Page 1 of 2
Unhandled hosts down
Posted: Fri Mar 27, 2015 9:16 am
by litsupport.box
Hello again,
I wan't to know if i can let nagios ignore hosts that are up but seem to be down because the mod gearman worker is off.
This is the view i get when my worker is down:

I only want to see the nagios worker
this is what my host says:
Code: Select all
host check orphaned, is the mod-gearman worker on queue 'hostgroup_za_hosts' running?
I only like to know if its possible to only show the worker when it's down and not it's tree of hosts under the parent host.
kind regards,
farid
Re: Unhandled hosts down
Posted: Fri Mar 27, 2015 1:03 pm
by jdalrymple
Right out of the gate I can say it's likely possible to suppress notifications and check executions with host and service dependencies. Suppressing viewing them in your XI console is going to be more tricky (less possible).
To achieve what I did mention define a host something like gearman_dummy and make it's check_command check_gearman, then of course narrow the scope of check_gearman to the appropriate worker. All of the proper hosts (that gearman's hostgroup presumably) could have a dependency on that fake host you made.
Something similar with services if you want, but no need to make a "dummy" service per-se since gearman is already a useful service to check.
I don't expect that answers your question fully, but it's some starter information to chew on...
Re: Unhandled hosts down
Posted: Fri Mar 27, 2015 2:38 pm
by BanditBBS
Dependencies are one way to go, but you could also set that worker as the parent to those other devices. Then if it ever goes down they'd be marked as unreachable instead of down and you'd also have a network blocking outage listed on the tactical overview.
Re: Unhandled hosts down
Posted: Mon Mar 30, 2015 9:11 am
by ssax
litsupport.box, will jdalrymple's or BanditBBS's solutions work for you?
Re: Unhandled hosts down
Posted: Wed Apr 01, 2015 2:02 am
by litsupport.box
BanditBBS wrote:Dependencies are one way to go, but you could also set that worker as the parent to those other devices. Then if it ever goes down they'd be marked as unreachable instead of down and you'd also have a network blocking outage listed on the tactical overview.
What do you mean exactly? I already have my worker setup as Parent of all devices on that location. It does it's job but it reports all hosts that are down, if the parent is down i want nagios to only report the parent instead of all hosts.
Re: Unhandled hosts down
Posted: Wed Apr 01, 2015 3:20 pm
by jdalrymple
If it's a parent - proper it should report them as UNREACHABLE instead of DOWN. Is that not the behavior you're seeing?
If it's working right - just disable notifications for U.
Re: Unhandled hosts down
Posted: Wed Apr 01, 2015 3:23 pm
by BanditBBS
jdalrymple wrote:If it's a parent - proper it should report them as UNREACHABLE instead of DOWN. Is that not the behavior you're seeing?
If it's working right - just disable notifications for U.
On top of that, I think(can't verify at the moment) that because they are UNREACHABLE instead of down that the do not show up on unhandled lists.
But yeah, what JD said, if it is set as parent then they should not be listed as down, if you aren't seeing that, then something else is wrong.
Re: Unhandled hosts down
Posted: Thu Apr 02, 2015 3:28 am
by litsupport.box
I'll test this again. i will update this post later today.
Re: Unhandled hosts down
Posted: Thu Apr 02, 2015 3:21 pm
by tmcdonald
Anything new to share?
Re: Unhandled hosts down
Posted: Tue Apr 07, 2015 8:35 am
by litsupport.box
I'm cleaning up nagios a little bit, when i'm done im going to configure nagios again to have its structure back with the parents etc. I hope this fixes my problem. I'll update as soon as possible.