When: During a pre-defined DBA maintenance windows
Affected wikis:: https://noc.wikimedia.org/conf/highlight.php?file=dblists/s4.dblist
Checklist:
NEW primary: db1160
OLD primary: db1138
- Check configuration differences between new and old primary:
$ sudo pt-config-diff --defaults-file /root/.my.cnf h=db1138.eqiad.wmnet h=db1160.eqiad.wmnet 16 config differences Variable db1138 db1160 ========================= ========================= ========================= general_log_file db1138.log db1160.log gtid_binlog_pos 171966513-171966513-28... 171966513-171966513-28... gtid_binlog_state 171966513-171966513-28... 171966513-171966513-28... gtid_current_pos 0-180359175-3368394787... 0-180359175-3368394787... gtid_domain_id 171978876 171966513 gtid_slave_pos 0-180359175-3368394787... 0-180359175-3368394787... hostname db1138 db1160 log_bin_basename /srv/sqldata/db1138-bin /srv/sqldata/db1160-bin log_bin_index /srv/sqldata/db1138-bi... /srv/sqldata/db1160-bi... pid_file /srv/sqldata/db1138.pid /srv/sqldata/db1160.pid report_host db1138.eqiad.wmnet db1160.eqiad.wmnet rpl_semi_sync_master_e... ON OFF rpl_semi_sync_slave_en... OFF ON server_id 171978876 171966513 slow_query_log_file db1138-slow.log db1160-slow.log wsrep_node_name db1138 db1160
Failover prep:
- Silence alerts on all hosts:
sudo cookbook sre.hosts.downtime --hours 1 -r "Primary switchover s4 T351184" 'A:db-section-s4'
- Set NEW primary with weight 0 (and depool it from API or vslow/dump groups if it is present).
sudo dbctl instance db1160 set-weight 0 sudo dbctl config commit -m "Set db1160 with weight 0 T351184"
- Topology changes, move all replicas under NEW primary
sudo db-switchover --timeout=25 --replicating-master --read-only-master --only-slave-move db1138 db1160
- Disable puppet on both nodes
sudo cumin 'db1138* or db1160*' 'disable-puppet "primary switchover T351184"'
- Merge gerrit puppet change to promote NEW primary: FIXME
Failover:
- Log the failover:
!log Starting s4 eqiad failover from db1138 to db1160 - T351184
- Switch primaries:
sudo db-switchover --replicating-master --read-only-master --skip-slave-move db1138 db1160 echo "===== db1138 (OLD)"; sudo db-mysql db1138 -e 'show slave status\G' echo "===== db1160 (NEW)"; sudo db-mysql db1160 -e 'show slave status\G'
- Promote NEW primary in dbctl
sudo dbctl --scope eqiad section s4 set-master db1160 sudo dbctl config commit -m "Promote db1160 to s4 primary T351184"
- Restart puppet on both hosts:
sudo cumin 'db1138* or db1160*' 'run-puppet-agent -e "primary switchover T351184"'
Clean up tasks:
- Clean up heartbeat table(s).
sudo db-mysql db1160 heartbeat -e "delete from heartbeat where file like 'db1138%';"
- change events for query killer:
events_coredb_master.sql on the new primary db1160 events_coredb_slave.sql on the new slave db1138
- Update candidate primary dbctl and orchestrator notes
sudo dbctl instance db1138 set-candidate-master --section s4 true sudo dbctl instance db1160 set-candidate-master --section s4 false (dborch1001): sudo orchestrator-client -c untag -i db1160 --tag name=candidate (dborch1001): sudo orchestrator-client -c tag -i db1138 --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 = 's4';"
- (If needed): Depool db1138 for maintenance.
sudo dbctl instance db1138 depool sudo dbctl config commit -m "Depool db1138 T351184"
- Change db1138 weight to mimic the previous weight db1160:
sudo dbctl instance db1138 edit
- Apply outstanding schema changes to db1138 (if any)
- Update/resolve this ticket.