[Linux-cluster] DLM Problem

[Brian Kroth]

Derek Anderson <Derek.Anderson at compellent.com>:
> 
> 
> > -----Original Message-----
> > From: linux-cluster-bounces at redhat.com [mailto:linux-cluster-
> > bounces at redhat.com] On Behalf Of isplist at logicore.net
> > Sent: Thursday, January 31, 2008 2:43 PM
> > To: linux clustering
> > Subject: RE: [Linux-cluster] DLM Problem
> > 
> > On Thu, 31 Jan 2008 14:16:49 -0600, Derek Anderson wrote:
> > > In your example run gfs_mkfs against /dev/sda1.
> > 
> > Fixed.
> > 
> > >> #gfs_mkfs -p lock_dlm -t compweb:web -j 8 /dev/sda
> > 
> > What's not clear to me now is how I can use the volume's cluster name
> > and FS
> > name for mounting across nodes.
> > 
> > For example, on other nodes, I can only see this as /dev/sdx rather
> > than the
> > less confusing and more easily dealt with cluster name/fs.
> 
> Yes.  And there's another catch.  If node 1 sees it as /dev/sda, it is
> possible that node 2 will see the same block device as /dev/sdb.  Naming
> consistency among cluster nodes is a nice thing that CLVM provides.  And
> it sounds like you are trying to move away from CLVM.  You found one of
> the rubs.  The first few paragraphs of gfs_mount(8) speak to this topic.

It's easy to get around this problem with scsi_id and udev.  It doesn't
necessitate clvm.

> > 
> > Since I created it using compweb:web, how can I now see them and mount
> > them as
> > such on the different nodes?
> 
> I am not aware of a way to mount by LockTableName.

I remember reading somewhere that label support will be added to gfs_mkfs
at some point.

Brian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 2192 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20080131/852afb29/attachment.bin>