This is a little How-To from oops-ups I have experienced.


gfs2 and drbd

Recovering files on gfs2 on a broken drbd

First: Bring up the cluster

drbdadm up <cluster name>

If the local node is secondary make it primary

drbdadm primary <cluster name>

Then disconnect it from its peer

drbdadm disconnect <cluster name>

Finally mount the file system. Note that mounting read-write did not work for me

mount -o localcaching,localflocks,lockproto=lock_nolock,ro <device> <mount point>