Am 26.07.2014 22:18 schrieb "Max Zhang" <[email protected]>:
>
> What is the proper way to get remote host check results without running
active checks for certain remote host on master node in a distributed
monitoring setup?
A "Master" does not run the check, but only receives results from check
satellites if configured that way.
A "slave" checker continues to run checks, even if the connection to the
master breaks. Once re-established, it replays the log (mysql binlog, pgsql
wals similarities). That way no historocal events are lost.
Forget about the ideas below. That's old 1.x and Nagios workaround
thinking. We've made our mind quite deeply about real distributed
architectures with different connection & checkresult routes in Icinga 2.
You can even run a webinterface on a checker, or send commands (acknowledge
problem, which gets replicated to the master for example)
And now stop asking, and keep trying & testing stuff instead.
> Do I need to use ExternalCommandListener or CheckResultReader ?
> or something like below? Do I actually need to install certain agents to
achieve the effect?
>
> object EventCommand "plugin-event-process-check-result" { import
"plugin-event-command" command = [ PluginDir + "/process_check_result",
"-H", "$host.name$", "-S", "$service.name$", "-c", LocalStateDir +
"/run/icinga2/cmd/icinga2.cmd", "-r", "0", "-o", "Event Handler triggered
in state '$service.state$' with output '$service.output$'." ] }
>
> I haven't really used passive checks or anything like that in Icinga 1.X,
please advise, I'd really appreciate the help.
>
> Thanks
>
> _______________________________________________
> icinga-users mailing list
> [email protected]
> https://lists.icinga.org/mailman/listinfo/icinga-users
>
_______________________________________________
icinga-users mailing list
[email protected]
https://lists.icinga.org/mailman/listinfo/icinga-users
