When: Anytime, writes will be disabled
Prerequisites: https://wikitech.wikimedia.org/wiki/MariaDB/Primary_switchover
Checklist:
NEW primary: es1037
OLD primary: es1038
- Check configuration differences between new and old primary:
sudo pt-config-diff --defaults-file /root/.my.cnf h=es1038.eqiad.wmnet h=es1037.eqiad.wmnet
- Disable writes in es6 with https://gerrit.wikimedia.org/r/c/operations/mediawiki-config/+/1041107:
(in deployment.eqiad.org):
scap backport 1041107
- Check es6 is indeed read-only
Failover prep:
- Silence alerts on all hosts:
sudo cookbook sre.hosts.downtime --hours 1 -r "Primary switchover es6 T367055" 'A:db-section-es6'
- Set NEW primary with weight 0 (and depool it from API or vslow/dump groups if it is present).
sudo dbctl instance es1037 set-weight 0 sudo dbctl config commit -m "Set es1037 with weight 0 T367055"
- Topology changes, move all replicas under NEW primary
sudo db-switchover --timeout=25 --only-slave-move es1038 es1037
- Disable puppet on both nodes
sudo cumin 'es1038* or es1037*' 'disable-puppet "primary switchover T367055"'
- Merge gerrit puppet change to promote NEW primary: https://gerrit.wikimedia.org/r/c/operations/puppet/+/1041092
Failover:
- Log the failover:
!log Starting es6 eqiad failover from es1038 to es1037 - T367055
- Switch primaries:
sudo db-switchover --skip-slave-move es1038 es1037 echo "===== es1038 (OLD)"; sudo db-mysql es1038 -e 'show slave status\G' echo "===== es1037 (NEW)"; sudo db-mysql es1037 -e 'show slave status\G'
- Promote NEW primary in dbctl, and remove read-only
sudo dbctl --scope eqiad section es6 set-master es1037 sudo dbctl config commit -m "Promote es1037 to es6 primary T367055"
- Restart puppet on both hosts:
sudo cumin 'es1038* or es1037*' 'run-puppet-agent -e "primary switchover T367055"'
Clean up tasks:
- Clean up heartbeat table(s).
sudo db-mysql es1037 heartbeat -e "delete from heartbeat where file like 'es1038%';"
- change events for query killer:
events_coredb_master.sql on the new primary es1037 events_coredb_slave.sql on the new slave es1038
- Update DNS: https://gerrit.wikimedia.org/r/c/operations/dns/+/1041093
- Update candidate primary dbctl and orchestrator notes
sudo dbctl instance es1038 set-candidate-master --section es6 true sudo dbctl instance es1037 set-candidate-master --section es6 false (dborch1001): sudo orchestrator-client -c untag -i es1037 --tag name=candidate (dborch1001): sudo orchestrator-client -c tag -i es1038 --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 = 'es6';"
- (If needed): Depool es1038 for maintenance.
sudo dbctl instance es1038 depool sudo dbctl config commit -m "Depool es1038 T367055"
- Change es1038 weight to mimic the previous weight es1037:
sudo dbctl instance es1038 edit
- Enable writes in es6 by merging and deploying revert of mediawiki config patch: https://gerrit.wikimedia.org/r/c/operations/mediawiki-config/+/1041401
(in deployment.eqiad.org)
scap backport 1041401
- Update/resolve this ticket.