Hi,

I'm trying to get icinga2 to work in a cluster configuration, but am having a lot of problems. Can someone please check if I am doing this right?

I have a zone with three endpoints. They all seem to connect to each other ok, and services without an endpoint seem to get run regularly. I've named the zone company-ops-admin and the endpoints are company-ops-admin-a, -b and -c.

I have several commands which are set to run on a specific endpoint, and these don't seem to work so well, but I can't make sense of what is going wrong.

I am connecting to just the -a host via the classic ui, and I can see all the checks which don't specify an endpoint updating regularly.

The checks which are set to run on -a all show as unknown. I can see in the "notice" logs that these checks are being run, but the state is not being saved locally.

The endpoint-specific checks which run on other hosts seem to be working fine, with the exception of one of them, which seems to work on -c but not on -b. I find this particularly baffling. I've deleted all data and restarted several times and it always does the same.

Here are some examples of the configuration. The users service seems to work just fine on -b and -c:

apply Service "users" {
        import "generic-service"
        check_command = "users"
        command_endpoint = host.name
        assign where host.vars.os == "Linux"
}

The disk space check only works on -c, however:

apply Service "disk /" {
        import "generic-service"
        check_command = "disk"
        command_endpoint = host.name
        vars.disk_partitions = "/"
        assign where host.vars.os == "Linux"
}

On -b, this check is showing "Remote Icinga instance 'wistla-ops-admin-b' is not connected.". The last check timestamp happily changes regularly. I can see the host is connected because it received heartbeat messages, and because the other checks work.

I'm guessing the fact it doesn't work on -a is to do with https://dev.icinga.org/issues/7863, which should be fixed, since I am running the latest 2.2.2.1 ubuntu packages. It looks like this is not fixed, at least in my case.

Here's my zone config:

object Zone "global-templates" {
        global = true
}
object Zone "company-ops" {
        endpoints = [
                "company-ops-admin-a",
                "company-ops-admin-b",
                "company-ops-admin-c",
        ]
}
object Endpoint "company-ops-admin-a" {
        host = "10.142.1.4"
}
object Endpoint "company-ops-admin-b" {
        host = "10.142.2.4"
}
object Endpoint "company-ops-admin-c" {
        host = "10.142.3.4"
}

Can anyone suggest any ideas? I can provide more information, but I wanted to check if there were any obvious problems with what I was doing first.

Thanks,
James
_______________________________________________
icinga-users mailing list
[email protected]
https://lists.icinga.org/mailman/listinfo/icinga-users

Reply via email to