Icinga Active Endpoint decision


I have 2 master HA + 2 different satellite zones.

Following the Icinga best Practice I only have the zones.d configuration on the Primary Master, which syncs it to the Secondary Master + the Satellites.

I’ve realised that The Primary Master was having sync problem, until I’ve realised that due to probably prior slow icinga reload, the active endpoint changed.
The strange part is that it switched not with the Secondary master, but a Satellite instead.

  1. Is it normal that a satellite becomes an active Endpoint?
  2. Is it normal that the Primary Master don’t sync anymore when is not the Active Endpoint? Icinga was up and running the whole time, except when the switch happened apparently.
  3. How can the Primary Master go back to be the active Endpoint without having to stop each single Icinga on the other nodes? This could be problematic as only the Master have the check, ido and notification features enabled.

EDIT: I actually see that the Active Endpoint keep on switching between the 2 satellites now, even though Icinga is up and running on each node, without any reboot/crash

Hello Daniel!

Please explain the exact symptoms.