Nagios Support Forum

I have a setup, whereby I have two master hosts, which collate input from a number of smaller nagios installations... which send their data using NSCA

All of the nagios instances are now running 4.0.1, and all of the hosts are running CentOs (mostly 6.4 - but the masters are running 5.8).

I have nagios set-up as a service, so in theory, I have a number of options to refresh nagios; In the past (Nagios 3.5.0 and previous), I always used kill -HUP - with no issues.

This is also the case with the smaller installations.... but in almost every case, when I do this with the master collectors I get a SIGSEGV error, and nagios is no longer running - it doesn't happen every time.... but it happens MOST times... it appears that this MIGHT be related to nsca processes still running ?

I don't get an issue if I stop / start nagios, as opposed to forcing it to re-read the config file (and I have confirmed that the new config is clean).

So, two questions...
- is there a "best" way to re-start nagios with a changed config ?
- assuming I can re-produce the core file - what other information is required, and what should I do with this ?

Thanks, Malcolm

I would suggest just running: If you did want to make sure that all nagios processes were closed, I'd run:

OK, thanks... that is certainly a workable solution.

Is there any mileage towards submitting a bug to the tracker with regards to the SIGSEGV / HUP / nsca combination ?

Its certainly not an urgent problem... but my logic is that I have a captive situation... and can reproduce almost at will... so does it make sense to create a new tracker entry ?

Malcolm

Yes. If the behavior changed and was not noted in the release notes, or was previously acknowledged as working as intended, opening a bug report at http://tracker.nagios.org is most likely justified.