Checks executed twice and no recovery notifications are sent about icinga2

[Al2Klimov]

Every time it happened in a small time frame (for e.g. between 8am and 9am)

Do those frames correlate with your Director deployments?

[danpicpic]

Every time it happened in a small time frame (for e.g. between 8am and 9am)

Do those frames correlate with your Director deployments?

We don't use Director. But anyway no, and it also happened other few times with less servers (even one single event at the time)

[danpicpic]

We have experienced this additional few times. We have no clue how to debug it (db query?), or how to replicate it, therefore any suggestions are appreciated.

Just to sum up, and add 2 additional new info:

• In the icingaweb history we see the checks attempts are displayed twice
• eventually there are 2 notifications that are actually sent
• no ok notification is ever sent, showing the the message This notification was not sent out to any contact. what is weird is that is always systematically like this (twice the critical/warning are sent but no ok is sent)
• we have recently realised that we had both satellites communicating with each other. We have now set the communication only from satelliteA to satelliteB as suggested by the documentation (even though it should only cost in terms of performance and not of functionality). As the problem is not constant, we have no clue to know if this could be the solution