Re: [Nagios-devel] Scheduled downtime not recognized after
Posted: Tue May 12, 2009 7:43 pm
Frater, Greg J wrote:
> Greetings All,
>
> I'm working through an upgrade from Nagios 1 to 3.0.6. I've had several
> situations where I've scheduled downtime for a host *after* it was
> already in Nagios as DOWN and when it comes up it alerted (even though
> it was in scheduled downtime)
Yes. This is because Nagios remembers that it has sent an alert for
the problem state, and therefore it sends the recovery alert as well.
IIRC, this was a feature requested on the list quite some time ago.
> or in my most recent case the host was up
> but a bunch of the service checks I'd added went critical because they
> use the NSClient++ agent which was not loaded yet. In this case after a
> bunch of the service checks had alerted as CRITICAL, I put the host in
> scheduled downtime (verified it by the web interface) and then I
> installed the nsclient++ agent. After it was installed I told Nagios to
> recheck all services on the host to validate the install and I got
> alerts for the recoveries of the service checks.
Same thing here, I'm afraid.
> I posted for this
> before on the users mailing list
> (http://article.gmane.org/gmane.network. ... user/61750), thinking
> there was a possible delay between the web ui and the engine when
> scheduling downtime, but now I'm wondering if it's a problem with the
> logic such that when a host or service is already in a down/critical
> state upon recovery it does not check for scheduled downtime before
> sending recovery alerts? Could this be, please advise.
>
Unfortunately I'm on my home windows box right now, so I can't check the
sources properly, but it goes something like this:
if (have_sent_problem_notification(host, contact))
send_recovery_notification(host, contact);
I'd be willing to take a patch to make this configurable, but you'd have
to come up with a neat name for the config option and some documentation
for it, cause I can't for the life of me think what to name or how to
explain something like that.
/Andreas
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]
> Greetings All,
>
> I'm working through an upgrade from Nagios 1 to 3.0.6. I've had several
> situations where I've scheduled downtime for a host *after* it was
> already in Nagios as DOWN and when it comes up it alerted (even though
> it was in scheduled downtime)
Yes. This is because Nagios remembers that it has sent an alert for
the problem state, and therefore it sends the recovery alert as well.
IIRC, this was a feature requested on the list quite some time ago.
> or in my most recent case the host was up
> but a bunch of the service checks I'd added went critical because they
> use the NSClient++ agent which was not loaded yet. In this case after a
> bunch of the service checks had alerted as CRITICAL, I put the host in
> scheduled downtime (verified it by the web interface) and then I
> installed the nsclient++ agent. After it was installed I told Nagios to
> recheck all services on the host to validate the install and I got
> alerts for the recoveries of the service checks.
Same thing here, I'm afraid.
> I posted for this
> before on the users mailing list
> (http://article.gmane.org/gmane.network. ... user/61750), thinking
> there was a possible delay between the web ui and the engine when
> scheduling downtime, but now I'm wondering if it's a problem with the
> logic such that when a host or service is already in a down/critical
> state upon recovery it does not check for scheduled downtime before
> sending recovery alerts? Could this be, please advise.
>
Unfortunately I'm on my home windows box right now, so I can't check the
sources properly, but it goes something like this:
if (have_sent_problem_notification(host, contact))
send_recovery_notification(host, contact);
I'd be willing to take a patch to make this configurable, but you'd have
to come up with a neat name for the config option and some documentation
for it, cause I can't for the life of me think what to name or how to
explain something like that.
/Andreas
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]