How to make a coredump

From Zarafa wiki

Latest revision as of 12:47, 3 January 2013

How to report segfaults and coredumps

When a Zarafa services creates an error, you may see the following error in your log:

Caught SIGSEGV, traceback:
0x0000000000481D87
...

To correctly report this issue to the Zarafa development team, you can write an email to support@zarafa.com with these lines from the log.

You must also add the following information:

• Zarafa version
• Distribution name and version
• Architecture (i386, x86_64)

To help the development team even further, you can create a coredump of the process that crashed. To to that, you need to add the following command before restarting the service again:

ulimit -c unlimited

Then start the service. When the services crashes again, you will have a file called 'core' or 'core.PID', where PID is the pid number the process had. This file may be found in the starting directory or the path given in the running_path option in the server.cfg file. Compress this file:

gzip core

And send the core.gz file to support@zarafa.com with the same information about versions and distributions used. If the compressed file is larger than 40 Mb, please upload it somewhere where we can download the file, and add a link in the email.

Adding ulimit to all Zarafa processes

Instead of manually running ulimit while you are logged in to the terminal.

You may also find that it is easier to add it to the init.d script of the process, or to the /etc/default/zarafa under Debian/Ubuntu bases distrobutions.

If the server process was restarted after adding ulimit a full 'unlimited' coredump file will be generated which will approximately have a file size of occupied memory space so make sure you have enough free disc space.

Which will generally make it much easier for Zarafa to debug the crash.

Add this line "case "$1" in", or add as first line in the /etc/default/zarafa file.

ulimit -c unlimited