[DRBD-user] drbd + nfs for rw access to mirrored partition fr om both nodes

Ziller, James James.Ziller at qg.com
Mon Mar 8 15:15:17 CET 2004

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Well as far as I can tell /dev/nb0 needs to be unmounted on node A before
switching drbd0 to secondary on node A - which needs to happen before the
node B can switch drbd0 to primary and mount it.

I think I have it working now with the scripts i wrote, but they are kind of
ugly and rely on "sleeps" to insure that things get mounted/unmounted in the
correct sequence on both nodes when failing over.  I was hoping someone
might have a "cleaner" way of accomplishing this, but it doesnt look like
anyone does.

-jz

-----Original Message-----
From: Curtis Tiffany [mailto:curtis at reecemarketing.com]
Sent: Friday, March 05, 2004 2:04 PM
To: 'drbd-user at lists.linbit.com'
Subject: Re: [DRBD-user] drbd + nfs for rw access to mirrored partition
from both nodes


I've been thinking of doing this as well.  I've not performed any tests, 
but why are you trying to unmount /dev/nb0 as a part of your failover? 
The secondary machine shouldn't have /dev/nb0 mounted.

If a is primary and b is secondary, a fails
When you are sure you want b to become the primary
	b unmounts /rep
	b sets itself as primary drbd node
	b mounts /dev/nb0 on /mnt/rep
	b mounts /mnt/rep on /rep via nfs

You say you're mounting over nfs using the cluster IP address 
(singular).  Would using static IPs for each node, for use with DRBD, 
help with your situation?

Are you sure you're not trying to unmount /dev/nb0 while it's mounted to 
/mnt/rep or in primary mode?


> Hello,
> 
> 	I am currently using drbd with heartbeat.  The goal is to create a
> cluster where the secondary node has rw access to the mirrored partition
by
> mounting it via nfs.  I am currently mounting /dev/nb0 on /mnt/rep and
then
> also mounting it via nfs (using the cluster IP address) on mountpoint /rep
> on both the primary and secondary node.  The way im doing this creates
> problems when failing over and back again - it appears everything has to
be
> timed just right or otherwise the system hangs when trying to unmount
> /dev/nb0. Does anyone know of an easier/more elegant way of doing this?
> 
> thanks,
> 
> James Ziller
> Information Systems
> 
> Quad/Graphics - Q/DS
> West Allis, Wisconsin
> james.ziller at qg.com
> 
> 
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user
> 

_______________________________________________
drbd-user mailing list
drbd-user at lists.linbit.com
http://lists.linbit.com/mailman/listinfo/drbd-user



More information about the drbd-user mailing list