TiDB Operator 1.1.8 Release Notes
Release date: December 21, 2020
TiDB Operator version: 1.1.8
New Features
- Support arbitrary Volume and VolumeMount for
PD
,TiDB
,TiKV
,TiFlash
,Backup
andRestore
, which enables using NFS or any other kubernetes supported volume source for backup/restore workflow (#3517, @dragonly)
Improvements
Support cluster and client TLS for
tidb-lightning
andtikv-importer
helm charts (#3598, @csuzhangxc)Support setting additional ports for the TiDB service. Users can utilize this feature to implement customized services, for example, additional health check (#3599, @handlerww)
Support skipping TLS when connecting
TidbInitializer
to TiDB Server (#3564, @LinuxGit)Support tableFilters for restoring data using TiDB Lightning (#3521, @sstubbs)
Support Prometheus to scrape metrics data from multiple TiDB clusters (#3622, @mikechengwei)
ACTION REQUIRED: If
TidbMonitor
CRs are deployed, update thespec.initializer.version
tov4.0.9
after upgrading TiDB Operator to v1.1.8, or some metrics will not be shown correctly in the Grafana dashboards. Prometheus crape job names are changed from${component}
to${namespace}-${TidbCluster Name}-${component}
.component
label is added to the scrape jobs of Prometheus inTidbMonitor
(#3609, @mikechengwei)
Bug Fixes
- Fix the issue that TiDB cluster fails to deploy if
spec.tikv.storageVolumes
is configured (#3586, @mikechengwei) - Fix codecs error for non-ASCII character password in the
TidbInitializer
job (#3569, @handlerww) - Fix the issue that TiFlash Pods are misrecognized as TiKV Pods. The original issue can potentially cause TiDB Operator to scale in TiKV Pods to a number smaller than
tikv.replicas
, when there are TiFlash Pods in theTidbCluster
(#3514, @handlerww) - Fix the issue that deploying
Backup
CR withoutspec.from
configured will crashtidb-controller-manager
Pod when TLS is enabled for TiDB client (#3535, @dragonly) - Fix the issue that TiDB Lightning doesn’t log to stdout (#3617, @csuzhangxc)