[Ocfs2-users] o2cb stack and kernel >= 2.6.37

[Werner Flamme]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Joel Becker [04.04.2011 18:15]:
> On Mon, Apr 04, 2011 at 03:20:43PM +0200, Jacek Stępniewski wrote:
>> Is there anybody who uses o2cb stack with mixed kernels (2.6.37 or more and less than 2.6.37)?
>> I looks like locking protocol version problem (2.6.37 has 1.1. dlm protocol).
>> Is there any patch for this?
> 
> 	The change is rolling upgrade compatible.  If the first machine
> to mount is the newer one, the newer protocol is selected.  If an older
> machine mounts first, the newer machines will join at the older
> protocol.
> 
> Joel
> 

Joel, I think I encountered otherwise :-(

In our OCFS2 cluster there are up to 15 active nodes. 7 of them were
running yesterday (2 with Oracle Linux, 5 with SLES 11 SP1 + SLE HAE
SP1). When applying the last patches of the SLE HAE, the patched nodes
talked dlm 1.1 and did not fallback. They silently unmunted three
volumes, the fourth volume stayed connected (there was no data on it).
In the logs, I found entries like

dlm_query_join_proto_check:734 Node 5 wanted to join with DLM locking
protocol 1.0, but we have 1.1, disallowing
o2net: connection to node vocnod9 (num 5) at 141.65.171.51:7777
shutdown, state 8
o2net: no longer connected to node vocnod9 (num 5) at 141.65.171.51:7777

and node 5 lost access to at least one volume straight after this.

Maybe a specialty of SUSE, I don't know. I could not get the nodes to
communicate with the rest of the cluster, I had to undo all of the
patches provided from HAE SP1-update repo and to reboot before it worked
again. Maybe the libdlm3 package was the culprit. I opened a support
case at Novell and will report back what they say...

Regards,
Werner
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.15 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/

iEYEARECAAYFAk2fI0AACgkQk33Krq8b42OXKACggP45Psn4/yqUs2TMIY9y8Vlk
HxoAn2bGrYhPyuIhXxfNf6fdivSC900R
=kKRY
-----END PGP SIGNATURE-----