HA - pg_auto_failover for two node active-passive

[ashwanidave]

Environment :

• Ubuntu 22.04, postgreSQL 14.18, OpenNMS Horizon 33.1.5-1,
• pg-auto-failover-cli, postgresql-auto-failover, all three nodes are SSH enabled
• keepalived (using different VIP)
• using RRD files at os
• lab setup - no data loss threat

Scenario :
Two node Active-Passive (postgresql 14 + opennms),
Third node Monitoring (multiple automatic failover/failback), no manual intervention
Floating IP (always refer active IP during failover/failback)

Architecture :
Serv1 - Active (postgresql 14, OpenNMS Horizon, pg-auto-failover-cli, postgresql-auto-failover)
Serv2 - Passive (postgresql 14, OpenNMS Horizon, pg-auto-failover-cli, postgresql-auto-failover)
Serv3 - Monitor (postgresql 14, pg-auto-failover-cli, postgresql-auto-failover)

Process :

1. enable Active-Passive
2. disable systemd of opennms for Serv2 (since active opennms will push data)
3. installed on Serv3 for monitor - pg-auto-failover-cli, postgresql-auto-failover
4. installed on Serv1, Serv2 for register - pg-auto-failover-cli, postgresql-auto-failover

Expectation :

• complete concise technical guide lines step-by-step in details (just to ensure of no missing step)
• challenges - now and then upgrade in minor versions for postgresql 14, opennms horizon which mismatch
  configuration and functionality
• real-time sync of RRD files over LAN/WAN with no delay
• any better appropriate solution for multiple failover/failback without manual intervention or this will suffice and
  serve the purpose

TIA