fix flush vs OpContext (Samuel Just)
• osd: fix MarkMeDown and other shutdown races (Sage Weil)
• osd: fix scrub vs cache bugs (Samuel Just)
• osd: fix trim of hitsets (Sage Weil)
• osd, msgr: fast-dispatch of OSD ops (Greg Farnum, Samuel Just)
• osd, objecter: resend ops on last_force_op_resend barrier; fix cache overlay op ordering (Sage Weil)
• osd: remove obsolete classic scrub code (David Zafman)
• osd: scrub PGs with invalid stats (Sage Weil)
• osd: simple snap trimmer throttle (Sage Weil)
• osd: use FIEMAP to inform copy_range (Haomai Wang)
• rbd-fuse: allow exposing single image (Stephen Taylor)
• rbd-fuse: fix unlink (Josh Durgin)
• removed mkcephfs (deprecated since dumpling)
• rgw: bucket link uses instance id (Yehuda Sadeh)
• rgw: fix memory leak following chunk read error (Yehuda Sadeh)
• rgw: fix URL escaping (Yehuda Sadeh)
• rgw: fix user manifest (Yehuda Sadeh)
• rgw: object and bucket rewrite functions to allow restriping old objects (Yehuda Sadeh)
• rgw: prevent multiobject PUT race (Yehuda Sadeh)
• rgw: send user manifest header (Yehuda Sadeh)
• test_librbd_fsx: test krbd as well as librbd (Ilya Dryomov)

v0.80.11 Firefly

This is a bugfix release for Firefly. This Firefly 0.80.x is nearingits planned end of life in January 2016 it may also be the last.

We recommend that all Firefly users upgrade.

For more detailed information, see thecompletechangelog.

Notable Changes

• build/ops: /etc/init.d/radosgw restart does not work correctly (issue#11140, pr#5831, Dmitry Yatsushkevich)
• build/ops: Fix -Wno-format and -Werror=format-security options clash (issue#13417, pr#6207, Boris Ranto)
• build/ops: ceph-common needs python-argparse on older distros, but doesn’t require it (issue#12034, pr#5217, Nathan Cutler)
• build/ops: ceph.spec.in running fdupes unnecessarily (issue#12301, pr#5224, Nathan Cutler)
• build/ops: ceph.spec.in: 50-rbd.rules conditional is wrong (issue#12166, pr#5225, Nathan Cutler)
• build/ops: ceph.spec.in: useless %py_requires breaks SLE11-SP3 build (issue#12351, pr#5394, Nathan Cutler)
• build/ops: fedora21 has junit, not junit4 (issue#10728, pr#6203, Ken Dreyer, Loic Dachary)
• build/ops: upstart: configuration is too generous on restarts (issue#11798, pr#5992, Sage Weil)
• common: Client admin socket leaks file descriptors (issue#11535, pr#4633, Jon Bernard)
• common: FileStore calls syncfs(2) even it is not supported (issue#12512, pr#5529, Danny Al-Gaaf, Kefu Chai, Jianpeng Ma)
• common: HeartBeat: include types (issue#13088, pr#6038, Sage Weil)
• common: Malformed JSON command output when non-ASCII strings are present (issue#7387, pr#4635, Kefu Chai, Tim Serong)
• common: Memory leak in Mutex.cc, pthread_mutexattr_init without pthread_mutexattr_destroy (issue#11762, pr#5403, Ketor Meng)
• common: Thread:pthread_attr_destroy(thread_attr) when done with it (issue#12570, pr#6325, Piotr Dałek, Zheng Qiankun)
• common: ThreadPool add/remove work queue methods not thread safe (issue#12662, pr#5991, Jason Dillaman)
• common: buffer: critical bufferlist::zero bug (issue#12252, pr#5388, Haomai Wang)
• common: log: take mutex while opening fd (issue#12465, pr#5406, Samuel Just)
• common: recursive lock of md_config_t (0) (issue#12614, pr#5814, Josh Durgin)
• crush: take crashes due to invalid arg (issue#11602, pr#4769, Sage Weil)
• doc: backport v0.80.10 release notes to firefly (issue#11090, pr#5307, Loic Dachary, Sage Weil)
• doc: update docs to point to download.ceph.com (issue#13162, pr#5993, Alfredo Deza)
• fs: MDSMonitor: handle MDSBeacon messages properly (issue#11590, pr#5199, Kefu Chai)
• fs: client nonce collision due to unshared pid namespaces (issue#13032, pr#6087, Josh Durgin, Sage Weil)
• librbd: Objectcacher setting max object counts too low (issue#7385, pr#4639, Jason Dillaman)
• librbd: aio calls may block (issue#11056, pr#4854, Haomai Wang, Sage Weil, Jason Dillaman)
• librbd: internal.cc: 1967: FAILED assert(watchers.size() == 1) (issue#12176, pr#5171, Jason Dillaman)
• mon: Clock skew causes missing summary and confuses Calamari (issue#11877, pr#4867, Thorsten Behrens)
• mon: EC pools are not allowed as cache pools, disallow in the mon (issue#11650, pr#5389, Samuel Just)
• mon: Make it more difficult to delete pools in firefly (issue#11800, pr#4788, Sage Weil)
• mon: MonitorDBStore: get_next_key() only if prefix matches (issue#11786, pr#5360, Joao Eduardo Luis)
• mon: PaxosService: call post_refresh() instead of post_paxos_update() (issue#11470, pr#5358, Joao Eduardo Luis)
• mon: add a cache layer over MonitorDBStore (issue#12638, pr#5698, Kefu Chai)
• mon: adding exsting pool as tier with –force-nonempty clobbers removed_snaps (issue#11493, pr#5236, Sage Weil, Samuel Just)
• mon: ceph fails to compile with boost 1.58 (issue#11576, pr#5129, Kefu Chai)
• mon: does not check for IO errors on every transaction (issue#13089, pr#6091, Sage Weil)
• mon: get pools health’info have error (issue#12402, pr#5410, renhwztetecs)
• mon: increase globalid default for firefly (issue#13255, pr#6010, Sage Weil)
• mon: pgmonitor: wrong at/near target max“ reporting (issue#12401, pr#5409, huangjun)
• mon: register_new_pgs() should check ruleno instead of its index (issue#12210, pr#5404, Xinze Chi)
• mon: scrub error (osdmap encoding mismatch?) upgrading from 0.80 to ~0.80.2 (issue#8815, issue#8674, issue#9064, pr#5200, Sage Weil, Zhiqiang Wang, Samuel Just)
• mon: the output is wrong when runing ceph osd reweight (issue#12251, pr#5408, Joao Eduardo Luis)
• objecter: can get stuck in redirect loop if osdmap epoch == last_force_op_resend (issue#11026, pr#4597, Jianpeng Ma, Sage Weil)
• objecter: pg listing can deadlock when throttling is in use (issue#9008, pr#5043, Guang Yang)
• objecter: resend linger ops on split (issue#9806, pr#5062, Josh Durgin, Samuel Just)
• osd: Cleanup boost optionals for boost 1.56 (issue#9983, pr#5039, William A. Kennington III)
• osd: LibRadosTwoPools[EC]PP.PromoteSnap failure (issue#10052, pr#5050, Sage Weil)
• osd: Mutex Assert from PipeConnection::try_get_pipe (issue#12437, pr#5815, David Zafman)
• osd: PG stuck with remapped (issue#9614, pr#5044, Guang Yang)
• osd: PG::handle_advance_map: on_pool_change after handling the map change (issue#12809, pr#5988, Samuel Just)
• osd: PGLog: split divergent priors as well (issue#11069, pr#4631, Samuel Just)
• osd: PGLog::proc_replica_log: correctly handle case where entries between olog.head and log.tail were split out (issue#11358, pr#5287, Samuel Just)
• osd: WBThrottle::clear_object: signal on cond when we reduce throttle values (issue#12223, pr#5822, Samuel Just)
• osd: cache full mode still skips young objects (issue#10006, pr#5051, Xinze Chi, Zhiqiang Wang)
• osd: crash creating/deleting pools (issue#12429, pr#5526, John Spray)
• osd: explicitly specify OSD features in MOSDBoot (issue#10911, pr#4960, Sage Weil)
• osd: is_new_interval() fixes (issue#11771, issue#10399, pr#5726, Samuel Just, Jason Dillaman)
• osd: make the all osd/filestore thread pool suicide timeouts separately configurable (issue#11439, pr#5823, Samuel Just)
• osd: object creation by write cannot use an offset on an erasure coded pool (issue#11507, pr#4632, Jianpeng Ma, Loic Dachary)
• osd: os/FileJournal: Fix journal write fail, align for direct io (issue#12943, pr#5619, Xie Rui)
• osd: osd/PGLog.cc: 732: FAILED assert(log.log.size() == log_keys_debug.size()) (issue#12652, pr#5820, Sage Weil)
• osd: read on chunk-aligned xattr not handled (issue#12309, pr#5235, Sage Weil)
• rgw: Change variable length array of std::strings (not legal in C++) to std::vector (issue#12467, pr#4583, Daniel J. Hofmann)
• rgw: Civetweb RGW appears to report full size of object as downloaded when only partially downloaded (issue#11851, pr#5234, Yehuda Sadeh)
• rgw: Keystone PKI token expiration is not enforced (issue#11367, pr#4765, Anton Aksola)
• rgw: Object copy bug (issue#11639, pr#4762, Javier M. Mellid)
• rgw: RGW returns requested bucket name raw in “Bucket” response header (issue#11860, issue#12537, pr#5730, Yehuda Sadeh, Wido den Hollander)
• rgw: Swift API: response for PUT on /container does not contain the mandatory Content-Length header when FCGI is used (issue#11036, pr#5170, Radoslaw Zarzynski)
• rgw: content length parsing calls strtol() instead of strtoll() (issue#10701, pr#5997, Yehuda Sadeh)
• rgw: delete bucket does not remove .bucket.meta file (issue#11149, pr#4641, Orit Wasserman)
• rgw: doesn’t return ‘x-timestamp’ in header which is used by ‘View Details’ of OpenStack (issue#8911, pr#4584, Yehuda Sadeh)
• rgw: init some manifest fields when handling explicit objs (issue#11455, pr#5729, Yehuda Sadeh)
• rgw: logfile does not get chowned properly (issue#12073, pr#5233, Thorsten Behrens)
• rgw: logrotate.conf calls service with wrong init script name (issue#12043, pr#5390, wuxingyi)
• rgw: quota not respected in POST object (issue#11323, pr#4642, Sergey Arkhipov)
• rgw: swift smoke test fails on TestAccountUTF8 (issue#11091, issue#11438, issue#12939, issue#12157, issue#12158, issue#12363, pr#5532, Radoslaw Zarzynski, Orit Wasserman, Robin H. Johnson)
• rgw: use correct objv_tracker for bucket instance (issue#11416, pr#4535, Yehuda Sadeh)
• tests: ceph-fuse crash in test_client_recovery (issue#12673, pr#5813, Loic Dachary)
• tests: kernel_untar_build fails on EL7 (issue#11758, pr#6000, Greg Farnum)
• tests: qemu workunit refers to apt-mirror.front.sepia.ceph.com (issue#13420, pr#6328, Yuan Zhou, Sage Weil)
• tools: src/ceph-disk : disk zap sgdisk invocation (issue#11143, pr#4636, Thorsten Behrens, Owen Synge)
• tools: ceph-disk: sometimes the journal symlink is not created (issue#10146, pr#5541, Dan van der Ster)
• tools: ceph-disk: support NVMe device partitions (issue#11612, pr#4771, Ilja Slepnev)
• tools: ceph-post-file fails on rhel7 (issue#11836, pr#5037, Joseph McDonald, Sage Weil)
• tools: ceph_argparse_flag has no regular 3rd parameter (issue#11543, pr#4582, Thorsten Behrens)
• tools: use a new disk as journal disk,ceph-disk prepare fail (issue#10983, pr#4630, Loic Dachary)

v0.80.10 Firefly

This is a bugfix release for Firefly.

We recommend that all Firefly users upgrade.

For more detailed information, see thecompletechangelog.

Notable Changes

• build/ops: ceph.spec.in: package mkcephfs on EL6 (issue#11955, pr#4924, Ken Dreyer)
• build/ops: debian: ceph-test and rest-bench debug packages should require their respective binary packages (issue#11673, pr#4766, Ken Dreyer)
• build/ops: run RGW as root (issue#11453, pr#4638, Ken Dreyer)
• common: messages/MWatchNotify: include an error code in the message (issue#9193, pr#3944, Sage Weil)
• common: Rados.shutdown() dies with Illegal instruction (core dumped) (issue#10153, pr#3963, Federico Simoncelli)
• common: SimpleMessenger: allow RESETSESSION whenever we forget an endpoint (issue#10080, pr#3915, Greg Farnum)
• common: WorkQueue: make wait timeout on empty queue configurable (issue#10817, pr#3941, Samuel Just)
• crush: set_choose_tries = 100 for erasure code rulesets (issue#10353, pr#3824, Loic Dachary)
• doc: backport ceph-disk man page to Firefly (issue#10724, pr#3936, Nilamdyuti Goswami)
• doc: Fix ceph command manpage to match ceph -h (issue#10676, pr#3996, David Zafman)
• fs: mount.ceph: avoid spurious error message (issue#10351, pr#3927, Yan, Zheng)
• librados: Fix memory leak in python rados bindings (issue#10723, pr#3935, Josh Durgin)
• librados: fix resources leakage in RadosClient::connect() (issue#10425, pr#3828, Radoslaw Zarzynski)
• librados: Translate operation flags from C APIs (issue#10497, pr#3930, Matt Richards)
• librbd: acquire cache_lock before refreshing parent (issue#5488, pr#4206, Jason Dillaman)
• librbd: snap_remove should ignore -ENOENT errors (issue#11113, pr#4245, Jason Dillaman)
• mds: fix assertion caused by system clock backwards (issue#11053, pr#3970, Yan, Zheng)
• mon: ignore osd failures from before up_from (issue#10762, pr#3937, Sage Weil)
• mon: MonCap: take EntityName instead when expanding profiles (issue#10844, pr#3942, Joao Eduardo Luis)
• mon: Monitor: fix timecheck rounds period (issue#10546, pr#3932, Joao Eduardo Luis)
• mon: OSDMonitor: do not trust small values in osd epoch cache (issue#10787, pr#3823, Sage Weil)
• mon: OSDMonitor: fallback to json-pretty in case of invalid formatter (issue#9538, pr#4475, Loic Dachary)
• mon: PGMonitor: several stats output error fixes (issue#10257, pr#3826, Joao Eduardo Luis)
• objecter: fix map skipping (issue#9986, pr#3952, Ding Dinghua)
• osd: cache tiering: fix the atime logic of the eviction (issue#9915, pr#3949, Zhiqiang Wang)
• osd: cancel_pull: requeue waiters (issue#11244, pr#4415, Samuel Just)
• osd: check that source OSD is valid for MOSDRepScrub (issue#9555, pr#3947, Sage Weil)
• osd: DBObjectMap: lock header_lock on sync() (issue#9891, pr#3948, Samuel Just)
• osd: do not ignore deleted pgs on startup (issue#10617, pr#3933, Sage Weil)
• osd: ENOENT on clone (issue#11199, pr#4385, Samuel Just)
• osd: erasure-code-profile set races with erasure-code-profile rm (issue#11144, pr#4383, Loic Dachary)
• osd: FAILED assert(soid < scrubber.start || soid >= scrubber.end) (issue#11156, pr#4185, Samuel Just)
• osd: FileJournal: fix journalq population in do_read_entry() (issue#6003, pr#3960, Samuel Just)
• osd: fix negative degraded objects during backfilling (issue#7737, pr#4021, Guang Yang)
• osd: get the currently atime of the object in cache pool for eviction (issue#9985, pr#3950, Sage Weil)
• osd: load_pgs: we need to handle the case where an upgrade from earlier versions which ignored non-existent pgs resurrects a pg with a prehistoric osdmap (issue#11429, pr#4556, Samuel Just)
• osd: ObjectStore: Don’t use largest_data_off to calc data_align. (issue#10014, pr#3954, Jianpeng Ma)
• osd: osd_types: op_queue_age_hist and fs_perf_stat should be in osd_stat_t::o… (issue#10259, pr#3827, Samuel Just)
• osd: PG::actingset should be used when checking the number of acting OSDs for… (issue#11454, pr#4453, Guang Yang)
• osd: PG::all_unfound_are_queried_or_lost for non-existent osds (issue#10976, pr#4416, Mykola Golub)
• osd: PG: always clear_primary_state (issue#10059, pr#3955, Samuel Just)
• osd: PGLog.h: 279: FAILED assert(log.log.size() == log_keys_debug.size()) (issue#10718, pr#4382, Samuel Just)
• osd: PGLog: include rollback_info_trimmed_to in (read|write)_log (issue#10157, pr#3964, Samuel Just)
• osd: pg stuck stale after create with activation delay (issue#11197, pr#4384, Samuel Just)
• osd: ReplicatedPG: fail a non-blocking flush if the object is being scrubbed (issue#8011, pr#3943, Samuel Just)
• osd: ReplicatedPG::on_change: clean up callbacks_for_degraded_object (issue#8753, pr#3940, Samuel Just)
• osd: ReplicatedPG::scan_range: an object can disappear between the list and t… (issue#10150, pr#3962, Samuel Just)
• osd: requeue blocked op before flush it was blocked on (issue#10512, pr#3931, Sage Weil)
• rgw: check for timestamp for s3 keystone auth (issue#10062, pr#3958, Abhishek Lekshmanan)
• rgw: civetweb should use unique request id (issue#11720, pr#4780, Orit Wasserman)
• rgw: don’t allow negative / invalid content length (issue#11890, pr#4829, Yehuda Sadeh)
• rgw: fail s3 POST auth if keystone not configured (issue#10698, pr#3966, Yehuda Sadeh)
• rgw: flush xml header on get acl request (issue#10106, pr#3961, Yehuda Sadeh)
• rgw: generate new tag for object when setting object attrs (issue#11256, pr#4571, Yehuda Sadeh)
• rgw: generate the “Date” HTTP header for civetweb. (issue#11871,11891, pr#4851, Radoslaw Zarzynski)
• rgw: keystone token cache does not work correctly (issue#11125, pr#4414, Yehuda Sadeh)
• rgw: merge manifests correctly when there’s prefix override (issue#11622, pr#4697, Yehuda Sadeh)
• rgw: send appropriate op to cancel bucket index pending operation (issue#10770, pr#3938, Yehuda Sadeh)
• rgw: shouldn’t need to disable rgw_socket_path if frontend is configured (issue#11160, pr#4275, Yehuda Sadeh)
• rgw: Swift API. Dump container’s custom metadata. (issue#10665, pr#3934, Dmytro Iurchenko)
• rgw: Swift API. Support for X-Remove-Container-Meta-{key} header. (issue#10475, pr#3929, Dmytro Iurchenko)
• rgw: use correct objv_tracker for bucket instance (issue#11416, pr#4379, Yehuda Sadeh)
• tests: force checkout of submodules (issue#11157, pr#4079, Loic Dachary)
• tools: Backport ceph-objectstore-tool changes to firefly (issue#12327, pr#3866, David Zafman)
• tools: ceph-objectstore-tool: Output only unsupported features when incomatible (issue#11176, pr#4126, David Zafman)
• tools: ceph-objectstore-tool: Use exit status 11 for incompatible import attemp… (issue#11139, pr#4129, David Zafman)
• tools: Fix do_autogen.sh so that -L is allowed (issue#11303, pr#4247, Alfredo Deza)

v0.80.9 Firefly

This is a bugfix release for firefly. It fixes a performanceregression in librbd, an important CRUSH misbehavior (see below), andseveral RGW bugs. We have also backported support for flock/fcntllocks to ceph-fuse and libcephfs.

We recommend that all Firefly users upgrade.

For more detailed information, see thecompletechangelog.

Adjusting CRUSH maps

• This point release fixes several issues with CRUSH that triggerexcessive data migration when adjusting OSD weights. These are mostobvious when a very small weight change (e.g., a change from 0 to.01) triggers a large amount of movement, but the same set of bugscan also lead to excessive (though less noticeable) movement inother cases.

However, because the bug may already have affected your cluster,fixing it may trigger movement back to the more correct location.For this reason, you must manually opt-in to the fixed behavior.

In order to set the new tunable to correct the behavior:

ceph osd crush set-tunable straw_calc_version 1

Note that this change will have no immediate effect. However, fromthis point forward, any ‘straw’ bucket in your CRUSH map that isadjusted will get non-buggy internal weights, and that transitionmay trigger some rebalancing.

You can estimate how much rebalancing will eventually be necessaryon your cluster with:

  ceph osd getcrushmap -o /tmp/cm
  crushtool -i /tmp/cm --num-rep 3 --test --show-mappings > /tmp/a 2>&1
  crushtool -i /tmp/cm --set-straw-calc-version 1 -o /tmp/cm2
  crushtool -i /tmp/cm2 --reweight -o /tmp/cm2
  crushtool -i /tmp/cm2 --num-rep 3 --test --show-mappings > /tmp/b 2>&1
  wc -l /tmp/a                          # num total mappings
  diff -u /tmp/a /tmp/b | grep -c ^+    # num changed mappings
 
Divide the number of changed lines by the total number of lines in
/tmp/a.  We've found that most clusters are under 10%.
 
You can force all of this rebalancing to happen at once with::
 
  ceph osd crush reweight-all
 
Otherwise, it will happen at some unknown point in the future when
CRUSH weights are next adjusted.

Notable Changes

• ceph-fuse: flock, fcntl lock support (Yan, Zheng, Greg Farnum)
• crush: fix straw bucket weight calculation, add straw_calc_version tunable (#10095 Sage Weil)
• crush: fix tree bucket (Rongzu Zhu)
• crush: fix underflow of tree weights (Loic Dachary, Sage Weil)
• crushtool: add –reweight (Sage Weil)
• librbd: complete pending operations before losing image (#10299 Jason Dillaman)
• librbd: fix read caching performance regression (#9854 Jason Dillaman)
• librbd: gracefully handle deleted/renamed pools (#10270 Jason Dillaman)
• mon: fix dump of chooseleaf_vary_r tunable (Sage Weil)
• osd: fix PG ref leak in snaptrimmer on peering (#10421 Kefu Chai)
• osd: handle no-op write with snapshot (#10262 Sage Weil)
• radosgw-admin: create subuser when creating user (#10103 Yehuda Sadeh)
• rgw: change multipart uplaod id magic (#10271 Georgio Dimitrakakis, Yehuda Sadeh)
• rgw: don’t overwrite bucket/object owner when setting ACLs (#10978 Yehuda Sadeh)
• rgw: enable IPv6 for embedded civetweb (#10965 Yehuda Sadeh)
• rgw: fix partial swift GET (#10553 Yehuda Sadeh)
• rgw: fix quota disable (#9907 Dong Lei)
• rgw: index swift keys appropriately (#10471 Hemant Burman, Yehuda Sadeh)
• rgw: make setattrs update bucket index (#5595 Yehuda Sadeh)
• rgw: pass civetweb configurables (#10907 Yehuda Sadeh)
• rgw: remove swift user manifest (DLO) hash calculation (#9973 Yehuda Sadeh)
• rgw: return correct len for 0-len objects (#9877 Yehuda Sadeh)
• rgw: S3 object copy content-type fix (#9478 Yehuda Sadeh)
• rgw: send ETag on S3 object copy (#9479 Yehuda Sadeh)
• rgw: send HTTP status reason explicitly in fastcgi (Yehuda Sadeh)
• rgw: set ulimit -n from sysvinit (el6) init script (#9587 Sage Weil)
• rgw: update swift subuser permission masks when authenticating (#9918 Yehuda Sadeh)
• rgw: URL decode query params correctly (#10271 Georgio Dimitrakakis, Yehuda Sadeh)
• rgw: use attrs when reading object attrs (#10307 Yehuda Sadeh)
• rgw: use rn for http headers (#9254 Benedikt Fraunhofer, Yehuda Sadeh)

v0.80.8 Firefly

This is a long-awaited bugfix release for firefly. It has severalimporant (but relatively rare) OSD peering fixes, performance issueswhen snapshots are trimmed, several RGW fixes, a paxos corner casefix, and some packaging updates.

We recommend that all users for v0.80.x firefly upgrade when it isconvenient to do so.

For more detailed information, see thecompletechangelog.

Notable Changes

• build: remove stack-execute bit from assembled code sections (#10114 Dan Mick)
• ceph-disk: fix dmcrypt key permissions (#9785 Loic Dachary)
• ceph-disk: fix keyring location (#9653 Loic Dachary)
• ceph-disk: make partition checks more robust (#9721 #9665 Loic Dachary)
• ceph: cleanly shut down librados context on shutdown (#8797 Dan Mick)
• common: add $cctid config metavariable (#6228 Adam Crume)
• crush: align rule and ruleset ids (#9675 Xiaoxi Chen)
• crush: fix negative weight bug during create_or_move_item (#9998 Pawel Sadowski)
• crush: fix potential buffer overflow in erasure rules (#9492 Johnu George)
• debian: fix python-ceph -> ceph file movement (Sage Weil)
• libcephfs,ceph-fuse: fix flush tid wraparound bug (#9869 Greg Farnum, Yan, Zheng)
• libcephfs: close fd befure umount (#10415 Yan, Zheng)
• librados: fix crash from C API when read timeout is enabled (#9582 Sage Weil)
• librados: handle reply race with pool deletion (#10372 Sage Weil)
• librbd: cap memory utilization for read requests (Jason Dillaman)
• librbd: do not close a closed parent image on failure (#10030 Jason Dillaman)
• librbd: fix diff tests (#10002 Josh Durgin)
• librbd: protect list_children from invalid pools (#10123 Jason Dillaman)
• make check improvemens (Loic Dachary)
• mds: fix ctime updates (#9514 Greg Farnum)
• mds: fix journal import tool (#10025 John Spray)
• mds: fix rare NULL deref in cap flush handler (Greg Farnum)
• mds: handle unknown lock messages (Yan, Zheng)
• mds: store backtrace for straydir (Yan, Zheng)
• mon: abort startup if disk is full (#9502 Joao Eduardo Luis)
• mon: add paxos instrumentation (Sage Weil)
• mon: fix double-free in rare OSD startup path (Sage Weil)
• mon: fix osdmap trimming (#9987 Sage Weil)
• mon: fix paxos corner cases (#9301 #9053 Sage Weil)
• osd: cancel callback on blacklisted watchers (#8315 Samuel Just)
• osd: cleanly abort set-alloc-hint operations during upgrade (#9419 David Zafman)
• osd: clear rollback PG metadata on PG deletion (#9293 Samuel Just)
• osd: do not abort deep scrub if hinfo is missing (#10018 Loic Dachary)
• osd: erasure-code regression tests (Loic Dachary)
• osd: fix distro metadata reporting for SUSE (#8654 Danny Al-Gaaf)
• osd: fix full OSD checks during backfill (#9574 Samuel Just)
• osd: fix ioprio parsing (#9677 Loic Dachary)
• osd: fix journal direct-io shutdown (#9073 Mark Kirkwood, Ma Jianpeng, Somnath Roy)
• osd: fix journal dump (Ma Jianpeng)
• osd: fix occasional stall during peering or activation (Sage Weil)
• osd: fix past_interval display bug (#9752 Loic Dachary)
• osd: fix rare crash triggered by admin socket dump_ops_in_filght (#9916 Dong Lei)
• osd: fix snap trimming performance issues (#9487 #9113 Samuel Just, Sage Weil, Dan van der Ster, Florian Haas)
• osd: fix snapdir handling on cache eviction (#8629 Sage Weil)
• osd: handle map gaps in map advance code (Sage Weil)
• osd: handle undefined CRUSH results in interval check (#9718 Samuel Just)
• osd: include shard in JSON dump of ghobject (#10063 Loic Dachary)
• osd: make backfill reservation denial handling more robust (#9626 Samuel Just)
• osd: make misdirected op checks handle EC + primary affinity (#9835 Samuel Just, Sage Weil)
• osd: mount XFS with inode64 by default (Sage Weil)
• osd: other misc bugs (#9821 #9875 Samuel Just)
• rgw: add .log to default log path (#9353 Alexandre Marangone)
• rgw: clean up fcgi request context (#10194 Yehuda Sadeh)
• rgw: convet header underscores to dashes (#9206 Yehuda Sadeh)
• rgw: copy object data if copy target is in different pool (#9039 Yehuda Sadeh)
• rgw: don’t try to authenticate CORS peflight request (#8718 Robert Hubbard, Yehuda Sadeh)
• rgw: fix civetweb URL decoding (#8621 Yehuda Sadeh)
• rgw: fix hash calculation during PUT (Yehuda Sadeh)
• rgw: fix misc bugs (#9089 #9201 Yehuda Sadeh)
• rgw: fix object tail test (#9226 Sylvain Munaut, Yehuda Sadeh)
• rgw: make sysvinit script run rgw under systemd context as needed (#10125 Loic Dachary)
• rgw: separate civetweb log from rgw log (Yehuda Sadeh)
• rgw: set length for keystone token validations (#7796 Mark Kirkwood, Yehuda Sadeh)
• rgw: subuser creation fixes (#8587 Yehuda Sadeh)
• rpm: misc packaging improvements (Sandon Van Ness, Dan Mick, Erik Logthenberg, Boris Ranto)
• rpm: use standard udev rules for CentOS7/RHEL7 (#9747 Loic Dachary)

v0.80.7 Firefly

This release fixes a few critical issues with v0.80.6, particularlywith clusters running mixed versions.

We recommend that all v0.80.x Firefly users upgrade to this release.

For more detailed information, see thecompletechangelog.

Notable Changes

• osd: fix invalid memory reference in log trimming (#9731 Samuel Just)
• osd: fix use-after-free in cache tiering code (#7588 Sage Weil)
• osd: remove bad backfill assertion for mixed-version clusters (#9696 Samuel Just)

v0.80.6 Firefly

This is a major bugfix release for firefly, fixing a range of issuesin the OSD and monitor, particularly with cache tiering. There arealso important fixes in librados, with the watch/notify mechanism usedby librbd, and in radosgw.

A few pieces of new functionality of been backported, including improved‘ceph df’ output (view amount of writeable space per pool), support fornon-default cluster names when using sysvinit or systemd, and improved(and fixed) support for dmcrypt.

We recommend that all v0.80.x Firefly users upgrade to this release.

For more detailed information, see thecompletechangelog.

Notable Changes

• build: fix atomic64_t on i386 (#8969 Sage Weil)
• build: fix build on alpha (Michael Cree, Dmitry Smirnov)
• build: fix build on hppa (Dmitry Smirnov)
• build: fix yasm detection on x32 arch (Sage Weil)
• ceph-disk: fix ‘list’ function with dmcrypt (Sage Weil)
• ceph-disk: fix dmcrypt support (Alfredo Deza)
• ceph: allow non-default cluster to be specified (#8944)
• common: fix dup log messages to mon (#9080 Sage Weil)
• global: write pid file when -f is used (systemd, upstart) (Alexandre Oliva)
• librados: fix crash when read timeout is enabled (#9362 Matthias Kiefer, Sage Weil)
• librados: fix lock leaks in error paths (#9022 Pavan Rallabhandi)
• librados: fix watch resend on PG acting set change (#9220 Samuel Just)
• librados: python: fix aio_read handling with 0 (Mohammad Salehe)
• librbd: add interface to invalidate cached data (Josh Durgin)
• librbd: fix crash when using clone of flattened image (#8845 Josh Durgin)
• librbd: fix error path cleanup on open (#8912 Josh Durgin)
• librbd: fix null pointer check (Danny Al-Gaaf)
• librbd: limit dirty object count (Haomai Wang)
• mds: fix rstats for root and mdsdir (Yan, Zheng)
• mon: add ‘get’ command for new cache tier pool properties (Joao Eduardo Luis)
• mon: add ‘osd pool get-quota’ (#8523 Joao Eduardo Luis)
• mon: add cluster fingerprint (Sage Weil)
• mon: disallow nonsensical cache-mode transitions (#8155 Joao Eduardo Luis)
• mon: fix cache tier rounding error on i386 (Sage Weil)
• mon: fix occasional memory leak (#9176 Sage Weil)
• mon: fix reported latency for ‘osd perf’ (#9269 Samuel Just)
• mon: include ‘max avail’ in ‘ceph df’ output (Sage Weil, Xioaxi Chen)
• mon: persistently mark pools where scrub may find incomplete clones (#8882 Sage Weil)
• mon: preload erasure plugins (Loic Dachary)
• mon: prevent cache-specific settings on non-tier pools (#8696 Joao Eduardo Luis)
• mon: reduce log spam (Aanchal Agrawal, Sage Weil)
• mon: warn when cache pools have no hit_sets enabled (Sage Weil)
• msgr: fix trivial memory leak (Sage Weil)
• osd: automatically scrub PGs with invalid stats (#8147 Sage Weil)
• osd: avoid sharing PG metadata that is not durable (Samuel Just)
• osd: cap hit_set size (#9339 Samuel Just)
• osd: create default erasure profile if needed (#8601 Loic Dachary)
• osd: dump tid as JSON int (not string) where appropriate (Joao Eduardo Luis)
• osd: encode blacklist in deterministic order (#9211 Sage Weil)
• osd: fix behavior when cache tier has no hit_sets enabled (#8982 Sage Weil)
• osd: fix cache tier flushing of snapshots (#9054 Samuel Just)
• osd: fix cache tier op ordering when going from full to non-full (#8931 Sage Weil)
• osd: fix crash on dup recovery reservation (#8863 Sage Weil)
• osd: fix division by zero when pg_num adjusted with no OSDs (#9052 Sage Weil)
• osd: fix hint crash in experimental keyvaluestore_dev backend (Hoamai Wang)
• osd: fix leak in copyfrom cancellation (#8894 Samuel Just)
• osd: fix locking for copyfrom finish (#8889 Sage Weil)
• osd: fix long filename handling in backend (#8701 Sage Weil)
• osd: fix min_size check with backfill (#9497 Samuel Just)
• osd: fix mount/remount sync race (#9144 Sage Weil)
• osd: fix object listing + erasure code bug (Guang Yang)
• osd: fix race on reconnect to failed OSD (#8944 Greg Farnum)
• osd: fix recovery reservation deadlock (Samuel Just)
• osd: fix tiering agent arithmetic for negative values (#9082 Karan Singh)
• osd: improve shutdown order (#9218 Sage Weil)
• osd: improve subop discard logic (#9259 Samuel Just)
• osd: introduce optional sleep, io priority for scrub and snap trim (Sage Weil)
• osd: make scrub check for and remove stale erasure-coded objects (Samuel Just)
• osd: misc fixes (#9481 #9482 #9179 Sameul Just)
• osd: mix keyvaluestore_dev improvements (Haomai Wang)
• osd: only require CRUSH features for rules that are used (#8963 Sage Weil)
• osd: preload erasure plugins on startup (Loic Dachary)
• osd: prevent PGs from falling behind when consuming OSDMaps (#7576 Sage Weil)
• osd: prevent old clients from using tiered pools (#8714 Sage Weil)
• osd: set min_size on erasure pools to data chunk count (Sage Weil)
• osd: trim old erasure-coded objects more aggressively (Samuel Just)
• rados: enforce erasure code alignment (Lluis Pamies-Juarez)
• rgw: align object stripes with erasure pool alignment (#8442 Yehuda Sadeh)
• rgw: don’t send error body on HEAD for civetweb (#8539 Yehuda Sadeh)
• rgw: fix crash in CORS preflight request (Yehuda Sadeh)
• rgw: fix decoding of + in URL (#8702 Brian Rak)
• rgw: fix object removal on object create (#8972 Patrycja Szabowska, Yehuda Sadeh)
• systemd: use systemd-run when starting radosgw (JuanJose Galvez)
• sysvinit: support non-default cluster name (Alfredo Deza)

v0.80.5 Firefly

This release fixes a few important bugs in the radosgw and fixesseveral packaging and environment issues, including OSD log rotation,systemd environments, and daemon restarts on upgrade.

We recommend that all v0.80.x Firefly users upgrade, particularly if theyare using upstart, systemd, or radosgw.

Notable Changes

• ceph-dencoder: do not needlessly link to librgw, librados, etc. (Sage Weil)
• do not needlessly link binaries to leveldb (Sage Weil)
• mon: fix mon crash when no auth keys are present (#8851, Joao Eduardo Luis)
• osd: fix cleanup (and avoid occasional crash) during shutdown (#7981, Sage Weil)
• osd: fix log rotation under upstart (Sage Weil)
• rgw: fix multipart upload when object has irregular size (#8846, Yehuda Sadeh, Sylvain Munaut)
• rgw: improve bucket listing S3 compatibility (#8858, Yehuda Sadeh)
• rgw: improve delimited bucket listing (Yehuda Sadeh)
• rpm: do not restart daemons on upgrade (#8849, Alfredo Deza)

For more detailed information, see thecompletechangelog.

v0.80.4 Firefly

This Firefly point release fixes an potential data corruption problemwhen ceph-osd daemons run on top of XFS and service Firefly librbdclients. A recently added allocation hint that RBD utilizes triggersan XFS bug on some kernels (Linux 3.2, and likely others) that leadsto data corruption and deep-scrub errors (and inconsistent PGs). Thisrelease avoids the situation by disabling the allocation hint until wecan validate which kernels are affected and/or are known to be safe touse the hint on.

We recommend that all v0.80.x Firefly users urgently upgrade,especially if they are using RBD.

Notable Changes

• osd: disable XFS extsize hint by default (#8830, Samuel Just)
• rgw: fix extra data pool default name (Yehuda Sadeh)

For more detailed information, see thecompletechangelog.

v0.80.3 Firefly

This is the third Firefly point release. It includes a single fixfor a radosgw regression that was discovered in v0.80.2 right after itwas released.

We recommand that all v0.80.x Firefly users upgrade.

Notable Changes

• radosgw: fix regression in manifest decoding (#8804, Sage Weil)

For more detailed information, see thecompletechangelog.

v0.80.2 Firefly

This is the second Firefly point release. It contains a range ofimportant fixes, including several bugs in the OSD cache tiering, somecompatibility checks that affect upgrade situations, several radosgwbugs, and an irritating and unnecessary feature bit check thatprevents older clients from communicating with a cluster with anyerasure coded pools.

One someone large change in this point release is that the ceph RPMpackage is separated into a ceph and ceph-common package, similar toDebian. The ceph-common package contains just the client librarieswithout any of the server-side daemons.

We recommend that all v0.80.x Firefly users skip this release and usev0.80.3.

Notable Changes

• ceph-disk: better debug logging (Alfredo Deza)
• ceph-disk: fix preparation of OSDs with dmcrypt (#6700, Stephen F Taylor)
• ceph-disk: partprobe on prepare to fix dm-crypt (#6966, Eric Eastman)
• do not require ERASURE_CODE feature from clients (#8556, Sage Weil)
• libcephfs-java: build with older JNI headers (Greg Farnum)
• libcephfs-java: fix build with gcj-jdk (Dmitry Smirnov)
• librados: fix osd op tid for redirected ops (#7588, Samuel Just)
• librados: fix rados_pool_list buffer bounds checks (#8447, Sage Weil)
• librados: resend ops when pool overlay changes (#8305, Sage Weil)
• librbd, ceph-fuse: reduce CPU overhead for clean object check in cache (Haomai Wang)
• mon: allow deletion of cephfs pools (John Spray)
• mon: fix default pool ruleset choice (#8373, John Spray)
• mon: fix health summary for mon low disk warning (Sage Weil)
• mon: fix ‘osd pool set cache_target_full_ratio’ (Geoffrey Hartz)
• mon: fix quorum feature check (Greg Farnum)
• mon: fix request forwarding in mixed firefly+dumpling clusters 9#8727, Joao Eduardo Luis)
• mon: fix rule vs ruleset check in ‘osd pool set … crush_ruleset’ command (John Spray)
• mon: make osd ‘down’ count accurate (Sage Weil)
• mon: set ‘next commit’ in primary-affinity reply (Ilya Dryomov)
• mon: verify CRUSH features are supported by all mons (#8738, Greg Farnum)
• msgr: fix sequence negotiation during connection reset (Guang Yang)
• osd: block scrub on blocked objects (#8011, Samuel Just)
• osd: call XFS hint ioctl less often (#8241, Ilya Dryomov)
• osd: copy xattr spill out marker on clone (Haomai Wang)
• osd: fix flush of snapped objects (#8334, Samuel Just)
• osd: fix hashindex restart of merge operation (#8332, Samuel Just)
• osd: fix osdmap subscription bug causing startup hang (Greg Farnum)
• osd: fix potential null deref (#8328, Sage Weil)
• osd: fix shutdown race (#8319, Sage Weil)
• osd: handle ‘none’ in CRUSH results properly during peering (#8507, Samuel Just)
• osd: set no spill out marker on new objects (Greg Farnum)
• osd: skip op ordering debug checks on tiered pools (#8380, Sage Weil)
• rados: enforce ‘put’ alignment (Lluis Pamies-Juarez)
• rest-api: fix for ‘rx’ commands (Ailing Zhang)
• rgw: calc user manifest etag and fix check (#8169, #8436, Yehuda Sadeh)
• rgw: fetch attrs on multipart completion (#8452, Yehuda Sadeh, Sylvain Munaut)
• rgw: fix buffer overflow for long instance ids (#8608, Yehuda Sadeh)
• rgw: fix entity permission check on metadata put (#8428, Yehuda Sadeh)
• rgw: fix multipart retry race (#8269, Yehuda Sadeh)
• rpm: split ceph into ceph and ceph-common RPMs (Sandon Van Ness, Dan Mick)
• sysvinit: continue startin daemons after failure doing mount (#8554, Sage Weil)For more detailed information, see thecompletechangelog.

v0.80.1 Firefly

This first Firefly point release fixes a few bugs, the most visiblebeing a problem that prevents scrub from completing in some cases.

Notable Changes

• osd: revert incomplete scrub fix (Samuel Just)
• rgw: fix stripe calculation for manifest objects (Yehuda Sadeh)
• rgw: improve handling, memory usage for abort reads (Yehuda Sadeh)
• rgw: send Swift user manifest HTTP header (Yehuda Sadeh)
• libcephfs, ceph-fuse: expose MDS session state via admin socket (Yan, Zheng)
• osd: add simple throttle for snap trimming (Sage Weil)
• monclient: fix possible hang from ill-timed monitor connection failure (Sage Weil)
• osd: fix trimming of past HitSets (Sage Weil)
• osd: fix whiteouts for non-writeback cache modes (Sage Weil)
• osd: prevent divide by zero in tiering agent (David Zafman)
• osd: prevent busy loop when tiering agent can do no work (David Zafman)

For more detailed information, see thecompletechangelog.

v0.80 Firefly

This release will form the basis for our long-term supported releaseFirefly, v0.80.x. The big new features are support for erasure codingand cache tiering, although a broad range of other features, fixes,and improvements have been made across the code base. Highlights include:

• Erasure coding: support for a broad range of erasure codes for lowerstorage overhead and better data durability.
• Cache tiering: support for creating ‘cache pools’ that store hot,recently accessed objects with automatic demotion of colder data toa base tier. Typically the cache pool is backed by faster storagedevices like SSDs.
• Primary affinity: Ceph now has the ability to skew selection ofOSDs as the “primary” copy, which allows the read workload to becheaply skewed away from parts of the cluster without migrating anydata.
• Key/value OSD backend (experimental): An alternative storage backendfor Ceph OSD processes that puts all data in a key/value database likeleveldb. This provides better performance for workloads dominated bykey/value operations (like radosgw bucket indices).
• Standalone radosgw (experimental): The radosgw process can now runin a standalone mode without an apache (or similar) web server orfastcgi. This simplifies deployment and can improve performance.

We expect to maintain a series of stable releases based on v0.80Firefly for as much as a year. In the meantime, development of Cephcontinues with the next release, Giant, which will feature work on theCephFS distributed file system, more alternative storage backends(like RocksDB and f2fs), RDMA support, support for pyramid erasurecodes, and additional functionality in the block device (RBD) likecopy-on-read and multisite mirroring.

Upgrade Sequencing

• If your existing cluster is running a version older than v0.67Dumpling, please first upgrade to the latest Dumpling release beforeupgrading to v0.80 Firefly. Please refer to the Dumpling upgradedocumentation.

• We recommand adding the following to the [mon] section of yourceph.conf prior to upgrade:

mon warn on legacy crush tunables = false

This will prevent health warnings due to the use of legacy CRUSHplacement. Although it is possible to rebalance existing dataacross your cluster (see the upgrade notes below), we do notnormally recommend it for production environments as a large amountof data will move and there is a significant performance impact fromthe rebalancing.

• Upgrade daemons in the following order:

1. Monitors
2. OSDs
3. MDSs and/or radosgw

If the ceph-mds daemon is restarted first, it will wait until allOSDs have been upgraded before finishing its startup sequence. Ifthe ceph-mon daemons are not restarted prior to the ceph-osddaemons, they will not correctly register their new capabilitieswith the cluster and new features may not be usable until they arerestarted a second time.

• Upgrade radosgw daemons together. There is a subtle change in behaviorfor multipart uploads that prevents a multipart request that was initiatedwith a new radosgw from being completed by an old radosgw.

Upgrading from v0.79

• OSDMap’s json-formatted dump changed for keys ‘full’ and ‘nearfull’.What was previously being outputted as ‘true’ or ‘false’ strings arenow being outputted ‘true’ and ‘false’ booleans according to json syntax.
• HEALTH_WARN on ‘mon osd down out interval == 0’. Having this option setto zero on the leader acts much like having the ‘noout’ flag set. Thiswarning will only be reported if the monitor getting the ‘health’ or‘status’ request has this option set to zero.
• Monitor ‘auth’ commands now require the mon ‘x’ capability. This matchesdumpling v0.67.x and earlier, but differs from emperor v0.72.x.
• A librados WATCH operation on a non-existent object now returns ENOENT;previously it did not.
• Librados interface change: As there are no partial writes, the rados_write()and rados_append() operations now return 0 on success like rados_write_full()always has. This includes the C++ interface equivalents and AIO returnvalues for the aio variants.
• The radosgw init script (sysvinit) how requires that the ‘host = …’ line inceph.conf, if present, match the short hostname (the output of ‘hostname -s’),not the fully qualified hostname or the (occasionally non-short) output of‘hostname’. Failure to adjust this when upgrading from emperor or dumplingmay prevent the radosgw daemon from starting.

Upgrading from v0.72 Emperor

• See notes above.
• The ‘ceph -s’ or ‘ceph status’ command’s ‘num_in_osds’ field in theJSON and XML output has been changed from a string to an int.
• The recently added ‘ceph mds set allow_new_snaps’ command’s syntaxhas changed slightly; it is now ‘ceph mds set allow_new_snaps true’.The ‘unset’ command has been removed; instead, set the value to‘false’.
• The syntax for allowing snapshots is now ‘mds set allow_new_snaps’ instead of ‘mds allow_new_snaps’.
• ‘rbd ls’ on a pool which never held rbd images now exits with code0. It outputs nothing in plain format, or an empty list innon-plain format. This is consistent with the behavior for a poolwhich used to hold images, but contains none. Scripts relying onthis behavior should be updated.
• The MDS requires a new OSD operation TMAP2OMAP, added in this release. Whenupgrading, be sure to upgrade and restart the ceph-osd daemons before theceph-mds daemon. The MDS will refuse to start if any up OSDs do not supportthe new feature.
• The ‘ceph mds set_max_mds N’ command is now deprecated in favor of‘ceph mds set max_mds N’.
• The ‘osd pool create …’ syntax has changed for erasure pools.
• The default CRUSH rules and layouts are now using the ‘bobtail’tunables and defaults. Upgaded clusters using the old values willnow present with a health WARN state. This can be disabled byadding ‘mon warn on legacy crush tunables = false’ to ceph.conf andrestarting the monitors. Alternatively, you can switch to the newtunables with ‘ceph osd crush tunables firefly,’ but keep in mindthat this will involve moving a significant portion of the dataalready stored in the cluster and in a large cluster may takeseveral days to complete. We do not recommend adjusting tunables on aproduction cluster.
• We now default to the ‘bobtail’ CRUSH tunable values that are first supportedby Ceph clients in bobtail (v0.56) and Linux kernel version v3.9. If youplan to access a newly created Ceph cluster with an older kernel client, youshould use ‘ceph osd crush tunables legacy’ to switch back to the legacybehavior. Note that making that change will likely result in some datamovement in the system, so adjust the setting before populating the newcluster with data.
• We now set the HASHPSPOOL flag on newly created pools (and newclusters) by default. Support for this flag first appeared inv0.64; v0.67 Dumpling is the first major release that supports it.It is first supported by the Linux kernel version v3.9. If you planto access a newly created Ceph cluster with an older kernel orclients (e.g, librados, librbd) from a pre-dumpling Ceph release,you should add ‘osd pool default flag hashpspool = false’ to the‘[global]’ section of your ‘ceph.conf’ prior to creating yourmonitors (e.g., after ‘ceph-deploy new’ but before ‘ceph-deploy moncreate …’).
• The configuration option ‘osd pool default crush rule’ is deprecatedand replaced with ‘osd pool default crush replicated ruleset’. ‘osdpool default crush rule’ takes precedence for backward compatibilityand a deprecation warning is displayed when it is used.
• As part of fix for #6796, ‘ceph osd pool set ’ nowreceives as an integer instead of a string. This affects how‘hashpspool’ flag is set/unset: instead of ‘true’ or ‘false’, it nowmust be ‘0’ or ‘1’.
• The behavior of the CRUSH ‘indep’ choose mode has been changed. Noceph cluster should have been using this behavior unless someone hasmanually extracted a crush map, modified a CRUSH rule to replace‘firstn’ with ‘indep’, recompiled, and reinjected the new map intothe cluster. If the ‘indep’ mode is currently in use on a cluster,the rule should be modified to use ‘firstn’ instead, and theadministrator should wait until any data movement completes beforeupgrading.
• The ‘osd dump’ command now dumps pool snaps as an array instead of anobject.

Upgrading from v0.67 Dumpling

• See notes above.
• ceph-fuse and radosgw now use the same default values for the adminsocket and log file paths that the other daemons (ceph-osd,ceph-mon, etc.) do. If you run these daemons as non-root, you mayneed to adjust your ceph.conf to disable these options or to adjustthe permissions on /var/run/ceph and /var/log/ceph.
• The MDS now disallows snapshots by default as they are notconsidered stable. The command ‘ceph mds set allow_snaps’ willenable them.
• For clusters that were created before v0.44 (pre-argonaut, Spring2012) and store radosgw data, the auto-upgrade from TMAP to OMAPobjects has been disabled. Before upgrading, make sure that anybuckets created on pre-argonaut releases have been modified (e.g.,by PUTing and then DELETEing an object from each bucket). Anycluster created with argonaut (v0.48) or a later release or notusing radosgw never relied on the automatic conversion and is notaffected by this change.
• Any direct users of the ‘tmap’ portion of the librados API should beaware that the automatic tmap -> omap conversion functionality hasbeen removed.
• Most output that used K or KB (e.g., for kilobyte) now uses alower-case k to match the official SI convention. Any scripts thatparse output and check for an upper-case K will need to be modified.
• librados::Rados::pool_create_async() and librados::Rados::pool_delete_async()don’t drop a reference to the completion object on error, caller needs to takecare of that. This has never really worked correctly and we were leaking anobject
• ‘ceph osd crush set ’ no longer adds the osd to thespecified location, as that’s a job for ‘ceph osd crush add’. It willhowever continue to work just the same as long as the osd already existsin the crush map.
• The OSD now enforces that class write methods cannot both mutate anobject and return data. The rbd.assign_bid method, the loneoffender, has been removed. This breaks compatibility withpre-bobtail librbd clients by preventing them from creating newimages.
• librados now returns on commit instead of ack for synchronous calls.This is a bit safer in the case where both OSDs and the client crash, andis probably how it should have been acting from the beginning. Users areunlikely to notice but it could result in lower performance in somecircumstances. Those who care should switch to using the async interfaces,which let you specify safety semantics precisely.
• The C++ librados AioComplete::get_version() method was incorrectlyreturning an int (usually 32-bits). To avoid breaking librarycompatibility, a get_version64() method is added that returns thefull-width value. The old method is deprecated and will be removedin a future release. Users of the C++ librados API that make use ofthe get_version() method should modify their code to avoid getting avalue that is truncated from 64 to to 32 bits.

Notable changes since v0.79

• ceph-fuse, libcephfs: fix several caching bugs (Yan, Zheng)
• ceph-fuse: trim inodes in response to mds memory pressure (Yan, Zheng)
• librados: fix inconsistencies in API error values (David Zafman)
• librados: fix watch operations with cache pools (Sage Weil)
• librados: new snap rollback operation (David Zafman)
• mds: fix respawn (John Spray)
• mds: misc bugs (Yan, Zheng)
• mds: misc multi-mds fixes (Yan, Zheng)
• mds: use shared_ptr for requests (Greg Farnum)
• mon: fix peer feature checks (Sage Weil)
• mon: require ‘x’ mon caps for auth operations (Joao Luis)
• mon: shutdown when removed from mon cluster (Joao Luis)
• msgr: fix locking bug in authentication (Josh Durgin)
• osd: fix bug in journal replay/restart (Sage Weil)
• osd: many many many bug fixes with cache tiering (Samuel Just)
• osd: track omap and hit_set objects in pg stats (Samuel Just)
• osd: warn if agent cannot enable due to invalid (post-split) stats (Sage Weil)
• rados bench: track metadata for multiple runs separately (Guang Yang)
• rgw: fixed subuser modify (Yehuda Sadeh)
• rpm: fix redhat-lsb dependency (Sage Weil, Alfredo Deza)

Notable changes since v0.72 Emperor

• buffer: some zero-copy groundwork (Josh Durgin)
• build: misc improvements (Ken Dreyer)
• ceph-conf: stop creating bogus log files (Josh Durgin, Sage Weil)
• ceph-crush-location: new hook for setting CRUSH location of osd daemons on start)
• ceph-disk: avoid fd0 (Loic Dachary)
• ceph-disk: generalize path names, add tests (Loic Dachary)
• ceph-disk: misc improvements for puppet (Loic Dachary)
• ceph-disk: several bug fixes (Loic Dachary)
• ceph-fuse: fix race for sync reads (Sage Weil)
• ceph-fuse, libcephfs: fix several caching bugs (Yan, Zheng)
• ceph-fuse: trim inodes in response to mds memory pressure (Yan, Zheng)
• ceph-kvstore-tool: expanded command set and capabilities (Joao Eduardo Luis)
• ceph.spec: fix build dependency (Loic Dachary)
• common: bloom filter improvements (Sage Weil)
• common: check preexisting admin socket for active daemon before removing (Loic Dachary)
• common: fix aligned buffer allocation (Loic Dachary)
• common: fix authentication on big-endian architectures (Dan Mick)
• common: fix config variable substitution (Loic Dachary)
• common: portability changes to support libc++ (Noah Watkins)
• common: switch to unordered_map from hash_map (Noah Watkins)
• config: recursive metavariable expansion (Loic Dachary)
• crush: default to bobtail tunables (Sage Weil)
• crush: fix off-by-one error in recent refactor (Sage Weil)
• crush: many additional tests (Loic Dachary)
• crush: misc fixes, cleanups (Loic Dachary)
• crush: new rule steps to adjust retry attempts (Sage Weil)
• crush, osd: s/rep/replicated/ for less confusion (Loic Dachary)
• crush: refactor descend_once behavior; support set_choose*_tries for replicated rules (Sage Weil)
• crush: usability and test improvements (Loic Dachary)
• debian: change directory ownership between ceph and ceph-common (Sage Weil)
• debian: integrate misc fixes from downstream packaging (James Page)
• doc: big update to install docs (John Wilkins)
• doc: many many install doc improvements (John Wilkins)
• doc: many many updates (John Wilkins)
• doc: misc fixes (David Moreau Simard, Kun Huang)
• erasure-code: improve buffer alignment (Loic Dachary)
• erasure-code: rewrite region-xor using vector operations (Andreas Peters)
• init: fix startup ordering/timeout problem with OSDs (Dmitry Smirnov)
• libcephfs: fix resource leak (Zheng Yan)
• librados: add C API coverage for atomic write operations (Christian Marie)
• librados: fix inconsistencies in API error values (David Zafman)
• librados: fix throttle leak (and eventual deadlock) (Josh Durgin)
• librados: fix watch operations with cache pools (Sage Weil)
• librados: new snap rollback operation (David Zafman)
• librados, osd: new TMAP2OMAP operation (Yan, Zheng)
• librados: read directly into user buffer (Rutger ter Borg)
• librbd: fix use-after-free aio completion bug #5426 (Josh Durgin)
• librbd: localize/distribute parent reads (Sage Weil)
• librbd: skip zeroes/holes when copying sparse images (Josh Durgin)
• mailmap: affiliation updates (Loic Dachary)
• mailmap updates (Loic Dachary)
• many portability improvements (Noah Watkins)
• many unit test improvements (Loic Dachary)
• mds: always store backtrace in default pool (Yan, Zheng)
• mds: cope with MDS failure during creation (John Spray)
• mds: fix cap migration behavior (Yan, Zheng)
• mds: fix client session flushing (Yan, Zheng)
• mds: fix crash from client sleep/resume (Zheng Yan)
• mds: fix many many multi-mds bugs (Yan, Zheng)
• mds: fix readdir end check (Zheng Yan)
• mds: fix Resetter locking (Alexandre Oliva)
• mds: fix respawn (John Spray)
• mds: inline data support (Li Wang, Yunchuan Wen)
• mds: misc bugs (Yan, Zheng)
• mds: misc fixes for directory fragments (Zheng Yan)
• mds: misc fixes for larger directories (Zheng Yan)
• mds: misc fixes for multiple MDSs (Zheng Yan)
• mds: misc multi-mds fixes (Yan, Zheng)
• mds: remove .ceph directory (John Spray)
• mds: store directories in omap instead of tmap (Yan, Zheng)
• mds: update old-format backtraces opportunistically (Zheng Yan)
• mds: use shared_ptr for requests (Greg Farnum)
• misc cleanups from coverity (Xing Lin)
• misc coverity fixes, cleanups (Danny Al-Gaaf)
• misc coverity fixes (Xing Lin, Li Wang, Danny Al-Gaaf)
• misc portability fixes (Noah Watkins, Alan Somers)
• misc portability fixes (Noah Watkins, Christophe Courtaut, Alan Somers, huanjun)
• misc portability work (Noah Watkins)
• mon: add erasure profiles and improve erasure pool creation (Loic Dachary)
• mon: add ‘mon getmap EPOCH’ (Joao Eduardo Luis)
• mon: allow adjustment of cephfs max file size via ‘ceph mds set max_file_size’ (Sage Weil)
• mon: allow debug quorum_{enter,exit} commands via admin socket
• mon: ‘ceph osd pg-temp …’ and primary-temp commands (Ilya Dryomov)
• mon: change mds allow_new_snaps syntax to be more consistent (Sage Weil)
• mon: clean up initial crush rule creation (Loic Dachary)
• mon: collect misc metadata about osd (os, kernel, etc.), new ‘osd metadata’ command (Sage Weil)
• mon: do not create erasure rules by default (Sage Weil)
• mon: do not generate spurious MDSMaps in certain cases (Sage Weil)
• mon: do not use keyring if auth = none (Loic Dachary)
• mon: fix peer feature checks (Sage Weil)
• mon: fix pg_temp leaks (Joao Eduardo Luis)
• mon: fix pool count in ‘ceph -s’ output (Sage Weil)
• mon: handle more whitespace (newline, tab) in mon capabilities (Sage Weil)
• mon: improve (replicate or erasure) pool creation UX (Loic Dachary)
• mon: infrastructure to handle mixed-version mon cluster and cli/rest API (Greg Farnum)
• mon: MForward tests (Loic Dachary)
• mon: mkfs now idempotent (Loic Dachary)
• mon: only seed new osdmaps to current OSDs (Sage Weil)
• mon, osd: create erasure style crush rules (Loic Dachary, Sage Weil)
• mon: ‘osd crush show-tunables’ (Sage Weil)
• mon: ‘osd dump’ dumps pool snaps as array, not object (Dan Mick)
• mon, osd: new ‘erasure’ pool type (still not fully supported)
• mon: persist quorum features to disk (Greg Farnum)
• mon: prevent extreme changes in pool pg_num (Greg Farnum)
• mon: require ‘x’ mon caps for auth operations (Joao Luis)
• mon: shutdown when removed from mon cluster (Joao Luis)
• mon: take ‘osd pool set …’ value as an int, not string (Joao Eduardo Luis)
• mon: track osd features in OSDMap (Joao Luis, David Zafman)
• mon: trim MDSMaps (Joao Eduardo Luis)
• mon: warn if crush has non-optimal tunables (Sage Weil)
• mount.ceph: add -n for autofs support (Steve Stock)
• msgr: fix locking bug in authentication (Josh Durgin)
• msgr: fix messenger restart race (Xihui He)
• msgr: improve connection error detection between clients and monitors (Greg Farnum, Sage Weil)
• osd: add/fix CPU feature detection for jerasure (Loic Dachary)
• osd: add HitSet tracking for read ops (Sage Weil, Greg Farnum)
• osd: avoid touching leveldb for some xattrs (Haomai Wang, Sage Weil)
• osd: backfill to multiple targets (David Zafman)
• osd: backfill to osds not in acting set (David Zafman)
• osd: cache pool support for snapshots (Sage Weil)
• osd: client IO path changes for EC (Samuel Just)
• osd: default to 3x replication
• osd: do not include backfill targets in acting set (David Zafman)
• osd: enable new hashpspool layout by default (Sage Weil)
• osd: erasure plugin benchmarking tool (Loic Dachary)
• osd: fix and cleanup misc backfill issues (David Zafman)
• osd: fix bug in journal replay/restart (Sage Weil)
• osd: fix copy-get omap bug (Sage Weil)
• osd: fix linux kernel version detection (Ilya Dryomov)
• osd: fix memstore segv (Haomai Wang)
• osd: fix object_info_t encoding bug from emperor (Sam Just)
• osd: fix omap_clear operation to not zap xattrs (Sam Just, Yan, Zheng)
• osd: fix several bugs with tier infrastructure
• osd: fix throttle thread (Haomai Wang)
• osd: fix XFS detection (Greg Farnum, Sushma Gurram)
• osd: generalize scrubbing infrastructure to allow EC (David Zafman)
• osd: handle more whitespace (newline, tab) in osd capabilities (Sage Weil)
• osd: ignore num_objects_dirty on scrub for old pools (Sage Weil)
• osd: improved scrub checks on clones (Sage Weil, Sam Just)
• osd: improve locking in fd lookup cache (Samuel Just, Greg Farnum)
• osd: include more info in pg query result (Sage Weil)
• osd, librados: fix full cluster handling (Josh Durgin)
• osd: many erasure fixes (Sam Just)
• osd: many many many bug fixes with cache tiering (Samuel Just)
• osd: move to jerasure2 library (Loic Dachary)
• osd: new ‘chassis’ type in default crush hierarchy (Sage Weil)
• osd: new keyvaluestore-dev backend based on leveldb (Haomai Wang)
• osd: new OSDMap encoding (Greg Farnum)
• osd: new tests for erasure pools (David Zafman)
• osd: preliminary cache pool support (no snaps) (Greg Farnum, Sage Weil)
• osd: reduce scrub lock contention (Guang Yang)
• osd: requery unfound on stray notify (#6909) (Samuel Just)
• osd: some PGBackend infrastructure (Samuel Just)
• osd: support for new ‘memstore’ (memory-backed) backend (Sage Weil)
• osd: track erasure compatibility (David Zafman)
• osd: track omap and hit_set objects in pg stats (Samuel Just)
• osd: warn if agent cannot enable due to invalid (post-split) stats (Sage Weil)
• rados: add ‘crush location’, smart replica selection/balancing (Sage Weil)
• rados bench: track metadata for multiple runs separately (Guang Yang)
• rados: some performance optimizations (Yehuda Sadeh)
• rados tool: fix listomapvals (Josh Durgin)
• rbd: add ‘rbdmap’ init script for mapping rbd images on book (Adam Twardowski)
• rbd: add rbdmap support for upstart (Laurent Barbe)
• rbd: expose kernel rbd client options via ‘rbd map’ (Ilya Dryomov)
• rbd: fix bench-write command (Hoamai Wang)
• rbd: make ‘rbd list’ return empty list and success on empty pool (Josh Durgin)
• rbd: prevent deletion of images with watchers (Ilya Dryomov)
• rbd: support for 4096 mapped devices, up from ~250 (Ilya Dryomov)
• rest-api: do not fail when no OSDs yet exist (Dan Mick)
• rgw: add ‘status’ command to sysvinit script (David Moreau Simard)
• rgw: allow multiple frontends (Yehuda Sadeh)
• rgw: allow use of an erasure data pool (Yehuda Sadeh)
• rgw: convert bucket info to new format on demand (Yehuda Sadeh)
• rgw: fixed subuser modify (Yehuda Sadeh)
• rgw: fix error setting empty owner on ACLs (Yehuda Sadeh)
• rgw: fix fastcgi deadlock (do not return data from librados callback) (Yehuda Sadeh)
• rgw: fix many-part multipart uploads (Yehuda Sadeh)
• rgw: fix misc CORS bugs (Robin H. Johnson)
• rgw: fix object placement read op (Yehuda Sadeh)
• rgw: fix reading bucket policy (#6940)
• rgw: fix read_user_buckets ‘max’ behavior (Yehuda Sadeh)
• rgw: fix several CORS bugs (Robin H. Johnson)
• rgw: fix use-after-free when releasing completion handle (Yehuda Sadeh)
• rgw: improve swift temp URL support (Yehuda Sadeh)
• rgw: make multi-object delete idempotent (Yehuda Sadeh)
• rgw: optionally defer to bucket ACLs instead of object ACLs (Liam Monahan)
• rgw: prototype mongoose frontend (Yehuda Sadeh)
• rgw: several doc fixes (Alexandre Marangone)
• rgw: support for password (instead of admin token) for keystone authentication (Christophe Courtaut)
• rgw: switch from mongoose to civetweb (Yehuda Sadeh)
• rgw: user quotas (Yehuda Sadeh)
• rpm: fix redhat-lsb dependency (Sage Weil, Alfredo Deza)
• specfile: fix RPM build on RHEL6 (Ken Dreyer, Derek Yarnell)
• specfile: ship libdir/ceph (Key Dreyer)
• sysvinit, upstart: prevent both init systems from starting the same daemons (Josh Durgin)

Notable changes since v0.67 Dumpling

• build cleanly under clang (Christophe Courtaut)
• build: Makefile refactor (Roald J. van Loon)
• build: fix [/usr]/sbin locations (Alan Somers)
• ceph-disk: fix journal preallocation
• ceph-fuse, radosgw: enable admin socket and logging by default
• ceph-fuse: fix problem with readahead vs truncate race (Yan, Zheng)
• ceph-fuse: trim deleted inodes from cache (Yan, Zheng)
• ceph-fuse: use newer fuse api (Jianpeng Ma)
• ceph-kvstore-tool: new tool for working with leveldb (copy, crc) (Joao Luis)
• ceph-post-file: new command to easily share logs or other files with ceph devs
• ceph: improve parsing of CEPH_ARGS (Benoit Knecht)
• ceph: make -h behave when monitors are down
• ceph: parse CEPH_ARGS env variable
• common: bloom_filter improvements, cleanups
• common: cache crc32c values where possible
• common: correct SI is kB not KB (Dan Mick)
• common: fix looping on BSD (Alan Somers)
• common: migrate SharedPtrRegistry to use boost::shared_ptr<> (Loic Dachary)
• common: misc portability fixes (Noah Watkins)
• crc32c: fix optimized crc32c code (it now detects arch support properly)
• crc32c: improved intel-optimized crc32c support (~8x faster on my laptop!)
• crush: fix name caching
• doc: erasure coding design notes (Loic Dachary)
• hadoop: removed old version of shim to avoid confusing users (Noah Watkins)
• librados, mon: ability to query/ping out-of-quorum monitor status (Joao Luis)
• librados: fix async aio completion wakeup
• librados: fix installed header #includes (Dan Mick)
• librados: get_version64() method for C++ API
• librados: hello_world example (Greg Farnum)
• librados: sync calls now return on commit (instead of ack) (Greg Farnum)
• librbd python bindings: fix parent image name limit (Josh Durgin)
• librbd, ceph-fuse: avoid some sources of ceph-fuse, rbd cache stalls
• mds: avoid leaking objects when deleting truncated files (Yan, Zheng)
• mds: fix F_GETLK (Yan, Zheng)
• mds: fix LOOKUPSNAP bug
• mds: fix heap profiler commands (Joao Luis)
• mds: fix locking deadlock (David Disseldorp)
• mds: fix many bugs with stray (unlinked) inodes (Yan, Zheng)
• mds: fix many directory fragmentation bugs (Yan, Zheng)
• mds: fix mds rejoin with legacy parent backpointer xattrs (Alexandre Oliva)
• mds: fix rare restart/failure race during fs creation
• mds: fix standby-replay when we fall behind (Yan, Zheng)
• mds: fix stray directory purging (Yan, Zheng)
• mds: notify clients about deleted files (so they can release from their cache) (Yan, Zheng)
• mds: several bug fixes with clustered mds (Yan, Zheng)
• mon, osd: improve osdmap trimming logic (Samuel Just)
• mon, osd: initial CLI for configuring tiering
• mon: a few ‘ceph mon add’ races fixed (command is now idempotent) (Joao Luis)
• mon: allow (un)setting HASHPSPOOL flag on existing pools (Joao Luis)
• mon: allow cap strings with . to be unquoted
• mon: allow logging level of cluster log (/var/log/ceph/ceph.log) to be adjusted
• mon: avoid rewriting full osdmaps on restart (Joao Luis)
• mon: continue to discover peer addr info during election phase
• mon: disallow CephFS snapshots until ‘ceph mds set allow_new_snaps’ (Greg Farnum)
• mon: do not expose uncommitted state from ‘osd crush {add,set} …’ (Joao Luis)
• mon: fix ‘ceph osd crush reweight …’ (Joao Luis)
• mon: fix ‘osd crush move …’ command for buckets (Joao Luis)
• mon: fix byte counts (off by factor of 4) (Dan Mick, Joao Luis)
• mon: fix paxos corner case
• mon: kv properties for pools to support EC (Loic Dachary)
• mon: make ‘osd pool rename’ idempotent (Joao Luis)
• mon: modify ‘auth add’ semantics to make a bit more sense (Joao Luis)
• mon: new ‘osd perf’ command to dump recent performance information (Samuel Just)
• mon: new and improved ‘ceph -s’ or ‘ceph status’ command (more info, easier to read)
• mon: some auth check cleanups (Joao Luis)
• mon: track per-pool stats (Joao Luis)
• mon: warn about pools with bad pg_num
• mon: warn when mon data stores grow very large (Joao Luis)
• monc: fix small memory leak
• new wireshark patches pulled into the tree (Kevin Jones)
• objecter, librados: redirect requests based on cache tier config
• objecter: fix possible hang when cluster is unpaused (Josh Durgin)
• osd, librados: add new COPY_FROM rados operation
• osd, librados: add new COPY_GET rados operations (used by COPY_FROM)
• osd: ‘osd recover clone overlap limit’ option to limit cloning during recovery (Samuel Just)
• osd: COPY_GET on-wire encoding improvements (Greg Farnum)
• osd: add ‘osd heartbeat min healthy ratio’ configurable (was hard-coded at 33%)
• osd: add option to disable pg log debug code (which burns CPU)
• osd: allow cap strings with . to be unquoted
• osd: automatically detect proper xattr limits (David Zafman)
• osd: avoid extra copy in erasure coding reference implementation (Loic Dachary)
• osd: basic cache pool redirects (Greg Farnum)
• osd: basic whiteout, dirty flag support (not yet used)
• osd: bloom_filter encodability, fixes, cleanups (Loic Dachary, Sage Weil)
• osd: clean up and generalize copy-from code (Greg Farnum)
• osd: cls_hello OSD class example
• osd: erasure coding doc updates (Loic Dachary)
• osd: erasure coding plugin infrastructure, tests (Loic Dachary)
• osd: experiemental support for ZFS (zfsonlinux.org) (Yan, Zheng)
• osd: fix RWORDER flags
• osd: fix exponential backoff of slow request warnings (Loic Dachary)
• osd: fix handling of racing read vs write (Samuel Just)
• osd: fix version value returned by various operations (Greg Farnum)
• osd: generalized temp object infrastructure
• osd: ghobject_t infrastructure for EC (David Zafman)
• osd: improvements for compatset support and storage (David Zafman)
• osd: infrastructure to copy objects from other OSDs
• osd: instrument peering states (David Zafman)
• osd: misc copy-from improvements
• osd: opportunistic crc checking on stored data (off by default)
• osd: properly enforce RD/WR flags for rados classes
• osd: reduce blocking on backing fs (Samuel Just)
• osd: refactor recovery using PGBackend (Samuel Just)
• osd: remove old magical tmap->omap conversion
• osd: remove old pg log on upgrade (Samuel Just)
• osd: revert xattr size limit (fixes large rgw uploads)
• osd: use fdatasync(2) instead of fsync(2) to improve performance (Sam Just)
• pybind: fix blacklisting nonce (Loic Dachary)
• radosgw-agent: multi-region replication/DR
• rgw: complete in-progress requests before shutting down
• rgw: default log level is now more reasonable (Yehuda Sadeh)
• rgw: fix S3 auth with response-* query string params (Sylvain Munaut, Yehuda Sadeh)
• rgw: fix a few minor memory leaks (Yehuda Sadeh)
• rgw: fix acl group check (Yehuda Sadeh)
• rgw: fix inefficient use of std::list::size() (Yehuda Sadeh)
• rgw: fix major CPU utilization bug with internal caching (Yehuda Sadeh, Mark Nelson)
• rgw: fix ordering of write operations (preventing data loss on crash) (Yehuda Sadeh)
• rgw: fix ordering of writes for mulitpart upload (Yehuda Sadeh)
• rgw: fix various CORS bugs (Yehuda Sadeh)
• rgw: fix/improve swift COPY support (Yehuda Sadeh)
• rgw: improve help output (Christophe Courtaut)
• rgw: misc fixes to support DR (Josh Durgin, Yehuda Sadeh)
• rgw: per-bucket quota (Yehuda Sadeh)
• rgw: validate S3 tokens against keystone (Roald J. van Loon)
• rgw: wildcard support for keystone roles (Christophe Courtaut)
• rpm: fix junit dependencies (Alan Grosskurth)
• sysvinit radosgw: fix status return code (Danny Al-Gaaf)
• sysvinit rbdmap: fix error ‘service rbdmap stop’ (Laurent Barbe)
• sysvinit: add condrestart command (Dan van der Ster)
• sysvinit: fix shutdown order (mons last) (Alfredo Deza)

v0.79

This release is intended to serve as a release candidate for firefly,which will hopefully be v0.80. No changes are being made to the codebase at this point except those that fix bugs. Please test thisrelease if you intend to make use of the new erasure-coded pools orcache tiers in firefly.

This release fixes a range of bugs found in v0.78 and streamlines theuser experience when creating erasure-coded pools. There is also araft of fixes for the MDS (multi-mds, directory fragmentation, andlarge directories). The main notable new piece of functionality is asmall change to allow radosgw to use an erasure-coded pool for objectdata.

Upgrading

• Erasure pools created with v0.78 will no longer function with v0.79. Youwill need to delete the old pool and create a new one.

• A bug was fixed in the authentication handshake with big-endianarchitectures that prevent authentication between big- andlittle-endian machines in the same cluster. If you have a clusterthat consists entirely of big-endian machines, you will need toupgrade all daemons and clients and restart.

• The ‘ceph.file.layout’ and ‘ceph.dir.layout’ extended attributes areno longer included in the listxattr(2) results to prevent problems with‘cp -a’ and similar tools.

• Monitor ‘auth’ read-only commands now expect the user to have ‘rx’ caps.This is the same behavior that was present in dumpling, but in emperorand more recent development releases the ‘r’ cap was sufficient. Theaffected commands are:

ceph auth export
ceph auth get
ceph auth get-key
ceph auth print-key
ceph auth list

Notable Changes

• ceph-conf: stop creating bogus log files (Josh Durgin, Sage Weil)
• common: fix authentication on big-endian architectures (Dan Mick)
• debian: change directory ownership between ceph and ceph-common (Sage Weil)
• init: fix startup ordering/timeout problem with OSDs (Dmitry Smirnov)
• librbd: skip zeroes/holes when copying sparse images (Josh Durgin)
• mds: cope with MDS failure during creation (John Spray)
• mds: fix crash from client sleep/resume (Zheng Yan)
• mds: misc fixes for directory fragments (Zheng Yan)
• mds: misc fixes for larger directories (Zheng Yan)
• mds: misc fixes for multiple MDSs (Zheng Yan)
• mds: remove .ceph directory (John Spray)
• misc coverity fixes, cleanups (Danny Al-Gaaf)
• mon: add erasure profiles and improve erasure pool creation (Loic Dachary)
• mon: ‘ceph osd pg-temp …’ and primary-temp commands (Ilya Dryomov)
• mon: fix pool count in ‘ceph -s’ output (Sage Weil)
• msgr: improve connection error detection between clients and monitors (Greg Farnum, Sage Weil)
• osd: add/fix CPU feature detection for jerasure (Loic Dachary)
• osd: improved scrub checks on clones (Sage Weil, Sam Just)
• osd: many erasure fixes (Sam Just)
• osd: move to jerasure2 library (Loic Dachary)
• osd: new tests for erasure pools (David Zafman)
• osd: reduce scrub lock contention (Guang Yang)
• rgw: allow use of an erasure data pool (Yehuda Sadeh)

v0.78

This development release includes two key features: erasure coding andcache tiering. A huge amount of code was merged for this release andseveral additional weeks were spent stabilizing the code base, and itis now in a state where it is ready to be tested by a broader userbase.

This is not the firefly release. Firefly will be delayed for atleast another sprint so that we can get some operational experiencewith the new code and do some additional testing before committing tolong term support.

Note

Please note that while it is possible to create and testerasure coded pools in this release, the pools will not beusable when you upgrade to v0.79 as the OSDMap encoding willsubtlely change. Please do not populate your test poolswith important data that can’t be reloaded.

Upgrading

• Upgrade daemons in the following order:

1. Monitors
2. OSDs
3. MDSs and/or radosgw

If the ceph-mds daemon is restarted first, it will wait until allOSDs have been upgraded before finishing its startup sequence. Ifthe ceph-mon daemons are not restarted prior to the ceph-osddaemons, they will not correctly register their new capabilitieswith the cluster and new features may not be usable until they arerestarted a second time.

• Upgrade radosgw daemons together. There is a subtle change in behaviorfor multipart uploads that prevents a multipart request that was initiatedwith a new radosgw from being completed by an old radosgw.

• CephFS recently added support for a new ‘backtrace’ attribute onfile data objects that is used for lookup by inode number (i.e., NFSreexport and hard links), and will later be used by fsck repair.This replaces the existing anchor table mechanism that is used forhard link resolution. In order to completely phase that out, anyinode that has an outdated backtrace attribute will get updated whenthe inode itself is modified. This will result in some extra workloadafter a legacy CephFS file system is upgraded.

• The per-op return code in librados’ ObjectWriteOperation interfaceis now filled in.

• The librados cmpxattr operation now handles xattrs containing null bytes asdata rather than null-terminated strings.

• Compound operations in librados that create and then delete the same objectare now explicitly disallowed (they fail with -EINVAL).

• The default leveldb cache size for the ceph-osd daemon has beenincreased from 4 MB to 128 MB. This will increase the memoryfootprint of that process but tends to increase performance of omap(key/value) objects (used for CephFS and the radosgw). If memory in yourdeployment is tight, you can preserve the old behavio by adding:

leveldb write buffer size = 0
leveldb cache size = 0

to your ceph.conf to get back the (leveldb) defaults.

Notable Changes

• ceph-brag: new client and server tools (Sebastien Han, Babu Shanmugam)
• ceph-disk: use partx on RHEL or CentOS instead of partprobe (Alfredo Deza)
• ceph: fix combination of ‘tell’ and interactive mode (Joao Eduardo Luis)
• ceph-fuse: fix bugs with inline data and multiple MDSs (Zheng Yan)
• client: fix getcwd() to use new LOOKUPPARENT operation (Zheng Yan)
• common: fall back to json-pretty for admin socket (Loic Dachary)
• common: fix ‘config dump’ debug prefix (Danny Al-Gaaf)
• common: misc coverity fixes (Danny Al-Gaaf)
• common: throtller, shared_cache performance improvements, TrackedOp (Greg Farnum, Samuel Just)
• crush: fix JSON schema for dump (John Spray)
• crush: misc cleanups, tests (Loic Dachary)
• crush: new vary_r tunable (Sage Weil)
• crush: prevent invalid buckets of type 0 (Sage Weil)
• keyvaluestore: add perfcounters, misc bug fixes (Haomai Wang)
• keyvaluestore: portability improvements (Noah Watkins)
• libcephfs: API changes to better support NFS reexport via Ganesha (Matt Benjamin, Adam Emerson, Andrey Kuznetsov, Casey Bodley, David Zafman)
• librados: API documentation improvements (John Wilkins, Josh Durgin)
• librados: fix object enumeration bugs; allow iterator assignment (Josh Durgin)
• librados: streamline tests (Josh Durgin)
• librados: support for atomic read and omap operations for C API (Josh Durgin)
• librados: support for osd and mon command timeouts (Josh Durgin)
• librbd: pass allocation hints to OSD (Ilya Dryomov)
• logrotate: fix bug that prevented rotation for some daemons (Loic Dachary)
• mds: avoid duplicated discovers during recovery (Zheng Yan)
• mds: fix file lock owner checks (Zheng Yan)
• mds: fix LOOKUPPARENT, new LOOKUPNAME ops for reliable NFS reexport (Zheng Yan)
• mds: fix xattr handling on setxattr (Zheng Yan)
• mds: fix xattrs in getattr replies (Sage Weil)
• mds: force backtrace updates for old inodes on update (Zheng Yan)
• mds: several multi-mds and dirfrag bug fixes (Zheng Yan)
• mon: encode erasure stripe width in pool metadata (Loic Dachary)
• mon: erasure code crush rule creation (Loic Dachary)
• mon: erasure code plugin support (Loic Dachary)
• mon: fix bugs in initial post-mkfs quorum creation (Sage Weil)
• mon: fix error output to terminal during startup (Joao Eduardo Luis)
• mon: fix legacy CRUSH tunables warning (Sage Weil)
• mon: fix osd_epochs lower bound tracking for map trimming (Sage Weil)
• mon: fix OSDMap encoding features (Sage Weil, Aaron Ten Clay)
• mon: fix ‘pg dump’ JSON output (John Spray)
• mon: include dirty stats in ‘ceph df detail’ (Sage Weil)
• mon: list quorum member names in quorum order (Sage Weil)
• mon: prevent addition of non-empty cache tier (Sage Weil)
• mon: prevent deletion of CephFS pools (John Spray)
• mon: warn when cache tier approaches ‘full’ (Sage Weil)
• osd: allocation hint, with XFS support (Ilya Dryomov)
• osd: erasure coded pool support (Samuel Just)
• osd: fix bug causing slow/stalled recovery (#7706) (Samuel Just)
• osd: fix bugs in log merging (Samuel Just)
• osd: fix/clarify end-of-object handling on read (Loic Dachary)
• osd: fix impolite mon session backoff, reconnect behavior (Greg Farnum)
• osd: fix SnapContext cache id bug (Samuel Just)
• osd: increase default leveldb cache size and write buffer (Sage Weil, Dmitry Smirnov)
• osd: limit size of ‘osd bench …’ arguments (Joao Eduardo Luis)
• osdmaptool: new –test-map-pgs mode (Sage Weil, Ilya Dryomov)
• osd, mon: add primary-affinity to adjust selection of primaries (Sage Weil)
• osd: new ‘status’ admin socket command (Sage Weil)
• osd: simple tiering agent (Sage Weil)
• osd: store checksums for erasure coded object stripes (Samuel Just)
• osd: tests for objectstore backends (Haomai Wang)
• osd: various refactoring and bug fixes (Samuel Just, David Zafman)
• rados: add ‘set-alloc-hint’ command (Ilya Dryomov)
• rbd-fuse: fix enumerate_images overflow, memory leak (Ilya Dryomov)
• rbdmap: fix upstart script (Stephan Renatus)
• rgw: avoid logging system events to usage log (Yehuda Sadeh)
• rgw: fix Swift range reponse (Yehuda Sadeh)
• rgw: improve scalability for manifest objects (Yehuda Sadeh)
• rgw: misc fixes for multipart objects, policies (Yehuda Sadeh)
• rgw: support non-standard MultipartUpload command (Yehuda Sadeh)

v0.77

This is the final development release before the Firefly featurefreeze. The main items in this release include some additionalrefactoring work in the OSD IO path (include some lockingimprovements), per-user quotas for the radosgw, a switch to civetwebfrom mongoose for the prototype radosgw standalone mode, and aprototype leveldb-based backend for the OSD. The C librados API alsogot support for atomic write operations (read side transactions willappear in v0.78).

Upgrading

• The ‘ceph -s’ or ‘ceph status’ command’s ‘num_in_osds’ field in theJSON and XML output has been changed from a string to an int.
• The recently added ‘ceph mds set allow_new_snaps’ command’s syntaxhas changed slightly; it is now ‘ceph mds set allow_new_snaps true’.The ‘unset’ command has been removed; instead, set the value to‘false’.
• The syntax for allowing snapshots is now ‘mds set allow_new_snaps’ instead of ‘mds allow_new_snaps’.

Notable Changes

• osd: client IO path changes for EC (Samuel Just)
• common: portability changes to support libc++ (Noah Watkins)
• common: switch to unordered_map from hash_map (Noah Watkins)
• rgw: switch from mongoose to civetweb (Yehuda Sadeh)
• osd: improve locking in fd lookup cache (Samuel Just, Greg Farnum)
• doc: many many updates (John Wilkins)
• rgw: user quotas (Yehuda Sadeh)
• mon: persist quorum features to disk (Greg Farnum)
• mon: MForward tests (Loic Dachary)
• mds: inline data support (Li Wang, Yunchuan Wen)
• rgw: fix many-part multipart uploads (Yehuda Sadeh)
• osd: new keyvaluestore-dev backend based on leveldb (Haomai Wang)
• rbd: prevent deletion of images with watchers (Ilya Dryomov)
• osd: avoid touching leveldb for some xattrs (Haomai Wang, Sage Weil)
• mailmap: affiliation updates (Loic Dachary)
• osd: new OSDMap encoding (Greg Farnum)
• osd: generalize scrubbing infrastructure to allow EC (David Zafman)
• rgw: several doc fixes (Alexandre Marangone)
• librados: add C API coverage for atomic write operations (Christian Marie)
• rgw: improve swift temp URL support (Yehuda Sadeh)
• rest-api: do not fail when no OSDs yet exist (Dan Mick)
• common: check preexisting admin socket for active daemon before removing (Loic Dachary)
• osd: handle more whitespace (newline, tab) in osd capabilities (Sage Weil)
• mon: handle more whitespace (newline, tab) in mon capabilities (Sage Weil)
• rgw: make multi-object delete idempotent (Yehuda Sadeh)
• crush: fix off-by-one error in recent refactor (Sage Weil)
• rgw: fix read_user_buckets ‘max’ behavior (Yehuda Sadeh)
• mon: change mds allow_new_snaps syntax to be more consistent (Sage Weil)

v0.76

This release includes another batch of updates for fireflyfunctionality. Most notably, the cache pool infrastructure nowsupport snapshots, the OSD backfill functionality has been generalizedto include multiple targets (necessary for the coming erasure pools),and there were performance improvements to the erasure code plugin oncapable processors. The MDS now properly utilizes (and seamlesslymigrates to) the OSD key/value interface (aka omap) for storing directoryobjects. There continue to be many other fixes and improvements forusability and code portability across the tree.

Upgrading

• ‘rbd ls’ on a pool which never held rbd images now exits with code0. It outputs nothing in plain format, or an empty list innon-plain format. This is consistent with the behavior for a poolwhich used to hold images, but contains none. Scripts relying onthis behavior should be updated.
• The MDS requires a new OSD operation TMAP2OMAP, added in this release. Whenupgrading, be sure to upgrade and restart the ceph-osd daemons before theceph-mds daemon. The MDS will refuse to start if any up OSDs do not supportthe new feature.
• The ‘ceph mds set_max_mds N’ command is now deprecated in favor of‘ceph mds set max_mds N’.

Notable Changes

• build: misc improvements (Ken Dreyer)
• ceph-disk: generalize path names, add tests (Loic Dachary)
• ceph-disk: misc improvements for puppet (Loic Dachary)
• ceph-disk: several bug fixes (Loic Dachary)
• ceph-fuse: fix race for sync reads (Sage Weil)
• config: recursive metavariable expansion (Loic Dachary)
• crush: usability and test improvements (Loic Dachary)
• doc: misc fixes (David Moreau Simard, Kun Huang)
• erasure-code: improve buffer alignment (Loic Dachary)
• erasure-code: rewrite region-xor using vector operations (Andreas Peters)
• librados, osd: new TMAP2OMAP operation (Yan, Zheng)
• mailmap updates (Loic Dachary)
• many portability improvements (Noah Watkins)
• many unit test improvements (Loic Dachary)
• mds: always store backtrace in default pool (Yan, Zheng)
• mds: store directories in omap instead of tmap (Yan, Zheng)
• mon: allow adjustment of cephfs max file size via ‘ceph mds set max_file_size’ (Sage Weil)
• mon: do not create erasure rules by default (Sage Weil)
• mon: do not generate spurious MDSMaps in certain cases (Sage Weil)
• mon: do not use keyring if auth = none (Loic Dachary)
• mon: fix pg_temp leaks (Joao Eduardo Luis)
• osd: backfill to multiple targets (David Zafman)
• osd: cache pool support for snapshots (Sage Weil)
• osd: fix and cleanup misc backfill issues (David Zafman)
• osd: fix omap_clear operation to not zap xattrs (Sam Just, Yan, Zheng)
• osd: ignore num_objects_dirty on scrub for old pools (Sage Weil)
• osd: include more info in pg query result (Sage Weil)
• osd: track erasure compatibility (David Zafman)
• rbd: make ‘rbd list’ return empty list and success on empty pool (Josh Durgin)
• rgw: fix object placement read op (Yehuda Sadeh)
• rgw: fix several CORS bugs (Robin H. Johnson)
• specfile: fix RPM build on RHEL6 (Ken Dreyer, Derek Yarnell)
• specfile: ship libdir/ceph (Key Dreyer)

v0.75

This is a big release, with lots of infrastructure going in forfirefly. The big items include a prototype standalone frontend forradosgw (which does not require apache or fastcgi), tracking for readactivity on the osds (to inform tiering decisions), preliminary cachepool support (no snapshots yet), and lots of bug fixes and other workacross the tree to get ready for the next batch of erasure codingpatches.

For comparison, here are the diff stats for the last few versions:

v0.75 291 files changed, 82713 insertions(+), 33495 deletions(-)
v0.74 192 files changed, 17980 insertions(+), 1062 deletions(-)
v0.73 148 files changed, 4464 insertions(+), 2129 deletions(-)

Upgrading

• The ‘osd pool create …’ syntax has changed for erasure pools.
• The default CRUSH rules and layouts are now using the latest andgreatest tunables and defaults. Clusters using the old values willnow present with a health WARN state. This can be disabled byadding ‘mon warn on legacy crush tunables = false’ to ceph.conf.

Notable Changes

• common: bloom filter improvements (Sage Weil)
• common: fix config variable substitution (Loic Dachary)
• crush, osd: s/rep/replicated/ for less confusion (Loic Dachary)
• crush: refactor descend_once behavior; support set_choose*_tries for replicated rules (Sage Weil)
• librados: fix throttle leak (and eventual deadlock) (Josh Durgin)
• librados: read directly into user buffer (Rutger ter Borg)
• librbd: fix use-after-free aio completion bug #5426 (Josh Durgin)
• librbd: localize/distribute parent reads (Sage Weil)
• mds: fix Resetter locking (Alexandre Oliva)
• mds: fix cap migration behavior (Yan, Zheng)
• mds: fix client session flushing (Yan, Zheng)
• mds: fix many many multi-mds bugs (Yan, Zheng)
• misc portability work (Noah Watkins)
• mon, osd: create erasure style crush rules (Loic Dachary, Sage Weil)
• mon: ‘osd crush show-tunables’ (Sage Weil)
• mon: clean up initial crush rule creation (Loic Dachary)
• mon: improve (replicate or erasure) pool creation UX (Loic Dachary)
• mon: infrastructure to handle mixed-version mon cluster and cli/rest API (Greg Farnum)
• mon: mkfs now idempotent (Loic Dachary)
• mon: only seed new osdmaps to current OSDs (Sage Weil)
• mon: track osd features in OSDMap (Joao Luis, David Zafman)
• mon: warn if crush has non-optimal tunables (Sage Weil)
• mount.ceph: add -n for autofs support (Steve Stock)
• msgr: fix messenger restart race (Xihui He)
• osd, librados: fix full cluster handling (Josh Durgin)
• osd: add HitSet tracking for read ops (Sage Weil, Greg Farnum)
• osd: backfill to osds not in acting set (David Zafman)
• osd: enable new hashpspool layout by default (Sage Weil)
• osd: erasure plugin benchmarking tool (Loic Dachary)
• osd: fix XFS detection (Greg Farnum, Sushma Gurram)
• osd: fix copy-get omap bug (Sage Weil)
• osd: fix linux kernel version detection (Ilya Dryomov)
• osd: fix memstore segv (Haomai Wang)
• osd: fix several bugs with tier infrastructure
• osd: fix throttle thread (Haomai Wang)
• osd: preliminary cache pool support (no snaps) (Greg Farnum, Sage Weil)
• rados tool: fix listomapvals (Josh Durgin)
• rados: add ‘crush location’, smart replica selection/balancing (Sage Weil)
• rados: some performance optimizations (Yehuda Sadeh)
• rbd: add rbdmap support for upstart (Laurent Barbe)
• rbd: expose kernel rbd client options via ‘rbd map’ (Ilya Dryomov)
• rbd: fix bench-write command (Hoamai Wang)
• rbd: support for 4096 mapped devices, up from ~250 (Ilya Dryomov)
• rgw: allow multiple frontends (Yehuda Sadeh)
• rgw: convert bucket info to new format on demand (Yehuda Sadeh)
• rgw: fix misc CORS bugs (Robin H. Johnson)
• rgw: prototype mongoose frontend (Yehuda Sadeh)

v0.74

This release includes a few substantial pieces for Firefly, includinga long-overdue switch to 3x replication by default and a switch to the“new” CRUSH tunables by default (supported since bobtail). There isalso a fix for a long-standing radosgw bug (stalled GET) that hasalready been backported to emperor and dumpling.

Upgrading

• We now default to the ‘bobtail’ CRUSH tunable values that are first supportedby Ceph clients in bobtail (v0.56) and Linux kernel version v3.9. If youplan to access a newly created Ceph cluster with an older kernel client, youshould use ‘ceph osd crush tunables legacy’ to switch back to the legacybehavior. Note that making that change will likely result in some datamovement in the system, so adjust the setting before populating the newcluster with data.
• We now set the HASHPSPOOL flag on newly created pools (and newclusters) by default. Support for this flag first appeared inv0.64; v0.67 Dumpling is the first major release that supports it.It is first supported by the Linux kernel version v3.9. If you planto access a newly created Ceph cluster with an older kernel orclients (e.g, librados, librbd) from a pre-dumpling Ceph release,you should add ‘osd pool default flag hashpspool = false’ to the‘[global]’ section of your ‘ceph.conf’ prior to creating yourmonitors (e.g., after ‘ceph-deploy new’ but before ‘ceph-deploy moncreate …’).
• The configuration option ‘osd pool default crush rule’ is deprecatedand replaced with ‘osd pool default crush replicated ruleset’. ‘osdpool default crush rule’ takes precedence for backward compatibilityand a deprecation warning is displayed when it is used.

Notable Changes

• buffer: some zero-copy groundwork (Josh Durgin)
• ceph-disk: avoid fd0 (Loic Dachary)
• crush: default to bobtail tunables (Sage Weil)
• crush: many additional tests (Loic Dachary)
• crush: misc fixes, cleanups (Loic Dachary)
• crush: new rule steps to adjust retry attempts (Sage Weil)
• debian: integrate misc fixes from downstream packaging (James Page)
• doc: big update to install docs (John Wilkins)
• libcephfs: fix resource leak (Zheng Yan)
• misc coverity fixes (Xing Lin, Li Wang, Danny Al-Gaaf)
• misc portability fixes (Noah Watkins, Alan Somers)
• mon, osd: new ‘erasure’ pool type (still not fully supported)
• mon: add ‘mon getmap EPOCH’ (Joao Eduardo Luis)
• mon: collect misc metadata about osd (os, kernel, etc.), new ‘osd metadata’ command (Sage Weil)
• osd: default to 3x replication
• osd: do not include backfill targets in acting set (David Zafman)
• osd: new ‘chassis’ type in default crush hierarchy (Sage Weil)
• osd: requery unfound on stray notify (#6909) (Samuel Just)
• osd: some PGBackend infrastructure (Samuel Just)
• osd: support for new ‘memstore’ (memory-backed) backend (Sage Weil)
• rgw: fix fastcgi deadlock (do not return data from librados callback) (Yehuda Sadeh)
• rgw: fix reading bucket policy (#6940)
• rgw: fix use-after-free when releasing completion handle (Yehuda Sadeh)

v0.73

This release, the first development release after emperor, includesmany bug fixes and a few additional pieces of functionality. Thefirst batch of larger changes will be landing in the next version,v0.74.

Upgrading

• As part of fix for #6796, ‘ceph osd pool set ’ nowreceives as an integer instead of a string. This affects how‘hashpspool’ flag is set/unset: instead of ‘true’ or ‘false’, it nowmust be ‘0’ or ‘1’.
• The behavior of the CRUSH ‘indep’ choose mode has been changed. Noceph cluster should have been using this behavior unless someone hasmanually extracted a crush map, modified a CRUSH rule to replace‘firstn’ with ‘indep’, recompiled, and reinjected the new map intothe cluster. If the ‘indep’ mode is currently in use on a cluster,the rule should be modified to use ‘firstn’ instead, and theadministrator should wait until any data movement completes beforeupgrading.
• The ‘osd dump’ command now dumps pool snaps as an array instead of anobject.
• The radosgw init script (sysvinit) how requires that the ‘host = …’ line inceph.conf, if present, match the short hostname (the output of ‘hostname -s’),not the fully qualified hostname or the (occasionally non-short) output of‘hostname’. Failure to adjust this when upgrading from emperor or dumplingmay prevent the radosgw daemon from starting.

Notable Changes

• ceph-crush-location: new hook for setting CRUSH location of osd daemons on start
• ceph-kvstore-tool: expanded command set and capabilities (Joao Eduardo Luis)
• ceph.spec: fix build dependency (Loic Dachary)
• common: fix aligned buffer allocation (Loic Dachary)
• doc: many many install doc improvements (John Wilkins)
• mds: fix readdir end check (Zheng Yan)
• mds: update old-format backtraces opportunistically (Zheng Yan)
• misc cleanups from coverity (Xing Lin)
• misc portability fixes (Noah Watkins, Christophe Courtaut, Alan Somers, huanjun)
• mon: ‘osd dump’ dumps pool snaps as array, not object (Dan Mick)
• mon: allow debug quorum_{enter,exit} commands via admin socket
• mon: prevent extreme changes in pool pg_num (Greg Farnum)
• mon: take ‘osd pool set …’ value as an int, not string (Joao Eduardo Luis)
• mon: trim MDSMaps (Joao Eduardo Luis)
• osd: fix object_info_t encoding bug from emperor (Sam Just)
• rbd: add ‘rbdmap’ init script for mapping rbd images on book (Adam Twardowski)
• rgw: add ‘status’ command to sysvinit script (David Moreau Simard)
• rgw: fix error setting empty owner on ACLs (Yehuda Sadeh)
• rgw: optionally defer to bucket ACLs instead of object ACLs (Liam Monahan)
• rgw: support for password (instead of admin token) for keystone authentication (Christophe Courtaut)
• sysvinit, upstart: prevent both init systems from starting the same daemons (Josh Durgin)

v0.72.3 Emperor (pending release)

Upgrading

• Monitor ‘auth’ read-only commands now expect the user to have ‘rx’ caps.This is the same behavior that was present in dumpling, but in emperorand more recent development releases the ‘r’ cap was sufficient. Note thatthis backported security fix will break mon keys that are using the followingcommands but do not have the ‘x’ bit in the mon capability:

ceph auth export
ceph auth get
ceph auth get-key
ceph auth print-key
ceph auth list

v0.72.2 Emperor

This is the second bugfix release for the v0.72.x Emperor series. Wehave fixed a hang in radosgw, and fixed (again) a problem with monitorCLI compatiblity with mixed version monitors. (In the future thiswill no longer be a problem.)

Upgrading

• The JSON schema for the ‘osd pool set …’ command changed slightly. Pleaseavoid issuing this particular command via the CLI while there is a mix ofv0.72.1 and v0.72.2 monitor daemons running.
• As part of fix for #6796, ‘ceph osd pool set ’ nowreceives as an integer instead of a string. This affects how‘hashpspool’ flag is set/unset: instead of ‘true’ or ‘false’, it nowmust be ‘0’ or ‘1’.

Changes

• mon: ‘osd pool set …’ syntax change
• osd: added test for missing on-disk HEAD object
• osd: fix osd bench block size argument
• rgw: fix hang on large object GET
• rgw: fix rare use-after-free
• rgw: various DR bug fixes
• rgw: do not return error on empty owner when setting ACL
• sysvinit, upstart: prevent starting daemons using both init systems

For more detailed information, see thecompletechangelog.

v0.72.1 Emperor

Important Note

When you are upgrading from Dumpling to Emperor, do not run any of the“ceph osd pool set” commands while your monitors are running separate versions.Doing so could result in inadvertently changing cluster configuration settingsthat exhaust compute resources in your OSDs.

Changes

• osd: fix upgrade bug #6761
• ceph_filestore_tool: introduced tool to repair errors caused by #6761

This release addresses issue #6761. Upgrading to Emperor can causereads to begin returning ENFILE (too many open files). v0.72.1 fixesthat upgrade issue and adds a tool ceph_filestore_tool to repair osdstores affected by this bug.

To repair a cluster affected by this bug:

• Upgrade all osd machines to v0.72.1

• Install the ceph-test package on each osd machine to get ceph_filestore_tool

• Stop all osd processes

• To see all lost objects, run the following on each osd with the osd stopped andthe osd data directory mounted:

ceph_filestore_tool --list-lost-objects=true --filestore-path=<path-to-osd-filestore> --journal-path=<path-to-osd-journal>

• To fix all lost objects, run the following on each osd with theosd stopped and the osd data directory mounted:

ceph_filestore_tool --fix-lost-objects=true --list-lost-objects=true --filestore-path=<path-to-osd-filestore> --journal-path=<path-to-osd-journal>

• Once lost objects have been repaired on each osd, you can restartthe cluster.

Note, the ceph_filestore_tool performs a scan of all objects on theosd and may take some time.

v0.72 Emperor

This is the fifth major release of Ceph, the fourth since adopting a3-month development cycle. This release brings several new features,including multi-datacenter replication for the radosgw, improvedusability, and lands a lot of incremental performance and internalrefactoring work to support upcoming features in Firefly.

Important Note

When you are upgrading from Dumpling to Emperor, do not run any of the“ceph osd pool set” commands while your monitors are running separate versions.Doing so could result in inadvertently changing cluster configuration settingsthat exhaust compute resources in your OSDs.

Highlights

• common: improved crc32c performance
• librados: new example client and class code
• mds: many bug fixes and stability improvements
• mon: health warnings when pool pg_num values are not reasonable
• mon: per-pool performance stats
• osd, librados: new object copy primitives
• osd: improved interaction with backend file system to reduce latency
• osd: much internal refactoring to support ongoing erasure coding and tiering support
• rgw: bucket quotas
• rgw: improved CORS support
• rgw: performance improvements
• rgw: validate S3 tokens against Keystone

Coincident with core Ceph, the Emperor release also brings:

• radosgw-agent: support for multi-datacenter replication for disaster recovery
• tgt: improved support for iSCSI via upstream tgt

Packages for both are available on ceph.com.

Upgrade sequencing

There are no specific upgrade restrictions on the order or sequence ofupgrading from 0.67.x Dumpling. However, you cannot run any of the“ceph osd pool set” commands while your monitors are running separate versions.Doing so could result in inadvertently changing cluster configuration settingsand exhausting compute resources in your OSDs.

It is also possible to do a rolling upgrade from 0.61.x Cuttlefish,but there are ordering restrictions. (This is the same set ofrestrictions for Cuttlefish to Dumpling.)

• Upgrade ceph-common on all nodes that will use the command line ‘ceph’ utility.
• Upgrade all monitors (upgrade ceph package, restart ceph-mondaemons). This can happen one daemon or host at a time. Note thatbecause cuttlefish and dumpling monitors can’t talk to each other,all monitors should be upgraded in relatively short succession tominimize the risk that an a untimely failure will reduceavailability.
• Upgrade all osds (upgrade ceph package, restart ceph-osd daemons).This can happen one daemon or host at a time.
• Upgrade radosgw (upgrade radosgw package, restart radosgw daemons).

Upgrading from v0.71

• ceph-fuse and radosgw now use the same default values for the adminsocket and log file paths that the other daemons (ceph-osd,ceph-mon, etc.) do. If you run these daemons as non-root, you mayneed to adjust your ceph.conf to disable these options or to adjustthe permissions on /var/run/ceph and /var/log/ceph.

Upgrading from v0.67 Dumpling

• ceph-fuse and radosgw now use the same default values for the adminsocket and log file paths that the other daemons (ceph-osd,ceph-mon, etc.) do. If you run these daemons as non-root, you mayneed to adjust your ceph.conf to disable these options or to adjustthe permissions on /var/run/ceph and /var/log/ceph.
• The MDS now disallows snapshots by default as they are notconsidered stable. The command ‘ceph mds set allow_snaps’ willenable them.
• For clusters that were created before v0.44 (pre-argonaut, Spring2012) and store radosgw data, the auto-upgrade from TMAP to OMAPobjects has been disabled. Before upgrading, make sure that anybuckets created on pre-argonaut releases have been modified (e.g.,by PUTing and then DELETEing an object from each bucket). Anycluster created with argonaut (v0.48) or a later release or notusing radosgw never relied on the automatic conversion and is notaffected by this change.
• Any direct users of the ‘tmap’ portion of the librados API should beaware that the automatic tmap -> omap conversion functionality hasbeen removed.
• Most output that used K or KB (e.g., for kilobyte) now uses alower-case k to match the official SI convention. Any scripts thatparse output and check for an upper-case K will need to be modified.
• librados::Rados::pool_create_async() and librados::Rados::pool_delete_async()don’t drop a reference to the completion object on error, caller needs to takecare of that. This has never really worked correctly and we were leaking anobject
• ‘ceph osd crush set ’ no longer adds the osd to thespecified location, as that’s a job for ‘ceph osd crush add’. It willhowever continue to work just the same as long as the osd already existsin the crush map.
• The OSD now enforces that class write methods cannot both mutate anobject and return data. The rbd.assign_bid method, the loneoffender, has been removed. This breaks compatibility withpre-bobtail librbd clients by preventing them from creating newimages.
• librados now returns on commit instead of ack for synchronous calls.This is a bit safer in the case where both OSDs and the client crash, andis probably how it should have been acting from the beginning. Users areunlikely to notice but it could result in lower performance in somecircumstances. Those who care should switch to using the async interfaces,which let you specify safety semantics precisely.
• The C++ librados AioComplete::get_version() method was incorrectlyreturning an int (usually 32-bits). To avoid breaking librarycompatibility, a get_version64() method is added that returns thefull-width value. The old method is deprecated and will be removedin a future release. Users of the C++ librados API that make use ofthe get_version() method should modify their code to avoid getting avalue that is truncated from 64 to to 32 bits.

Notable Changes since v0.71

• build: fix [/usr]/sbin locations (Alan Somers)
• ceph-fuse, radosgw: enable admin socket and logging by default
• ceph: make -h behave when monitors are down
• common: cache crc32c values where possible
• common: fix looping on BSD (Alan Somers)
• librados, mon: ability to query/ping out-of-quorum monitor status (Joao Luis)
• librbd python bindings: fix parent image name limit (Josh Durgin)
• mds: avoid leaking objects when deleting truncated files (Yan, Zheng)
• mds: fix F_GETLK (Yan, Zheng)
• mds: fix many bugs with stray (unlinked) inodes (Yan, Zheng)
• mds: fix many directory fragmentation bugs (Yan, Zheng)
• mon: allow (un)setting HASHPSPOOL flag on existing pools (Joao Luis)
• mon: make ‘osd pool rename’ idempotent (Joao Luis)
• osd: COPY_GET on-wire encoding improvements (Greg Farnum)
• osd: bloom_filter encodability, fixes, cleanups (Loic Dachary, Sage Weil)
• osd: fix handling of racing read vs write (Samuel Just)
• osd: reduce blocking on backing fs (Samuel Just)
• radosgw-agent: multi-region replication/DR
• rgw: fix/improve swift COPY support (Yehuda Sadeh)
• rgw: misc fixes to support DR (Josh Durgin, Yehuda Sadeh)
• rgw: per-bucket quota (Yehuda Sadeh)
• rpm: fix junit dependencies (Alan Grosskurth)

Notable Changes since v0.67 Dumpling

• build cleanly under clang (Christophe Courtaut)
• build: Makefile refactor (Roald J. van Loon)
• build: fix [/usr]/sbin locations (Alan Somers)
• ceph-disk: fix journal preallocation
• ceph-fuse, radosgw: enable admin socket and logging by default
• ceph-fuse: fix problem with readahead vs truncate race (Yan, Zheng)
• ceph-fuse: trim deleted inodes from cache (Yan, Zheng)
• ceph-fuse: use newer fuse api (Jianpeng Ma)
• ceph-kvstore-tool: new tool for working with leveldb (copy, crc) (Joao Luis)
• ceph-post-file: new command to easily share logs or other files with ceph devs
• ceph: improve parsing of CEPH_ARGS (Benoit Knecht)
• ceph: make -h behave when monitors are down
• ceph: parse CEPH_ARGS env variable
• common: bloom_filter improvements, cleanups
• common: cache crc32c values where possible
• common: correct SI is kB not KB (Dan Mick)
• common: fix looping on BSD (Alan Somers)
• common: migrate SharedPtrRegistry to use boost::shared_ptr<> (Loic Dachary)
• common: misc portability fixes (Noah Watkins)
• crc32c: fix optimized crc32c code (it now detects arch support properly)
• crc32c: improved intel-optimized crc32c support (~8x faster on my laptop!)
• crush: fix name caching
• doc: erasure coding design notes (Loic Dachary)
• hadoop: removed old version of shim to avoid confusing users (Noah Watkins)
• librados, mon: ability to query/ping out-of-quorum monitor status (Joao Luis)
• librados: fix async aio completion wakeup
• librados: fix installed header #includes (Dan Mick)
• librados: get_version64() method for C++ API
• librados: hello_world example (Greg Farnum)
• librados: sync calls now return on commit (instead of ack) (Greg Farnum)
• librbd python bindings: fix parent image name limit (Josh Durgin)
• librbd, ceph-fuse: avoid some sources of ceph-fuse, rbd cache stalls
• mds: avoid leaking objects when deleting truncated files (Yan, Zheng)
• mds: fix F_GETLK (Yan, Zheng)
• mds: fix LOOKUPSNAP bug
• mds: fix heap profiler commands (Joao Luis)
• mds: fix locking deadlock (David Disseldorp)
• mds: fix many bugs with stray (unlinked) inodes (Yan, Zheng)
• mds: fix many directory fragmentation bugs (Yan, Zheng)
• mds: fix mds rejoin with legacy parent backpointer xattrs (Alexandre Oliva)
• mds: fix rare restart/failure race during fs creation
• mds: fix standby-replay when we fall behind (Yan, Zheng)
• mds: fix stray directory purging (Yan, Zheng)
• mds: notify clients about deleted files (so they can release from their cache) (Yan, Zheng)
• mds: several bug fixes with clustered mds (Yan, Zheng)
• mon, osd: improve osdmap trimming logic (Samuel Just)
• mon, osd: initial CLI for configuring tiering
• mon: a few ‘ceph mon add’ races fixed (command is now idempotent) (Joao Luis)
• mon: allow (un)setting HASHPSPOOL flag on existing pools (Joao Luis)
• mon: allow cap strings with . to be unquoted
• mon: allow logging level of cluster log (/var/log/ceph/ceph.log) to be adjusted
• mon: avoid rewriting full osdmaps on restart (Joao Luis)
• mon: continue to discover peer addr info during election phase
• mon: disallow CephFS snapshots until ‘ceph mds set allow_new_snaps’ (Greg Farnum)
• mon: do not expose uncommitted state from ‘osd crush {add,set} …’ (Joao Luis)
• mon: fix ‘ceph osd crush reweight …’ (Joao Luis)
• mon: fix ‘osd crush move …’ command for buckets (Joao Luis)
• mon: fix byte counts (off by factor of 4) (Dan Mick, Joao Luis)
• mon: fix paxos corner case
• mon: kv properties for pools to support EC (Loic Dachary)
• mon: make ‘osd pool rename’ idempotent (Joao Luis)
• mon: modify ‘auth add’ semantics to make a bit more sense (Joao Luis)
• mon: new ‘osd perf’ command to dump recent performance information (Samuel Just)
• mon: new and improved ‘ceph -s’ or ‘ceph status’ command (more info, easier to read)
• mon: some auth check cleanups (Joao Luis)
• mon: track per-pool stats (Joao Luis)
• mon: warn about pools with bad pg_num
• mon: warn when mon data stores grow very large (Joao Luis)
• monc: fix small memory leak
• new wireshark patches pulled into the tree (Kevin Jones)
• objecter, librados: redirect requests based on cache tier config
• objecter: fix possible hang when cluster is unpaused (Josh Durgin)
• osd, librados: add new COPY_FROM rados operation
• osd, librados: add new COPY_GET rados operations (used by COPY_FROM)
• osd: ‘osd recover clone overlap limit’ option to limit cloning during recovery (Samuel Just)
• osd: COPY_GET on-wire encoding improvements (Greg Farnum)
• osd: add ‘osd heartbeat min healthy ratio’ configurable (was hard-coded at 33%)
• osd: add option to disable pg log debug code (which burns CPU)
• osd: allow cap strings with . to be unquoted
• osd: automatically detect proper xattr limits (David Zafman)
• osd: avoid extra copy in erasure coding reference implementation (Loic Dachary)
• osd: basic cache pool redirects (Greg Farnum)
• osd: basic whiteout, dirty flag support (not yet used)
• osd: bloom_filter encodability, fixes, cleanups (Loic Dachary, Sage Weil)
• osd: clean up and generalize copy-from code (Greg Farnum)
• osd: cls_hello OSD class example
• osd: erasure coding doc updates (Loic Dachary)
• osd: erasure coding plugin infrastructure, tests (Loic Dachary)
• osd: experiemental support for ZFS (zfsonlinux.org) (Yan, Zheng)
• osd: fix RWORDER flags
• osd: fix exponential backoff of slow request warnings (Loic Dachary)
• osd: fix handling of racing read vs write (Samuel Just)
• osd: fix version value returned by various operations (Greg Farnum)
• osd: generalized temp object infrastructure
• osd: ghobject_t infrastructure for EC (David Zafman)
• osd: improvements for compatset support and storage (David Zafman)
• osd: infrastructure to copy objects from other OSDs
• osd: instrument peering states (David Zafman)
• osd: misc copy-from improvements
• osd: opportunistic crc checking on stored data (off by default)
• osd: properly enforce RD/WR flags for rados classes
• osd: reduce blocking on backing fs (Samuel Just)
• osd: refactor recovery using PGBackend (Samuel Just)
• osd: remove old magical tmap->omap conversion
• osd: remove old pg log on upgrade (Samuel Just)
• osd: revert xattr size limit (fixes large rgw uploads)
• osd: use fdatasync(2) instead of fsync(2) to improve performance (Sam Just)
• pybind: fix blacklisting nonce (Loic Dachary)
• radosgw-agent: multi-region replication/DR
• rgw: complete in-progress requests before shutting down
• rgw: default log level is now more reasonable (Yehuda Sadeh)
• rgw: fix S3 auth with response-* query string params (Sylvain Munaut, Yehuda Sadeh)
• rgw: fix a few minor memory leaks (Yehuda Sadeh)
• rgw: fix acl group check (Yehuda Sadeh)
• rgw: fix inefficient use of std::list::size() (Yehuda Sadeh)
• rgw: fix major CPU utilization bug with internal caching (Yehuda Sadeh, Mark Nelson)
• rgw: fix ordering of write operations (preventing data loss on crash) (Yehuda Sadeh)
• rgw: fix ordering of writes for mulitpart upload (Yehuda Sadeh)
• rgw: fix various CORS bugs (Yehuda Sadeh)
• rgw: fix/improve swift COPY support (Yehuda Sadeh)
• rgw: improve help output (Christophe Courtaut)
• rgw: misc fixes to support DR (Josh Durgin, Yehuda Sadeh)
• rgw: per-bucket quota (Yehuda Sadeh)
• rgw: validate S3 tokens against keystone (Roald J. van Loon)
• rgw: wildcard support for keystone roles (Christophe Courtaut)
• rpm: fix junit dependencies (Alan Grosskurth)
• sysvinit radosgw: fix status return code (Danny Al-Gaaf)
• sysvinit rbdmap: fix error ‘service rbdmap stop’ (Laurent Barbe)
• sysvinit: add condrestart command (Dan van der Ster)
• sysvinit: fix shutdown order (mons last) (Alfredo Deza)

v0.71

This development release includes a significant amount of new code andrefactoring, as well as a lot of preliminary functionality that will be neededfor erasure coding and tiering support. There are also several significantpatch sets improving this with the MDS.

Upgrading

• The MDS now disallows snapshots by default as they are notconsidered stable. The command ‘ceph mds set allow_snaps’ willenable them.
• For clusters that were created before v0.44 (pre-argonaut, Spring2012) and store radosgw data, the auto-upgrade from TMAP to OMAPobjects has been disabled. Before upgrading, make sure that anybuckets created on pre-argonaut releases have been modified (e.g.,by PUTing and then DELETEing an object from each bucket). Anycluster created with argonaut (v0.48) or a later release or notusing radosgw never relied on the automatic conversion and is notaffected by this change.
• Any direct users of the ‘tmap’ portion of the librados API should beaware that the automatic tmap -> omap conversion functionality hasbeen removed.
• Most output that used K or KB (e.g., for kilobyte) now uses alower-case k to match the official SI convention. Any scripts thatparse output and check for an upper-case K will need to be modified.

Notable Changes

• build: Makefile refactor (Roald J. van Loon)
• ceph-disk: fix journal preallocation
• ceph-fuse: trim deleted inodes from cache (Yan, Zheng)
• ceph-fuse: use newer fuse api (Jianpeng Ma)
• ceph-kvstore-tool: new tool for working with leveldb (copy, crc) (Joao Luis)
• common: bloom_filter improvements, cleanups
• common: correct SI is kB not KB (Dan Mick)
• common: misc portability fixes (Noah Watkins)
• hadoop: removed old version of shim to avoid confusing users (Noah Watkins)
• librados: fix installed header #includes (Dan Mick)
• librbd, ceph-fuse: avoid some sources of ceph-fuse, rbd cache stalls
• mds: fix LOOKUPSNAP bug
• mds: fix standby-replay when we fall behind (Yan, Zheng)
• mds: fix stray directory purging (Yan, Zheng)
• mon: disallow CephFS snapshots until ‘ceph mds set allow_new_snaps’ (Greg Farnum)
• mon, osd: improve osdmap trimming logic (Samuel Just)
• mon: kv properties for pools to support EC (Loic Dachary)
• mon: some auth check cleanups (Joao Luis)
• mon: track per-pool stats (Joao Luis)
• mon: warn about pools with bad pg_num
• osd: automatically detect proper xattr limits (David Zafman)
• osd: avoid extra copy in erasure coding reference implementation (Loic Dachary)
• osd: basic cache pool redirects (Greg Farnum)
• osd: basic whiteout, dirty flag support (not yet used)
• osd: clean up and generalize copy-from code (Greg Farnum)
• osd: erasure coding doc updates (Loic Dachary)
• osd: erasure coding plugin infrastructure, tests (Loic Dachary)
• osd: fix RWORDER flags
• osd: fix exponential backoff of slow request warnings (Loic Dachary)
• osd: generalized temp object infrastructure
• osd: ghobject_t infrastructure for EC (David Zafman)
• osd: improvements for compatset support and storage (David Zafman)
• osd: misc copy-from improvements
• osd: opportunistic crc checking on stored data (off by default)
• osd: refactor recovery using PGBackend (Samuel Just)
• osd: remove old magical tmap->omap conversion
• pybind: fix blacklisting nonce (Loic Dachary)
• rgw: default log level is now more reasonable (Yehuda Sadeh)
• rgw: fix acl group check (Yehuda Sadeh)
• sysvinit: fix shutdown order (mons last) (Alfredo Deza)