[linux-lvm] Persistent minor number problem in lvm2

Chris Osicki osk at admin.swisscom-mobile.ch
Mon Jan 8 11:24:21 UTC 2007


On Fri, 5 Jan 2007 08:01:46 +0100

Hi Paul

Don't let OS generate filesystem handles from major/minor numbers.
Have a look at fsid option for exportfs (man 5 exports). It works. 
It's how I implemented it on my failover NFS clusters.
BTW, persistent minors works as well, I found fsid easier to manage, though.

Hope it helps.

Regards
Chris

Rainer Krienke <krienke at uni-koblenz.de> wrote:

> Am Donnerstag, 4. Januar 2007 21:37 schrieb Mark H. Wood:
> > Mount by LABEL= or UUID=?
> >
> > Device node numbers are good for physical things like serial lines,
> > but not so good for logical things like storage volumes.  If you can
> > label it, do, and always identify it by your label.
> 
> I doubt that this would solve the problem since in the scenario I had, the 
> client already had mounted a directory from a server. Then on the server the 
> minor ids of exported NFS volumes changed and then the client did see 
> the "wrong" contents in the "right" NFS share mounted from the server. The 
> problem is that NFS relies on major, minor numbers in its protocol, I think. 
> So UUIDs could only help  in an initial mount but not later.
> 
> Have a nice day
> Rainer
> -- 
> ---------------------------------------------------------------------------
> Rainer Krienke, Universitaet Koblenz, Rechenzentrum, Raum A022
> Universitaetsstrasse 1, 56070 Koblenz, Tel: +49 261287 -1312, Fax: -1001312
> Mail: krienke at uni-koblenz.de, Web: http://www.uni-koblenz.de/~krienke
> Get my public PGP key: http://www.uni-koblenz.de/~krienke/mypgp.html
> ---------------------------------------------------------------------------




More information about the linux-lvm mailing list