Replacing a failed etcd member

If some etcd members fail, but you still have a quorum of etcd members, you can use the remaining etcd members and the data that they contain to add more etcd members without etcd or cluster downtime.

Removing a failed etcd node

Before you add a new etcd node, remove the failed one.

Procedure

  1. From an active etcd host, remove the failed etcd node:

    1. # etcdctl -C https://<surviving host IP>:2379 \
    2.   --ca-file=/etc/etcd/ca.crt     \
    3.   --cert-file=/etc/etcd/peer.crt     \
    4.   --key-file=/etc/etcd/peer.key cluster-health
    6. # etcdctl -C https://<surviving host IP>:2379 \
    7.   --ca-file=/etc/etcd/ca.crt     \
    8.   --cert-file=/etc/etcd/peer.crt     \
    9.   --key-file=/etc/etcd/peer.key member remove <failed member identifier>

  2. Stop the etcd service on the failed etcd member by removing the etcd pod definition:

    1. # mkdir -p /etc/origin/node/pods-stopped
    2. # mv /etc/origin/node/pods/* /etc/origin/node/pods-stopped/

  3. Remove the contents of the etcd directory:

    It is recommended to back up this directory to an off-cluster location before removing the contents. You can remove this backup after a successful restore.

    1. # rm -rf /var/lib/etcd/*

Adding an etcd member

You can add an etcd host either by using an Ansible playbook or by manual steps.

Adding a new etcd host using Ansible

Procedure

  1. In the Ansible inventory file, create a new group named [new_etcd] and add the new host. Then, add the new_etcd group as a child of the [OSEv3] group:

    1. [OSEv3:children]
    2. masters
    3. nodes
    4. etcd
    5. new_etcd (1)
    7. ... [OUTPUT ABBREVIATED] ...
    9. [etcd]
    10. master-0.example.com
    11. master-1.example.com
    12. master-2.example.com
    14. [new_etcd] (1)
    15. etcd0.example.com (1)
    1Add these lines.

  2. From the host that installed OKD and hosts the Ansible inventory file, change to the playbook directory and run the etcd scaleup playbook:

    1. $ cd /usr/share/ansible/openshift-ansible
    2. $ ansible-playbook  playbooks/openshift-etcd/scaleup.yml

  3. After the playbook runs, modify the inventory file to reflect the current status by moving the new etcd host from the [new_etcd] group to the [etcd] group:

    1. [OSEv3:children]
    2. masters
    3. nodes
    4. etcd
    5. new_etcd
    7. ... [OUTPUT ABBREVIATED] ...
    9. [etcd]
    10. master-0.example.com
    11. master-1.example.com
    12. master-2.example.com
    13. etcd0.example.com

  4. If you use Flannel, modify the flanneld service configuration on every OKD host, located at /etc/sysconfig/flanneld, to include the new etcd host:

    1. FLANNEL_ETCD_ENDPOINTS=https://master-0.example.com:2379,https://master-1.example.com:2379,https://master-2.example.com:2379,https://etcd0.example.com:2379

  5. Restart the flanneld service:

    1. # systemctl restart flanneld.service

Manually adding a new etcd host

If you do not run etcd as static pods on master nodes, you might need to add another etcd host.

Procedure

Modify the current etcd cluster

To create the etcd certificates, run the openssl command, replacing the values with those from your environment.

  1. Create some environment variables:

    1. export NEW_ETCD_HOSTNAME="*etcd0.example.com*"
    2. export NEW_ETCD_IP="192.168.55.21"
    4. export CN=$NEW_ETCD_HOSTNAME
    5. export SAN="IP:${NEW_ETCD_IP}, DNS:${NEW_ETCD_HOSTNAME}"
    6. export PREFIX="/etc/etcd/generated_certs/etcd-$CN/"
    7. export OPENSSLCFG="/etc/etcd/ca/openssl.cnf"

    The custom openssl extensions used as etcdv3_ca* include the $SAN environment variable as subjectAltName. See /etc/etcd/ca/openssl.cnf for more information.

  2. Create the directory to store the configuration and certificates:

    1. # mkdir -p ${PREFIX}

  3. Create the server certificate request and sign it: (server.csr and server.crt)

    1. # openssl req -new -config ${OPENSSLCFG} \
    2.     -keyout ${PREFIX}server.key  \
    3.     -out ${PREFIX}server.csr \
    4.     -reqexts etcd_v3_req -batch -nodes \
    5.     -subj /CN=$CN
    7. # openssl ca -name etcd_ca -config ${OPENSSLCFG} \
    8.     -out ${PREFIX}server.crt \
    9.     -in ${PREFIX}server.csr \
    10.     -extensions etcd_v3_ca_server -batch

  4. Create the peer certificate request and sign it: (peer.csr and peer.crt)

    1. # openssl req -new -config ${OPENSSLCFG} \
    2.     -keyout ${PREFIX}peer.key \
    3.     -out ${PREFIX}peer.csr \
    4.     -reqexts etcd_v3_req -batch -nodes \
    5.     -subj /CN=$CN
    7. # openssl ca -name etcd_ca -config ${OPENSSLCFG} \
    8.   -out ${PREFIX}peer.crt \
    9.   -in ${PREFIX}peer.csr \
    10.   -extensions etcd_v3_ca_peer -batch

  5. Copy the current etcd configuration and ca.crt files from the current node as examples to modify later:

    1. # cp /etc/etcd/etcd.conf ${PREFIX}
    2. # cp /etc/etcd/ca.crt ${PREFIX}

  6. While still on the surviving etcd host, add the new host to the cluster. To add additional etcd members to the cluster, you must first adjust the default localhost peer in the **peerURLs** value for the first member:

    1. Get the member ID for the first member using the member list command:

      1. # etcdctl --cert-file=/etc/etcd/peer.crt \
      2.     --key-file=/etc/etcd/peer.key \
      3.     --ca-file=/etc/etcd/ca.crt \
      4.     --peers="https://172.18.1.18:2379,https://172.18.9.202:2379,https://172.18.0.75:2379" \ (1)
      5.     member list
      1Ensure that you specify the URLs of only active etcd members in the —peers parameter value.

    2. Obtain the IP address where etcd listens for cluster peers:

      1. $ ss -l4n | grep 2380

    3. Update the value of **peerURLs** using the etcdctl member update command by passing the member ID and IP address obtained from the previous steps:

      1. # etcdctl --cert-file=/etc/etcd/peer.crt \
      2.     --key-file=/etc/etcd/peer.key \
      3.     --ca-file=/etc/etcd/ca.crt \
      4.     --peers="https://172.18.1.18:2379,https://172.18.9.202:2379,https://172.18.0.75:2379" \
      5.     member update 511b7fb6cc0001 https://172.18.1.18:2380

    4. Re-run the member list command and ensure the peer URLs no longer include localhost.

  1. Add the new host to the etcd cluster. Note that the new host is not yet configured, so the status stays as unstarted until the you configure the new host.

    You must add each member and bring it online one at a time. When you add each additional member to the cluster, you must adjust the peerURLs list for the current peers. The peerURLs list grows by one for each member added. The etcdctl member add command outputs the values that you must set in the etcd.conf file as you add each member, as described in the following instructions.

    1. # etcdctl -C https://${CURRENT_ETCD_HOST}:2379 \
    2.   --ca-file=/etc/etcd/ca.crt     \
    3.   --cert-file=/etc/etcd/peer.crt     \
    4.   --key-file=/etc/etcd/peer.key member add ${NEW_ETCD_HOSTNAME} https://${NEW_ETCD_IP}:2380 (1)
    6. Added member named 10.3.9.222 with ID 4e1db163a21d7651 to cluster
    8. ETCD_NAME="<NEW_ETCD_HOSTNAME>"
    9. ETCD_INITIAL_CLUSTER="<NEW_ETCD_HOSTNAME>=https://<NEW_HOST_IP>:2380,<CLUSTERMEMBER1_NAME>=https:/<CLUSTERMEMBER2_IP>:2380,<CLUSTERMEMBER2_NAME>=https:/<CLUSTERMEMBER2_IP>:2380,<CLUSTERMEMBER3_NAME>=https:/<CLUSTERMEMBER3_IP>:2380"
    10. ETCD_INITIAL_CLUSTER_STATE="existing"
    1In this line, 10.3.9.222 is a label for the etcd member. You can specify the host name, IP address, or a simple name.

  2. Update the sample ${PREFIX}/etcd.conf file.

    1. Replace the following values with the values generated in the previous step:

      • ETCD_NAME

      • ETCD_INITIAL_CLUSTER

      • ETCD_INITIAL_CLUSTER_STATE

  1. 2.  Modify the following variables with the new host IP from the output of the previous step. You can use `${NEW_ETCD_IP}` as the value.
  3.     ```
  4.     ETCD_LISTEN_PEER_URLS
  5.     ETCD_LISTEN_CLIENT_URLS
  6.     ETCD_INITIAL_ADVERTISE_PEER_URLS
  7.     ETCD_ADVERTISE_CLIENT_URLS
  8.     ```
  10. 3.  If you previously used the member system as an etcd node, you must overwrite the current values in the ***/etc/etcd/etcd.conf*** file.
  12. 4.  Check the file for syntax errors or missing IP addresses, otherwise the etcd service might fail:
  14.     ```
  15.     # vi ${PREFIX}/etcd.conf
  16.     ```

  1. On the node that hosts the installation files, update the [etcd] hosts group in the /etc/ansible/hosts inventory file. Remove the old etcd hosts and add the new ones.

  2. Create a tgz file that contains the certificates, the sample configuration file, and the ca and copy it to the new host:

    1. # tar -czvf /etc/etcd/generated_certs/${CN}.tgz -C ${PREFIX} .
    2. # scp /etc/etcd/generated_certs/${CN}.tgz ${CN}:/tmp/
Modify the new etcd host

  1. Install iptables-services to provide iptables utilities to open the required ports for etcd:

    1. # yum install -y iptables-services

  2. Create the OS_FIREWALL_ALLOW firewall rules to allow etcd to communicate:

    • Port 2379/tcp for clients

    • Port 2380/tcp for peer communication

      1. # systemctl enable iptables.service --now
      2. # iptables -N OS_FIREWALL_ALLOW
      3. # iptables -t filter -I INPUT -j OS_FIREWALL_ALLOW
      4. # iptables -A OS_FIREWALL_ALLOW -p tcp -m state --state NEW -m tcp --dport 2379 -j ACCEPT
      5. # iptables -A OS_FIREWALL_ALLOW -p tcp -m state --state NEW -m tcp --dport 2380 -j ACCEPT
      6. # iptables-save | tee /etc/sysconfig/iptables

      In this example, a new chain OS_FIREWALL_ALLOW is created, which is the standard naming the OKD installer uses for firewall rules.

      If the environment is hosted in an IaaS environment, modify the security groups for the instance to allow incoming traffic to those ports as well.

  1. Install etcd:

    1. # yum install -y etcd

    Ensure version etcd-2.3.7-4.el7.x86_64 or greater is installed,

  2. Ensure the etcd service is not running by removing the etcd pod definition:

    1. # mkdir -p /etc/origin/node/pods-stopped
    2. # mv /etc/origin/node/pods/* /etc/origin/node/pods-stopped/

  3. Remove any etcd configuration and data:

    1. # rm -Rf /etc/etcd/*
    2. # rm -Rf /var/lib/etcd/*

  4. Extract the certificates and configuration files:

    1. # tar xzvf /tmp/etcd0.example.com.tgz -C /etc/etcd/

  5. Start etcd on the new host:

    1. # systemctl enable etcd --now

  6. Verify that the host is part of the cluster and the current cluster health:

    • If you use the v2 etcd api, run the following command:

      1. # etcdctl --cert-file=/etc/etcd/peer.crt \
      2.           --key-file=/etc/etcd/peer.key \
      3.           --ca-file=/etc/etcd/ca.crt \
      4.           --peers="https://*master-0.example.com*:2379,\
      5.           https://*master-1.example.com*:2379,\
      6.           https://*master-2.example.com*:2379,\
      7.           https://*etcd0.example.com*:2379"\
      8.           cluster-health
      9. member 5ee217d19001 is healthy: got healthy result from https://192.168.55.12:2379
      10. member 2a529ba1840722c0 is healthy: got healthy result from https://192.168.55.8:2379
      11. member 8b8904727bf526a5 is healthy: got healthy result from https://192.168.55.21:2379
      12. member ed4f0efd277d7599 is healthy: got healthy result from https://192.168.55.13:2379
      13. cluster is healthy

    • If you use the v3 etcd api, run the following command:

      1. # ETCDCTL_API=3 etcdctl --cert="/etc/etcd/peer.crt" \
      2.           --key=/etc/etcd/peer.key \
      3.           --cacert="/etc/etcd/ca.crt" \
      4.           --endpoints="https://*master-0.example.com*:2379,\
      5.             https://*master-1.example.com*:2379,\
      6.             https://*master-2.example.com*:2379,\
      7.             https://*etcd0.example.com*:2379"\
      8.             endpoint health
      9. https://master-0.example.com:2379 is healthy: successfully committed proposal: took = 5.011358ms
      10. https://master-1.example.com:2379 is healthy: successfully committed proposal: took = 1.305173ms
      11. https://master-2.example.com:2379 is healthy: successfully committed proposal: took = 1.388772ms
      12. https://etcd0.example.com:2379 is healthy: successfully committed proposal: took = 1.498829ms
Modify each OKD master

  1. Modify the master configuration in the etcClientInfo section of the /etc/origin/master/master-config.yaml file on every master. Add the new etcd host to the list of the etcd servers OKD uses to store the data, and remove any failed etcd hosts:

    1. etcdClientInfo:
    2.   ca: master.etcd-ca.crt
    3.   certFile: master.etcd-client.crt
    4.   keyFile: master.etcd-client.key
    5.   urls:
    6.     - https://master-0.example.com:2379
    7.     - https://master-1.example.com:2379
    8.     - https://master-2.example.com:2379
    9.     - https://etcd0.example.com:2379

  2. Restart the master API service:

    • On every master:

      1. # master-restart api
      2. # master-restart controllers

      The number of etcd nodes must be odd, so you must add at least two hosts.

  1. If you use Flannel, modify the flanneld service configuration located at /etc/sysconfig/flanneld on every OKD host to include the new etcd host:

    1. FLANNEL_ETCD_ENDPOINTS=https://master-0.example.com:2379,https://master-1.example.com:2379,https://master-2.example.com:2379,https://etcd0.example.com:2379

  2. Restart the flanneld service:

    1. # systemctl restart flanneld.service