Nagios3 checks not scheduled after clocks change from DST

[Chris May]

Binary package hint: nagios3

http://labs.opsview.com/2009/10/nagios-scheduling-bug/ describes a bug which we also encountered on nagios 3.2 in lucid. After the clocks are set back 1 hour for Daylight Saving Time (last weekend, in the UK), nagios fails to schedule any service checks for the next 24 hours. Manually rescheduling a check causes that check to start being checked properly again.

I believe this was fixed in nagios v 3.2.2; the changelog (http://www.nagios.org/projects/nagioscore/history/core-3x) includes "Fix for choosing next valid time on day of DST change when clocks go one hour backwards". However I can't find a corresponding bug in nagios's bugtracker.

Would it be possible to apply the same patch to the 3.2.0 source in Lucid? Or failing that, consider upgrading to 3.2.2 in some future release of Ubuntu, to avoid this problem?

[James Page (james-page)]

Natty should push the version of Nagios3 up to the latest (v3.2.3) but bearing in mind that Lucid is a LTS release I think it would make sense to request a stable release update to resolve this issue.

It would be good to get confirmation of this build from another source to support the case for an SRU.

[jon (jb-subs)]

We have been using Nagios3 on Ubuntu 10.04 LTS for the past 6 months or so and ran into this issue shortly afterwards in November 2010 (our last daylight savings change). Nagios had stopped scheduling any more checks and so was not doing anything, but this was not immediately obvious.

Found the nagios 3.2.0 daylight savings bug described here, in an article dated Oct 2009 (!)
http://www.chown-r.co.uk/2009/10/nagios-schedule-checks-do-not-run-after-daylight-savings-change/

We worked around it as they suggested (set use_retained_scheduling_info=0 and performed nagios restart):

However as Ubuntu 10.04 LTS is still not distributing any version newer than 3.2.0 we are still exposed to this bug and are about to be hit by it again as another daylight savings change is about to happen.

Can an updated version of nagios3 *please* be distributed.

[Paul Crawford (psc-sat)]

Same here - would have thought that an update should be out by now.

It is supposed to be supported in LTS for 5 years and I presume 'support' includes pushing out established bug-fixes!

[Paul Crawford (psc-sat)]

I see no activity here. Why is this not being maintained? It is not like there is no bug fix - can someone at Canonical actually do something useful about updating a *supported* LTS package?

[Paul Crawford (psc-sat)]

So folks, what IS happening about this bug, already fixed, for a package that is SUPPOSED to be supported in 10.04?

[jon (jb-subs)]

Another year, another daylight savings change, another failure of our Nagios system due to a bug that was fixed in the Nagios code a LONG time ago - yet has STILL not been pushed out through the Ubuntu update manager for 10.04 LTS.

What is the point of "Long Term Support" exactly??

[James Sellman (wd-jim-qp)]

We just got bit by this.

It looks like there *WAS* a fix in the Lucid package at one point, but then it was removed because:

debian/patches/96_fix_daylight-saving.dpatch: Already upstream.

Well, it is NOT upstream in 3.2.0. =/

[jon (jb-subs)]

Another daylight savings change has just happened, and once again our Nagios monitoring system silently failed to run any checks after the change, needing to be restarted.

Why does nobody out there care that the repositories for an LTS release of Ubuntu STILL only offer the same buggy version of Nagios (3.2.0) - They are up to 3.4.1 now and the bug was fixed a LONG time ago!!

Clearly LTS means absolutely NOTHING whatsoever.

[Colin Coghill (cqlin)]

Argh, just lost a good portion of a day trying to figure out why Nagios had stopped monitoring.