[Linux-cluster] clvm errors

[Jeff Dinisco]

I'm in the testing phase w/ gfs.  I have 2 nodes in a cluster which are
manually fenced.  I'm working on the basics, creating/removing volume
groups and volumes, creating filesystsems and testing performance.

I created a 20gb volume on a 956gb lun, created a gfs filesystem on it,
and mounted it on both nodes.  I ran some performance tests (creating
10,000 files, removing them, tar, untar, etc.).  The results looked ok
and everything appeared normal.  After running the tests a few times, I
unmounted the fs, and tried to deactivate the volume w/ "vgchnage -aln".
I got this error...

/dev/sdb: lseek 135222817456128 failed: Invalid argument

lvdisplay didn't work but vgdisplay did.  I was able to mount the fs
again.  After umounting, I tried stopping the clvmd daemon and a few
other things, and now I'm at the point where the vol group is not even
found.  

I can easily start from scratch, but this behavior has me concerned.
Any ideas?

 - Jeff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20051118/45cb4e16/attachment.htm>