[Ocfs2-users] Kernel panic when deleting a file

[Kristiansen Morten]

Hi,

We have three RHEL4 servers with ocfs-1.2.4 installed. Now I want to upgrade both RedHat and ocfs2 in this way:
<on all nodes>
Reinstall with RHEL5
Install ocfs2-1.2.9
Reinstall oracle with clusterware and submit to the Oracle cluster </on all nodes>

<on all nodes>
Upgrade ocfs2 from 1.2.9 to 1.4.7
</on all nodes>

I have reinstalled one of the servers with RHEL5 and then tried different things because I wasn't shure of how to do this. Ended up with the list above. The things I've done is:
	Installed rpm packages for ocfs2-1.4.4 
	Deinstalled rpm packages for ocfs2-1.4.4
	Downloaded, built and installed source packages for ocfs2-1.4.7
	Installed rpm packages for ocfs2-1.4.7 (without deinstall of the selfbuild source of ocfs2-1.4.7, don't know how)
	Deinstalled rpm packages for ocfs2-1.4.7
	Downloaded, built and installed source packages for ocfs2-1.2.9

In the end I was able to mount the ocfs2 disks. So I moved on installing all needed Oracle software, including clusterware. When trying to start the Oracle CRS, the server rebooted and kept rebooting until I was able to disable Oracle CRS automatic start.



Today I tried to create a file on one of the ocfs2 filesystem: 
# echo mktest>/disk07/mktest.txt		

This worked well, and then I tried to delete it:
# rm /disk07/mktest.txt

Then I lost connection and a "kernel panic" error was displayed on the console screen. Now I'm wondering what could be wrong? Could it be that the selfbuild source installation of ocfs2-1.4.7 isn't overwritten/deleted and causing this problem? Does anybody know how to deinstall this? Any suggestions what to do?

I've tried to post this message twice before, but it didn't show up. Is it because I attached a wordpad-file with a screendump of the "kernel panic". If so, how can I attach a file and which files isn't stopped by a filter?

Regards
Morten K