post crash core removal

[Cody Godines_1]

Hi everyone,
I'm on an HP c240 visualize. I have a graphical application that crashes and the system creates a core file. I'd like to have it automatically removed. Is this possible?

-Cody

[Pete Randall]

Cody,

You can do this with a crontab entry like this:

00 02 * * * /usr/bin/find / -name "core" -mtime +2 -exec rm {} \;

This would remove any core files which are over 2 days old.

Pete

[Ted Ellis_2]

only one core file (or directory) can exist under any one part of the directory trees... we protect our / mount point by also creating a /core directory. This prevents any core file from being dropped at the top level /

Ted

[Cody Godines_1]

Those are good ways to do it. But I think there is a feature that can be turned off or something. That is what I'm looking for.
-Cody

[Ted Ellis_2]

link with some good info and a way to prevent them, though in some cases you may want them to help analyze what is causing the problem

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x947a7d4cf554d611abdb0090277a778c,00.html

[John Poff]

Hi,

I think you can do a 'ulimit -c 0' in your profile which specifies the number of 512 byte blocks to use for core dumps, but that only works for the Posix shell. I don't think you can set it in ksh, and I don't know about other shell flavors (csh).

JP

[Pete Randall · Accepted solution]

Cody,

Have a look at /etc/rc.config.d/savecrash. In particular check the SAVECRASH parameter:

# SAVECRASH: Set to 0 to disable saving system crash dumps.

Pete

[Pete Randall]

Sorry, Cody, I got interrupted in mid-thought there. You could mess with SAVECRASH, but in my opinion, you're better off to use my earlier "find" approach. It will keep them cleaned out but still allow you to examine one should you have a problem.

Pete