When: During a pre-defined DBA maintenance windows
Prerequisites: https://wikitech.wikimedia.org/wiki/MariaDB/Primary_switchover
- Team calendar invite
Affected wikis:: TO-DO
Checklist:
NEW primary: db1258
OLD primary: db1255
- Check configuration differences between new and old primary:
sudo pt-config-diff --defaults-file /root/.my.cnf h=db1255.eqiad.wmnet h=db1258.eqiad.wmnet
Failover prep:
- Silence alerts on all hosts:
sudo cookbook sre.hosts.downtime --hours 1 -r "Primary switchover x3 T399699" 'A:db-section-x3'
- Set NEW primary with weight 0
sudo dbctl instance db1258 set-weight 0 sudo dbctl config commit -m "Set db1258 with weight 0 T399699"
- Depool NEW from any specific group (API, vslow, dump) if present.
sudo dbctl instance db1258 edit # If some changes were made: sudo dbctl config commit -m "Remove db1258 from API/vslow/dump T399699"
- Topology changes, move all replicas under NEW primary
sudo db-switchover --timeout=25 --only-slave-move db1255 db1258
- Disable puppet on both nodes
sudo cumin 'db1255* or db1258*' 'disable-puppet "primary switchover T399699"'
- Merge gerrit puppet change to promote NEW primary: https://gerrit.wikimedia.org/r/c/operations/puppet/+/1170098
Failover:
- Log the failover:
!log Starting x3 eqiad failover from db1255 to db1258 - T399699
- Set section read-only:
sudo dbctl --scope eqiad section x3 ro "Maintenance until 06:15 UTC - T399699" sudo dbctl --scope codfw section x3 ro "Maintenance until 06:15 UTC - T399699" sudo dbctl config commit -m "Set x3 eqiad as read-only for maintenance - T399699"
- Check x3 is indeed read-only
- Switch primaries:
sudo db-switchover --skip-slave-move db1255 db1258 echo "===== db1255 (OLD)"; sudo db-mysql db1255 -e 'show slave status\G' echo "===== db1258 (NEW)"; sudo db-mysql db1258 -e 'show slave status\G'
- Promote NEW primary in dbctl, and remove read-only
sudo dbctl --scope eqiad section x3 set-master db1258 sudo dbctl --scope eqiad section x3 rw sudo dbctl --scope codfw section x3 rw sudo dbctl config commit -m "Promote db1258 to x3 primary and set section read-write T399699"
- Clean up heartbeat table(s).
sudo db-mysql db1258 heartbeat -e "delete from heartbeat where file like 'db1255%';"
- Restart puppet on both hosts:
sudo cumin 'db1255* or db1258*' 'run-puppet-agent -e "primary switchover T399699"'
Clean up tasks:
- 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 db1258 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 db1255
- Merge DNS change: https://gerrit.wikimedia.org/r/c/operations/dns/+/1170099
sudo authdns-update
- Update candidate primary dbctl and orchestrator notes
sudo dbctl instance db1255 set-candidate-master --section x3 true sudo dbctl instance db1258 set-candidate-master --section x3 false sudo cumin 'dborch*' 'orchestrator-client -c untag -i db1258 --tag name=candidate' sudo cumin 'dborch*' 'orchestrator-client -c tag -i db1255 --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 = 'x3';"
- (If needed): Depool db1255 for maintenance.
sudo dbctl instance db1255 depool sudo dbctl config commit -m "Depool db1255 T399699"
- Change db1255 weight to mimic the previous weight db1258 (main/api/vslow/dumps):
sudo dbctl instance db1255 edit
- Update/resolve this ticket.