When: During a pre-defined DBA maintenance windows
Prerequisites: https://wikitech.wikimedia.org/wiki/MariaDB/Primary_switchover
Affected wikis:: https://noc.wikimedia.org/conf/highlight.php?file=dblists/s8.dblist
Checklist:
NEW primary: db2165
OLD primary: db2161
- Check configuration differences between new and old primary:
sudo pt-config-diff --defaults-file /root/.my.cnf h=db2161.codfw.wmnet h=db2165.codfw.wmnet
Failover prep:
- Silence alerts on all hosts:
sudo cookbook sre.hosts.downtime --hours 1 -r "Primary switchover s8 T415748" 'A:db-section-s8'
- Set NEW primary with weight 0
sudo dbctl instance db2165 set-weight 0 sudo dbctl config commit -m "Set db2165 with weight 0 T415748"
- Topology changes, move all replicas under NEW primary, open orchestartor to monitor the process and check it at the end.
sudo db-switchover --timeout=25 --only-slave-move db2161 db2165
- Disable puppet on both nodes
sudo cumin 'db2161* or db2165*' 'disable-puppet "primary switchover T415748"'
- Merge gerrit puppet change to promote NEW primary: https://gerrit.wikimedia.org/r/c/operations/puppet/+/1234158
Failover:
- Log the failover:
!log Starting s8 codfw failover from db2161 to db2165 - T415748
- Set section read-only:
sudo dbctl --scope eqiad section s8 ro "Maintenance until 06:15 UTC - T415748" sudo dbctl --scope codfw section s8 ro "Maintenance until 06:15 UTC - T415748" sudo dbctl config commit -m "Set s8 codfw as read-only for maintenance - T415748"
- Check s8 is indeed read-only
- Switch primaries:
sudo db-switchover --skip-slave-move db2161 db2165 echo "===== db2161 (OLD)"; sudo db-mysql db2161 -e 'show slave status\G' echo "===== db2165 (NEW)"; sudo db-mysql db2165 -e 'show slave status\G'
- Promote NEW primary in dbctl, and remove read-only
sudo dbctl --scope codfw section s8 set-master db2165 sudo dbctl --scope eqiad section s8 rw sudo dbctl --scope codfw section s8 rw sudo dbctl config commit -m "Promote db2165 to s8 primary and set section read-write T415748"
- Clean up heartbeat table(s).
sudo db-mysql db2165 heartbeat -e "delete from heartbeat where file like 'db2161%';"
- Restart puppet on both hosts:
sudo cumin 'db2161* or db2165*' 'run-puppet-agent -e "primary switchover T415748"'
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 db2165 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 db2161
- Merge DNS change (ssh ns0.wikimedia.org): https://gerrit.wikimedia.org/r/c/operations/dns/+/1234159
sudo authdns-update
- Update candidate primary dbctl and orchestrator notes
sudo dbctl instance db2161 set-candidate-master --section s8 true sudo dbctl instance db2165 set-candidate-master --section s8 false sudo cumin 'dborch*' 'orchestrator-client -c untag -i db2165 --tag name=candidate' sudo cumin 'dborch*' 'orchestrator-client -c tag -i db2161 --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';"
- Depool db2161 for maintenance.
sudo dbctl instance db2161 depool sudo dbctl config commit -m "Depool db2161 T415748"
- Change db2161 weight to mimic the previous weight db2165:
sudo dbctl instance db2161 edit
- Update/resolve this ticket.