Page MenuHomePhabricator

Switchover s3 master (db1223 -> db1189)
Closed, ResolvedPublic

Description

When: During a pre-defined DBA maintenance windows

Prerequisites: https://wikitech.wikimedia.org/wiki/MariaDB/Primary_switchover

  • Team calendar invite

Affected wikis:: https://noc.wikimedia.org/conf/highlight.php?file=dblists/s3.dblist

Checklist:

NEW primary: db1189
OLD primary: db1223

  • Check configuration differences between new and old primary:
sudo pt-config-diff --defaults-file /root/.my.cnf h=db1223.eqiad.wmnet h=db1189.eqiad.wmnet

Failover prep:

  • Silence alerts on all hosts:
sudo cookbook sre.hosts.downtime --hours 1 -r "Primary switchover s3 T372393" 'A:db-section-s3'
  • Set NEW primary with weight 0
sudo dbctl instance db1189 set-weight 0
sudo dbctl config commit -m "Set db1189 with weight 0 T372393"
  • Depool NEW from any specific group (API, vslow, dump) if present.
sudo dbctl instance db1189 edit
# If some changes were made:
sudo dbctl config commit -m "Remove db1189 from API/vslow/dump T372393"
  • Topology changes, move all replicas under NEW primary
sudo db-switchover --timeout=25 --only-slave-move db1223 db1189
  • Disable puppet on both nodes
sudo cumin 'db1223* or db1189*' 'disable-puppet "primary switchover T372393"'

Failover:

  • Log the failover:
!log Starting s3 eqiad failover from db1223 to db1189 - T372393
  • Set section read-only:
sudo dbctl --scope eqiad section s3 ro "Maintenance until 06:15 UTC - T372393"
sudo dbctl --scope codfw section s3 ro "Maintenance until 06:15 UTC - T372393"
sudo dbctl config commit -m "Set s3 eqiad as read-only for maintenance - T372393"
  • Check s3 is indeed read-only
  • Switch primaries:
sudo db-switchover --skip-slave-move db1223 db1189
echo "===== db1223 (OLD)"; sudo db-mysql db1223 -e 'show slave status\G'
echo "===== db1189 (NEW)"; sudo db-mysql db1189 -e 'show slave status\G'
  • Promote NEW primary in dbctl, and remove read-only
sudo dbctl --scope eqiad section s3 set-master db1189
sudo dbctl --scope eqiad section s3 rw
sudo dbctl --scope codfw section s3 rw
sudo dbctl config commit -m "Promote db1189 to s3 primary and set section read-write T372393"
  • Restart puppet on both hosts:
sudo cumin 'db1223* or db1189*' 'run-puppet-agent -e "primary switchover T372393"'

Clean up tasks:

  • Clean up heartbeat table(s).
sudo db-mysql db1189 heartbeat -e "delete from heartbeat where file like 'db1223%';"
  • change events for query killer:
curl -sS 'https://gerrit.wikimedia.org/r/plugins/gitiles/operations/software/+/refs/heads/master/dbtools/events_coredb_master.sql?format=TEXT' | base64 -d | sudo db-mysql db1189
curl -sS 'https://gerrit.wikimedia.org/r/plugins/gitiles/operations/software/+/refs/heads/master/dbtools/events_coredb_slave.sql?format=TEXT' | base64 -d | sudo db-mysql db1223
sudo dbctl instance db1223 set-candidate-master --section s3 true
sudo dbctl instance db1189 set-candidate-master --section s3 false
sudo cumin 'dborch*' 'orchestrator-client -c untag -i db1189 --tag name=candidate'
sudo cumin 'dborch*' 'orchestrator-client -c tag -i db1223 --tag name=candidate'
sudo db-mysql db1215 zarcillo -e "select * from masters where section = 's3';"
  • (If needed): Depool db1223 for maintenance.
sudo dbctl instance db1223 depool
sudo dbctl config commit -m "Depool db1223 T372393"
  • Change db1223 weight to mimic the previous weight db1189:
sudo dbctl instance db1223 edit
  • Update/resolve this ticket.

Event Timeline

Change #1062381 had a related patch set uploaded (by Gerrit maintenance bot; author: Gerrit maintenance bot):

[operations/puppet@production] mariadb: Promote db1189 to s3 master

https://gerrit.wikimedia.org/r/1062381

Change #1062382 had a related patch set uploaded (by Gerrit maintenance bot; author: Gerrit maintenance bot):

[operations/dns@master] wmnet: Update s3-master alias

https://gerrit.wikimedia.org/r/1062382

Marostegui triaged this task as Medium priority.
Marostegui moved this task from Triage to Ready on the DBA board.
Marostegui updated the task description. (Show Details)

Mentioned in SAL (#wikimedia-operations) [2024-08-15T04:49:22Z] <marostegui@cumin1002> START - Cookbook sre.hosts.downtime for 1:00:00 on 24 hosts with reason: Primary switchover s3 T372393

Mentioned in SAL (#wikimedia-operations) [2024-08-15T04:49:30Z] <marostegui@cumin1002> dbctl commit (dc=all): 'Set db1189 with weight 0 T372393', diff saved to https://phabricator.wikimedia.org/P67323 and previous config saved to /var/cache/conftool/dbconfig/20240815-044929-root.json

Mentioned in SAL (#wikimedia-operations) [2024-08-15T04:49:42Z] <marostegui@cumin1002> END (PASS) - Cookbook sre.hosts.downtime (exit_code=0) for 1:00:00 on 24 hosts with reason: Primary switchover s3 T372393

Change #1062381 merged by Marostegui:

[operations/puppet@production] mariadb: Promote db1189 to s3 master

https://gerrit.wikimedia.org/r/1062381

Mentioned in SAL (#wikimedia-operations) [2024-08-15T05:03:56Z] <marostegui> Starting s3 eqiad failover from db1223 to db1189 - T372393

Mentioned in SAL (#wikimedia-operations) [2024-08-15T05:04:11Z] <marostegui@cumin1002> dbctl commit (dc=all): 'Set s3 eqiad as read-only for maintenance - T372393', diff saved to https://phabricator.wikimedia.org/P67324 and previous config saved to /var/cache/conftool/dbconfig/20240815-050410-root.json

Mentioned in SAL (#wikimedia-operations) [2024-08-15T05:04:29Z] <marostegui@cumin1002> dbctl commit (dc=all): 'Promote db1189 to s3 primary and set section read-write T372393', diff saved to https://phabricator.wikimedia.org/P67325 and previous config saved to /var/cache/conftool/dbconfig/20240815-050428-root.json

Change #1062382 merged by Marostegui:

[operations/dns@master] wmnet: Update s3-master alias

https://gerrit.wikimedia.org/r/1062382

Mentioned in SAL (#wikimedia-operations) [2024-08-15T05:06:13Z] <marostegui@cumin1002> dbctl commit (dc=all): 'Depool db1223 T372393', diff saved to https://phabricator.wikimedia.org/P67326 and previous config saved to /var/cache/conftool/dbconfig/20240815-050613-root.json

Leaving this open because the backup source still needs to be moved when the backup finishes.

db1150:3313 moved to replicate from the new master.
This is all done