Created attachment 317435 [details] init from the generated Kdump initramfs Description of problem: From kdump.conf(5), default <reboot | shell> ... reboot: If the default action is reboot simply reboot the system and loose the core that you are trying to retrieve." However, I have seen Kdump still saved the vmcore. From the generated initramfs, I have seen, mount -t ext3 $DUMPDEV /mnt if [ $? == 0 ] then mkdir -p /mnt//var/crash/127.0.0.1-$DATE VMCORE=/mnt//var/crash/127.0.0.1-$DATE/vmcore export VMCORE monitor_cp_progress $VMCORE-incomplete & cp /proc/vmcore $VMCORE-incomplete >/dev/null exitcode=$? if [ $exitcode == 0 ] then mv $VMCORE-incomplete $VMCORE echo -e "\\033[0JSaving core complete" fi fi umount /mnt [ $exitcode == 0 ] && reboot -f reboot -f I was using the following kdump.conf, ext3 /dev/mapper/VolGroup00-LogVol00 default reboot If this behaviour is expected, we may need to fix the manpage. BTW, there looks like a typo in the manpage, - Action to preform instead of mounting root fs and running init process reboot: If the default action is reboot simply reboot the system and loose the core that you are trying to retrieve. + Action to preform instead of mounting root fs and running init process. reboot: If the default action is reboot, simply reboot the system and lose the core that you are trying to retrieve. Version-Release number of selected component (if applicable): kexec-tools-1.102pre-40.el5 How reproducible: always
"default: shell" has the same problem. It probably does not make any sense by dropping to a shell after the vmcore has already been saved.
This is really a disconnect between an old change in implementation and behavior. We changed the initrd so the default action was preformed after the core was captured regardless of success or failure. It dovetailed with the desire for a reboot to happen afterward in most cases. It was also good because dropping to a shell (or halting) sometimes does make perfectly good sense. People may to preform manual recovery actions on system components from other vendors, or otherwise interrogate a system before resetting it. I'll update the documentation to be more clear about this in bz 463404 *** This bug has been marked as a duplicate of bug 463404 ***