Cannot mount GFS2 in F8 - /sbin/mount.gfs2: waiting for gfs_controld to start

[Albert Graham]

Hi,

I'm trying mount a GFS2 disk using the clustering in Fedora 8 but 
gfs_controld dies (see below).

For example:

# service cman start (ver. 6.0.1 config 20)

Starting cluster:
   Loading modules... done
   Mounting configfs... done
   Starting ccsd... done
   Starting cman... done
   Starting daemons... done
   Starting fencing... done
                                                           [  OK  ]

# mkfs.gfs2 -t home:gfs -p lock_dlm -j 2  /dev/drbd0

This will destroy any data on /dev/drbd0.

Are you sure you want to proceed? [y/n] y

Device:                    /dev/drbd0
Blocksize:                 4096
Device Size                4.00 GB (1048535 blocks)
Filesystem Size:           4.00 GB (1048532 blocks)
Journals:                  2
Resource Groups:           16
Locking Protocol:          "lock_dlm"
Lock Table:                "home:gfs"


; Check that gfs_controld is running (as you see it is)

ps -ef|grep gfs
root      4915     2  0 08:54 ?        00:00:00 [gfs2_scand]
root      6081     1  0 09:32 ?        00:00:00 /sbin/gfs_controld


; Now try to mount it:

# mount -t gfs2  /dev/drbd0 /net -o 
locktable=home:gfs,lockproto=lock_dlm -vv

mount: no LABEL=, no UUID=, going to mount /dev/drbd0 by path
/sbin/mount.gfs2: mount /dev/drbd0 /net
/sbin/mount.gfs2: parse_opts: opts = 
"rw,locktable=home:gfs,lockproto=lock_dlm"
/sbin/mount.gfs2:   clear flag 1 for "rw", flags = 0
/sbin/mount.gfs2:   add extra locktable=home:gfs
/sbin/mount.gfs2:   add extra lockproto=lock_dlm
/sbin/mount.gfs2: parse_opts: flags = 0
/sbin/mount.gfs2: parse_opts: extra = 
"locktable=home:gfs,lockproto=lock_dlm"
/sbin/mount.gfs2: parse_opts: hostdata = ""
/sbin/mount.gfs2: parse_opts: lockproto = "lock_dlm"
/sbin/mount.gfs2: parse_opts: locktable = "home:gfs"
/sbin/mount.gfs2: waiting for gfs_controld to start  <-- DIES HERE
/sbin/mount.gfs2: waiting for gfs_controld to start
/sbin/mount.gfs2: waiting for gfs_controld to start
/sbin/mount.gfs2: waiting for gfs_controld to start
/sbin/mount.gfs2: waiting for gfs_controld to start
/sbin/mount.gfs2: waiting for gfs_controld to start
/sbin/mount.gfs2: waiting for gfs_controld to start
/sbin/mount.gfs2: waiting for gfs_controld to start
/sbin/mount.gfs2: waiting for gfs_controld to start
/sbin/mount.gfs2: waiting for gfs_controld to start
/sbin/mount.gfs2: gfs_controld not running
/sbin/mount.gfs2: error mounting lockproto lock_dlm

; At this point gfs_controld is not in memory (as it segfaulted)

' At this point, run gfs_controld in debug mode, and try to mount again 
and get the following:-

# /sbin/gfs_controld -DP # (ver. gfs_controld (built Feb 27 2007 17:14:07)
1208162228 listen 3
1208162228 cpg 6
1208162228 groupd 8
1208162228 uevent 9
1208162228 plocks 12
1208162228 setup done
1208162232 client 6: join /net gfs2 lock_dlm gfs 
rw,locktable=gfs,cluster=home /dev/drbd0
1208162232 mount: /net gfs2 lock_dlm gfs rw,locktable=gfs,cluster=home 
/dev/drbd0
Segmentation fault


; My kernel is:

"2.6.24.4-64.fc8PAE" but get the same results on another machine 
running: "2.6.23.8-63.local.fc8"

The drive DOES mount OK with "lockproto=lock_nolock"

The underlying device is DRBD block device (ver 8.2.5) which is backed 
by a /dev/loop3 file device, but I'm sure that has nothing to do with 
this problem (as ext3, XFS work OK as well as GFS2 with lock_nolock).


Any clues ?

TIA,

Albert.