When maintenance is required, you can perform that maintenance without disrupting replication, and with minimal impact to your UBA environment and users.

Follow these steps for both primary and/ or standby systems:

1. Check replication table and logs to ensure replication is active and same cycle ids:
   1. Postgres node: psql -d caspidadb -c 'select * from replication'
   2. Splunk UBA management node:tail -f /var/log/caspida/replication/replication.log
   Note: In 20-node clusters, Postgres services run on node 2 instead of node 1.
2. Stop all Splunk UBA services on management node of Standby UBA system:
   /opt/caspida/bin/Caspida stop-all
3. Perform maintenance on the affected node.
   CAUTION: Complete this task as soon as possible, ideally in less than four hours.
4. Start all Splunk UBA services on management node of standby UBA system:
   1. For primary systems: /opt/caspida/bin/Caspida start-all
   2. For standby systems: /opt/caspida/bin/Caspida start-all --no-caspida
5. Check replication table and logs to ensure replication is still active and same cycle ids:
   1. Postgres node: psql -d caspidadb -c 'select * from replication'
   2. Splunk UBA management node: tail -f /var/log/caspida/replication/replication.log
   Note: In 20-node clusters, Postgres services run on node 2 instead of node 1.