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: db1209
OLD primary: db1192
- Check configuration differences between new and old primary:
sudo pt-config-diff --defaults-file /root/.my.cnf h=db1192.eqiad.wmnet h=db1209.eqiad.wmnet
Failover prep:
- Silence alerts on all hosts:
sudo cookbook sre.hosts.downtime --hours 1 -r "Primary switchover s8 T368371" 'A:db-section-s8'
- Set NEW primary with weight 0
sudo dbctl instance db1209 set-weight 0 sudo dbctl config commit -m "Set db1209 with weight 0 T368371"
- Depool NEW from any specific group (API, vslow, dump) if present.
sudo dbctl instance db1209 edit # If some changes were made: sudo dbctl config commit -m "Remove db1209 from API/vslow/dump T368371"
- Topology changes, move all replicas under NEW primary
sudo db-switchover --timeout=25 --only-slave-move db1192 db1209
- Disable puppet on both nodes
sudo cumin 'db1192* or db1209*' 'disable-puppet "primary switchover T368371"'
- Merge gerrit puppet change to promote NEW primary: FIXME
Failover:
- Log the failover:
!log Starting s8 eqiad failover from db1192 to db1209 - T368371
- Set section read-only:
sudo dbctl --scope eqiad section s8 ro "Maintenance until 06:15 UTC - T368371" sudo dbctl --scope codfw section s8 ro "Maintenance until 06:15 UTC - T368371" sudo dbctl config commit -m "Set s8 eqiad as read-only for maintenance - T368371"
- Check s8 is indeed read-only
- Switch primaries:
sudo db-switchover --skip-slave-move db1192 db1209 echo "===== db1192 (OLD)"; sudo db-mysql db1192 -e 'show slave status\G' echo "===== db1209 (NEW)"; sudo db-mysql db1209 -e 'show slave status\G'
- Promote NEW primary in dbctl, and remove read-only
sudo dbctl --scope eqiad section s8 set-master db1209 sudo dbctl --scope eqiad section s8 rw sudo dbctl --scope codfw section s8 rw sudo dbctl config commit -m "Promote db1209 to s8 primary and set section read-write T368371"
- Restart puppet on both hosts:
sudo cumin 'db1192* or db1209*' 'run-puppet-agent -e "primary switchover T368371"'
Clean up tasks:
- Clean up heartbeat table(s).
sudo db-mysql db1209 heartbeat -e "delete from heartbeat where file like 'db1192%';"
- change events for query killer:
events_coredb_master.sql on the new primary db1209 events_coredb_slave.sql on the new slave db1192
- Update DNS: FIXME
- Update candidate primary dbctl and orchestrator notes
sudo dbctl instance db1192 set-candidate-master --section s8 true sudo dbctl instance db1209 set-candidate-master --section s8 false (dborch1001): sudo orchestrator-client -c untag -i db1209 --tag name=candidate (dborch1001): sudo orchestrator-client -c tag -i db1192 --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 db1192 for maintenance.
sudo dbctl instance db1192 depool sudo dbctl config commit -m "Depool db1192 T368371"
- Change db1192 weight to mimic the previous weight db1209:
sudo dbctl instance db1192 edit
- Apply outstanding schema changes to db1192 (if any)
- Update/resolve this ticket.