We’ve been having some issues with a 3rd party driver (EMC powerpath) that causes our SLES11 SP1 boxes to crash under certain circumstances. All of the configuration settings are in place to capture a crash dump. Except that the /var filesystem does not have enough space to write the dump files. I thought it would be a simple matter to add /var/crash as a separate filesystem with enough room to hold the files (16GB). Unfortunately, the crash kernel doesn’t notice this a a distinct mount point and still tries to write under the original /var . It appears that the crash kernel is LVM aware since all of our OS filesystems are built on LVM. It’s also reading /etc/fstab or it wouldn’t know to mount /var And, yes, this /var/crash is mounted at the time of a crash.
How to get it to take that extra step and look for /var/crash in /etc/fstab ?? I would prefer not to just expand /var to hold the dump files since that would become available space that regular processes could potentially consume.