When: During a pre-defined DBA maintenance windows
Affected wikis:: https://noc.wikimedia.org/conf/highlight.php?file=dblists/s7.dblist
Checklist:
NEW primary: db1136
OLD primary: db1181
- Check configuration differences between new and old primary:
sudo pt-config-diff --defaults-file /root/.my.cnf h=db1181.eqiad.wmnet h=db1136.eqiad.wmnet
Failover prep:
- Silence alerts on all hosts:
sudo cookbook sre.hosts.downtime --hours 1 -r "Primary switchover s7 T334569" 'A:db-section-s7'
- Set NEW primary with weight 0 (and depool it from API or vslow/dump groups if it is present).
sudo dbctl instance db1136 set-weight 0 sudo dbctl config commit -m "Set db1136 with weight 0 T334569"
- Topology changes, move all replicas under NEW primary
sudo db-switchover --timeout=25 --replicating-master --read-only-master --only-slave-move db1181 db1136
- Disable puppet on both nodes
sudo cumin 'db1181* or db1136*' 'disable-puppet "primary switchover T334569"'
- Merge gerrit puppet change to promote NEW primary: FIXME
Failover:
- Log the failover:
!log Starting s7 eqiad failover from db1181 to db1136 - T334569
- Switch primaries:
sudo db-switchover --replicating-master --read-only-master --skip-slave-move db1181 db1136 echo "===== db1181 (OLD)"; sudo db-mysql db1181 -e 'show slave status\G' echo "===== db1136 (NEW)"; sudo db-mysql db1136 -e 'show slave status\G'
- Promote NEW primary in dbctl
sudo dbctl --scope eqiad section s7 set-master db1136 sudo dbctl config commit -m "Promote db1136 to s7 primary T334569"
- Restart puppet on both hosts:
sudo cumin 'db1181* or db1136*' 'run-puppet-agent -e "primary switchover T334569"'
Clean up tasks:
- Clean up heartbeat table(s).
sudo db-mysql db1136 heartbeat -e "delete from heartbeat where file like 'db1181%';"
- change events for query killer:
events_coredb_master.sql on the new primary db1136 events_coredb_slave.sql on the new slave db1181
- Update candidate primary dbctl and orchestrator notes
sudo dbctl instance db1181 set-candidate-master --section s7 true sudo dbctl instance db1136 set-candidate-master --section s7 false (dborch1001): sudo orchestrator-client -c untag -i db1136 --tag name=candidate (dborch1001): sudo orchestrator-client -c tag -i db1181 --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 = 's7';"
- (If needed): Depool db1181 for maintenance.
sudo dbctl instance db1181 depool sudo dbctl config commit -m "Depool db1181 T334569"
- Change db1181 weight to mimic the previous weight db1136:
sudo dbctl instance db1181 edit
- Apply outstanding schema changes to db1181 (if any)
- Update/resolve this ticket.