seligman at nevis
Nov 18, 2011, 9:41 AM
Views: 373
Permalink
|
I just read the (excellent) document on setting up NFS storage with DRBD and
Pacemaker:
<http://www.linbit.com/en/education/tech-guides/highly-available-nfs-with-drbd-and-pacemaker/>
What's in that guide is not the way I did it. I'd like to know if I made a mistake.
In my current Pacemaker+DRBD+NFS cluster, I export directories by editing lines
in /etc/exportfs on the server, and have a cron job copy /etc/exportfs from my
primary to my secondary in case I make any changes. The NFS server is under the
control of Pacemaker.
Initially I had problems with failover, and conclude (wrongly?) that the issue
was the NFS server state in /var/lib/nfs. I added that directory to the DRBD
volume, so that /var/lib/nfs would be consistent between primary and secondary
during failover. This works, mostly; I can "service corosync stop" on the
primary and everything transfers smoothly to the secondary, but I have to reboot
the primary before can I transfer resources back to that same box; otherwise I
get NFS errors.
While reading the above document, I just learned about the Pacemaker exportfs
resource. *headdesk* Is that the correct way to do it? If I used that resource,
would I have to worry about maintaining state information in /var/lib/nfs across
my cluster nodes?
--
Bill Seligman | Phone: (914) 591-2823
Nevis Labs, Columbia Univ | mailto://seligman [at] nevis
PO Box 137 |
Irvington NY 10533 USA | http://www.nevis.columbia.edu/~seligman/
|
smime.p7s
DRBD and /var/lib/nfs
