Logic with cluster Master and Satellite

This forum was archived to /woltlab and is now in read-only mode. Please register a new account on our new community platform.

You can create a thread on the new site and link to an archived thread. This archive is available as knowledge base, safe and secured.

More details here.
  • Just want to get my logic right here as I think this is what is happening with my setup:

    One master node, primarily configured to handle all notifications, checks, web interface, etc. So as a stand alone machine that does the entire job.
    I want to help even out the load as the machine can run pretty busy (846 hosts, 10163 services) so went with a cluster setup.

    My thinking was to keep the master with it's notifications, web interface, checks, etc, but to add a satellite to help even out the load. This I think is WRONG thinking on my part.

    I setup the master:

    And it's the same on the satellite zones.conf file. Now when I start up the master, all is fine it starts doing checks, notifications, etc.
    When I start the satellite up it fails and complains:

    This is because I've not specified zones in any of my config. I want Icinga2 to do the work here and split it's resources up. This I think is my mistake and it won't do this. I either have to run them as both "master" and they will split all jobs, tasks, etc, up equally (tricky as both machines aren't equal in resources) OR I have to specify in templates/config what zone I want the tasks to be performed on (master or satellite)

    Am I correct here?

    Is there a simpler way of doing what I want, whereby still keeping the master as the primary and it can do everything, but it can sub-let jobs out to a satellite to help ease the workload.

    Thank you!

  • I just found an error in my config on the satellite which was causing that error. I've fixed that (had mismatched zones.conf files, they're now identical on the master and satellite).

    Now my Satellite appears to be sat there doing nothing. Daemon running and logs show OK but it's not taking any actions or doing anything which would hold weight with my theory that because I've not set specific hosts/services to be checked using that node it won't do anything.
    (Also if I did that, would that not be a single point of failure? i.e. I set a host/service to use the satellite, what if the satellite goes offline, does the master take over it's jobs automatically?)