In response to [email protected]:
> Has anyone had problems obtaining core dumps from the main Nagios daemon
> when it segfaults?
>
> I've enabled the option in the config file "daemon_dumps_core" but, after
> the daemon dies due to a segfault (confirmed via dmesg,) a core file never
> seems to be produced.
>
> And yes, I've set the global ulimit for core dumps to unlimited, and yes,
> there is enough space on my disk, and yes, I've scoured the disk with
> "find" looking for files with the name of "core" or "core.", but to
> no avail.
>
> I also completely disabled the segfault handling logic in the signal
> handler, but nothing's changed.
>
> What am I missing here?
>
> I'm considering running nagios in the foreground (non-daemon mode) just to
> see if I can catch the magically-missing core file.
What OS is this on?
FreeBSD has sysctls that control how core dumps are handled, and can even
disable them at the OS level. I assume that other POSIX systems have
similar capabilities:
$ sysctl -a | grep core
kern.sugid_coredump: 0
kern.coredump: 1
kern.nodump_coredump: 0
kern.corefile: %N.core
$ sysctl -d kern.sugid_coredump
kern.sugid_coredump: Enable coredumping set user/group ID processes
Notice that the default in FreeBSD is to _not_ allow setu/gid programs to
create core files ...
--
Bill Moran
Collaborative Fusion Inc.
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]