[Linux-cluster] GFS mounting problem.

[David Teigland]

On Tue, May 22, 2007 at 05:02:31PM +0200, Dan Deshayes wrote:
> Hello,
> I've just updated my two Centos5 nodes, with this update it installed a 
> new kenerl (2.6.18-8.1.4)
> and removed kmod-gfs and kmod-gnbd.
> But now I can't mount my gfs volumes with the new kernel.
> 
> on the dbLV it says error 16 (Device or resource busy)
> # mount -t gfs -v /dev/clusterVG/db /cluster/db/
> /sbin/mount.gfs: mount /dev/clusterVG/db /cluster/db
> /sbin/mount.gfs: parse_opts: opts = "rw"
> /sbin/mount.gfs:   clear flag 1 for "rw", flags = 0
> /sbin/mount.gfs: parse_opts: flags = 0
> /sbin/mount.gfs: parse_opts: extra = ""
> /sbin/mount.gfs: parse_opts: hostdata = ""
> /sbin/mount.gfs: parse_opts: lockproto = ""
> /sbin/mount.gfs: parse_opts: locktable = ""
> /sbin/mount.gfs: message to gfs_controld: asking to join mountgroup:
> /sbin/mount.gfs: write "join /cluster/db gfs lock_dlm acl002:cluster_db 
> rw /dev/clusterVG/db"
> /sbin/mount.gfs: lock_dlm_join: gfs_controld join error: -16
> /sbin/mount.gfs: error mounting lockproto lock_dlm

This error is usually due to a previous mount that failed but didn't get
cleaned up properly.  Run 'group_tool' and you'll probably see a
cluster_db gfs group that should have been cleared previously.  Restarting
the cluster on that node is surest way to fixing it up.

(Try not to kill mount, e.g. with control-c, since that can lead to things
not being cleaned up.)


> on the other two LVs it says error 19 (No such device)
> # mount -t gfs -v /dev/syncaderoVG/web /syncadero/web/
> /sbin/mount.gfs: mount /dev/syncaderoVG/web /syncadero/web
> /sbin/mount.gfs: parse_opts: opts = "rw"
> /sbin/mount.gfs:   clear flag 1 for "rw", flags = 0
> /sbin/mount.gfs: parse_opts: flags = 0
> /sbin/mount.gfs: parse_opts: extra = ""
> /sbin/mount.gfs: parse_opts: hostdata = ""
> /sbin/mount.gfs: parse_opts: lockproto = ""
> /sbin/mount.gfs: parse_opts: locktable = ""
> /sbin/mount.gfs: message to gfs_controld: asking to join mountgroup:
> /sbin/mount.gfs: write "join /syncadero/web gfs lock_dlm 
> acl002:cluster_web rw /dev/clusterVG/web"
> /sbin/mount.gfs: message from gfs_controld: response to join request:
> /sbin/mount.gfs: lock_dlm_join: read "0"
> /sbin/mount.gfs: message from gfs_controld: mount options:
> /sbin/mount.gfs: lock_dlm_join: read "hostdata=jid=0:id=589825:first=1"
> /sbin/mount.gfs: lock_dlm_join: hostdata: "hostdata=jid=0:id=589825:first=1"
> /sbin/mount.gfs: lock_dlm_join: extra_plus: 
> "hostdata=jid=0:id=589825:first=1"
> /sbin/mount.gfs: mount(2) failed error -1 errno 19

This error is usually due to the gfs kernel module not being loaded,
lsmod | grep gfs, and make sure both gfs and gfs2 modules are loaded.

Dave