contrib/job-scripts/coredump-handling/README
These are a set of scripts that will:
* Setup the core pattern to the "coredumps" sub directory of the
current job and sets the core file size limit to "unlimited"
* Restore the original settings once the job finishes running tests
* Links the core files (if any) on the job level "coredumps" subdir
into the test result specific "coredumps" dir based on the time
stamps of the test directories and the coredump files.
If your system lacks the prlimit utility, then you have to set core
file size limits yourself. Options to do so include running:
$ ulimit -c unlimited
Or setting it permanently with an entry in /etc/security/limits.conf,
or adding a file to /etc/security/limits.d.
To enable these scripts permanently, copy them to the Avocado
systemwide job scripts directories, located by default at
/etc/avocado/scripts/job, respecting the pre and post directories.
To enable these scripts during a development session, set the
following configuration content:
[plugins.jobscripts]
pre = <THIS_FOLDER_PATH>/pre.d/
post = <THIS_FOLDER_PATH>/post.d/
warn_non_existing_dir = True
warn_non_zero_status = True
WARNING: Since the core pattern is a system wide setting, if you run
multiple parallel jobs with these scripts, the result is pretty much
unpredictable and prone to errors.