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/s8.dblist
Checklist:
NEW primary: db1193
OLD primary: db1209
- Check configuration differences between new and old primary:
sudo pt-config-diff --defaults-file /root/.my.cnf h=db1209.eqiad.wmnet h=db1193.eqiad.wmnet
Failover prep:
- Silence alerts on all hosts:
sudo cookbook sre.hosts.downtime --hours 1 -r "Primary switchover s8 T371368" 'A:db-section-s8'
- Set NEW primary with weight 0
sudo dbctl instance db1193 set-weight 0 sudo dbctl config commit -m "Set db1193 with weight 0 T371368"
- Depool NEW from any specific group (API, vslow, dump) if present.
sudo dbctl instance db1193 edit # If some changes were made: sudo dbctl config commit -m "Remove db1193 from API/vslow/dump T371368"
- Topology changes, move all replicas under NEW primary
sudo db-switchover --timeout=25 --only-slave-move db1209 db1193
- Disable puppet on both nodes
sudo cumin 'db1209* or db1193*' 'disable-puppet "primary switchover T371368"'
- Merge gerrit puppet change to promote NEW primary: https://gerrit.wikimedia.org/r/c/operations/puppet/+/1058149
Failover:
- Log the failover:
!log Starting s8 eqiad failover from db1209 to db1193 - T371368
- Set section read-only:
sudo dbctl --scope eqiad section s8 ro "Maintenance until 06:15 UTC - T371368" sudo dbctl --scope codfw section s8 ro "Maintenance until 06:15 UTC - T371368" sudo dbctl config commit -m "Set s8 eqiad as read-only for maintenance - T371368"
- Check s8 is indeed read-only
- Switch primaries:
sudo db-switchover --skip-slave-move db1209 db1193 echo "===== db1209 (OLD)"; sudo db-mysql db1209 -e 'show slave status\G' echo "===== db1193 (NEW)"; sudo db-mysql db1193 -e 'show slave status\G'
- Promote NEW primary in dbctl, and remove read-only
sudo dbctl --scope eqiad section s8 set-master db1193 sudo dbctl --scope eqiad section s8 rw sudo dbctl --scope codfw section s8 rw sudo dbctl config commit -m "Promote db1193 to s8 primary and set section read-write T371368"
- Restart puppet on both hosts:
sudo cumin 'db1209* or db1193*' 'run-puppet-agent -e "primary switchover T371368"'
Clean up tasks:
- Clean up heartbeat table(s).
sudo db-mysql db1193 heartbeat -e "delete from heartbeat where file like 'db1209%';"
- 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 db1193 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 db1209
- Update DNS: https://gerrit.wikimedia.org/r/c/operations/dns/+/1058150
- Update candidate primary dbctl and orchestrator notes
sudo dbctl instance db1209 set-candidate-master --section s8 true sudo dbctl instance db1193 set-candidate-master --section s8 false sudo cumin 'dborch*' 'orchestrator-client -c untag -i db1193 --tag name=candidate' sudo cumin 'dborch*' 'orchestrator-client -c tag -i db1209 --tag name=candidate'
- Check zarcillo was updated
- db-switchover should do this. If it fails, do it manually: https://phabricator.wikimedia.org/P13956
sudo db-mysql db1215 zarcillo -e "select * from masters where section = 's8';"
- (If needed): Depool db1209 for maintenance.
sudo dbctl instance db1209 depool sudo dbctl config commit -m "Depool db1209 T371368"
- Change db1209 weight to mimic the previous weight db1193:
sudo dbctl instance db1209 edit
- Update/resolve this ticket.