Cannot get a core dump

[BuSchu]

We have an old server running CentOS 6.10. Now I want to have a core dump file if our application crashes, maybe because of a segmentation fault. I configured the following:

/etc/sysctl.conf:
kernel.core_pattern = /var/crash/core-%e-%s-%u-%g-%p-%t

/etc/sysconfig/init:
DAEMON_COREFILE_LIMIT='unlimited'

Then I called "sysctl -p". Next I sent our program a SIGSEGV signal. It was terminated and automatically restarted. But even after repeating this after a reboot there is even no directory /var/crash. And in /var/log/messages I read only:
<process> (<id>) terminated with status 139

What should I do to get a dump file?

Regards
Burkhard

[Errosion]

Just curious.

If you do:

Code: Select all

ulimit -a | grep core 

Do you see your core file size set to unlimited?

[BuSchu]

[root@mail2 ~]# ulimit -a | grep core
core file size (blocks, -c) 0
[root@mail2 ~]# su - email
[email@mail2 ~]$ ulimit -a | grep core
core file size (blocks, -c) unlimited

The process is running under the email user. I think, the setting is OK! Or am I wrong?

Regards
Burkhard

[rmhartman]

in /var/log/messages is there something from abrtd saying something about "XXX doesn’t belong to any package and ProcessUnpackaged is set to ‘no’" ?

sudo vim /etc/abrt/abrt-action-save-package-data.conf
(change setting ProcessUnpackaged from no to yes)
sudo service abrtd restart

I think after that the cores end up in /var/spool/abrt, but there should be some messages in /var/log/messages that will tell you where it is.