[Linux-cluster] clvmd version mismatch

[Dan Brown]

I'm trying to get a two node GFS2 cluster running on a pair of test
machines using GFS2 on top of a clustered lvm2 disk replicated over a
DRBD 8.2.0 active/active link. 
After starting up drbd and cman, I attempt to start up rgmanager and
clvmd and they both fail with this message in /var/log/syslog.

Nov 13 22:25:36 seahawk kernel: dlm: process clurgmgrd (19787) version
mismatch user (5.0.0) kernel (6.0.0)
Nov 13 22:26:06 seahawk kernel: dlm: process clvmd (20143) version
mismatch user (5.0.0) kernel (6.0.0)

After trying to tune in various parameters to get more logging or
debugging, apart from a somewhat uninformative strace, I can't seem to
find a whole lot more about which userspace items are conflicting with
which kernel space items. 

Oh, to make things a little worse, this is on a pair of Fedora FC6
machines rather than RedHat EL or CentOS.

----
Dan Brown
dcmbrown at shaw.ca