[Linux-cluster] GNBD & Network Outage

[Benjamin Marzinski]

On Fri, Feb 04, 2005 at 10:26:36AM +0000, Nigel Jewell wrote:
> Dear Ben,

Oh yeah, or you could take a look at http://www.drbd.org/

-Ben
 
> Thank you for your detailed reply.  It is always refreshing to get a 
> decent response on a mailing list ;) .
> 
> >Sure. You see the -c in you export line.  Don't put it there.  That puts
> >the device in (the very poorly named) uncached mode.  This does two 
> >things.
> >One: It causes the server to use direct IO to write to the exported 
> >device,
> >so your read performance will take a hit.  Two: It will time out after
> >a period (default to 10 sec).  After gnbd times out, it must be able 
> >to fence
> >the server before it will let the requests fail.  This is so that you 
> >know
> >that the server isn't simply stalled and might write out the requests 
> >later
> >(if gnbd failed out, and the requests were rerouted to the backend 
> >storage over
> >another gnbd server, if the first server wrote it's requests out 
> >later, it
> >could cause data corruption).
> > 
> >
> 
> My understanding was that the "-c" put the device in cached mode, as 
> described here:
> 
> http://www.redhat.com/docs/manuals/csgfs/admin-guide/s1-gnbd-commands.html
> 
> Or are you saying that by not putting the "-c" put its in uncached mode?
> 
> >This means that to run in uncached mode, you need to have a cluster 
> >manager and
> >fencing devices, which I'm not certain that you have.
> > 
> >
> 
> No we don't as we didn't really see the need, given what we want to do.
> 
> >I've got some questions about your setup.  Will this be part of a 
> >clustered
> >filesystem setup? If it will, I see some problems with your mirror.  When
> >other nodes (including the gnbd server node A) write to the exported 
> >device,
> >these writes will not appear on the local partion of B.  So won't your 
> >mirror
> >get out of sync?  If only B will write to the exported device, (and 
> >that's
> >the only way I see this working) you can probably get by with nbd, which
> >simply fails out if it loses connection.
> > 
> >
> 
> The intention of the setup was to have two hosts both exporting an 
> unmounted device, and the alternative device using it as a RAID-1 
> device.  Then to use heartbeat to mount and unmount the partitions as 
> required.  For example:
> 
> HOST A:
> 
> /dev/hda1 (md0, ext3, mounted)
> /dev/hda2 (ext3, unmounted, gnbd_exported as A)
> /dev/gnbd/B (md0, ext3, mounted)
> 
> HOST B:
> 
> /dev/hda1 (ext3, unmounted, gnbd_exported as B)
> /dev/hda2 (md0, ext3, mounted)
> /dev/gnbd/A (md0, ext3, mounted)
> 
> I hope that makes sense.
> 
> If so, does what we are trying to achieve sound sensible?  Any 
> gotchas/advice?
> 
> -- 
> Nige.
> 
> PixExcel Limited
> URL: http://www.pixexcel.co.uk
> 
> --
> Linux-cluster mailing list
> Linux-cluster at redhat.com
> http://www.redhat.com/mailman/listinfo/linux-cluster