CentOS - Sep 2010 - A question about DRBD and nfs

Hi,

I been working on for about a month and half on setup drbd and nfs. I keep
running into issue with the way heartbeat/pacemaker handles nfs. Does anyone
know a good way  to set up a HA NFS server with DRBD and Heartbeat and NFS. I am
willing to share my pain in setting it up.


Chuck Payne
NEO Linux System Engineer
chuck.payne at vocolacity.com


-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.centos.org/pipermail/centos/attachments/20100928/8976906d/attachment.html>

>I been working on for about a month and half on setup drbd and nfs. I keep
running
>into issue with the way heartbeat/pacemaker handles nfs. Does anyone know a
good
>way? to set up a HA NFS server with DRBD and Heartbeat and NFS. I am willing
to share
>my pain in setting it up. 
It involves a bit of manual work which all depends on how seamless you want
failover.
Which I assume is your problem so far. Client mount options play a more
significant
role in this setup.

The drbd part is trivially provided by the included scripts. It will
promote/demote from
sec to pri as needed. The NFS part involves some work.

http://sources.redhat.com/cluster/doc/nfscookbook.pdf

If I can offer a small suggestion, if you're on CentOS, use the boxed RHCS,
its very good
to understate it. Drop pacemaker, but that's my $0.02 CAN:) It's as
simple as yum install
with some cluster.conf tweaking and voila to setup a cluster. If you don't
care about the
locks/states being available on the secondary when it gets promoted (I assume
you're not
using a cluster FS and running pri/pri) then it's easy. When I was learning
RHCS I recall
symlinking the appropriate dirs on the replicated block device.

HTH,
jlc