I use nagios-3.0b5 on FreeBSD 6.2. It works fine, but sometimes goes into endless loop, consuming cpu and causing high load average (up to 100 and even more). At this time, there may be 600 nagios processes, despite the limit of 40.
Well, I’ve tried to debug the issue, and at first set daemon_dumps_core=1. But core didn’t appear. Little investigation shows that nagios looks for environment variable ‘HOME’ and creates core there, using / if no ‘HOME’ found. In my case, HOME was not set, and core was not created because user nagios can’t write to /.
It’s somewhat strange to use environment to determine homedir, so I’ve created a patch to determine homedir through getpwnam(3) call. I’ll be happy to give the patch to developers for review and merging into the source.
With best regards,