Hello all
I'm building a drbd - pacemaker - nfs Server in active - passive mode, but
there remain some open bits.
I have two IP'Adresses on my server, how would I do that ? The example (see
below) to work, that means, it doesn't produce any errors whereas the params
nfs_ip line written with a space instead of a comma produces this error :
lrmd: [1436]: info: RA output: (P_NFSSERVER:monitor:stderr)
/usr/lib/ocf/resource.d//heartbeat/nfsserver: line 229: [: 10.1.4.11: binary
operator expected
However, it seemed possible to access the server in spite of that from both
subnets so I wondered what the "nfs_ip" does and what it is needed for.
Would the below example be correct for two IP adresses ? At least, I can't
see any error.
As primitive, I'm using ocf :heartbeat :nfsserver
primitive P_NFSSERVER ocf:heartbeat:nfsserver \
params nfs_init_script="/etc/init.d/nfs" \
params nfs_notify_cmd="/sbin/rpc.statd" \
params nfs_shared_infodir="/var/lib/nfs" \
params nfs_ip="10.1.4.11,10.1.0.75" \
op monitor interval="15"
primitive P_NFSLOCK lsb:nfslock
I used http://linux-ha.org/HaNFS for setting up the shared directory on
drbd. However, I didn't fully understand the STATD_HOSTNAME part.
It's written that it is used for lock recover - how would I do that for my
two IP adresses?
STATD_HOSTNAME=10.1.4.11 10.1.0.75
Would that be correct? And how could I find out that it works correctly - I
mean, where do I see a difference when I do not set this parameter at all?
By the way: I'm using NFSv3 not NFSv4 because of this:
http://www.mail-archive.com/[email protected]/msg09445.html
which is exactly the same for me.
With NFSv3 I have a total failover time of about 10 seconds, with NFSv4
about 90 seconds.
Lowering the grace period isn't recommended in any case, see here:
http://www.spinics.net/lists/linux-nfs/msg22713.html
Also, see here:
http://wiki.linux-nfs.org/wiki/index.php/NFS_Recovery_and_Client_Migration
and here:
http://serverfault.com/questions/20465/nfs-v4-ha-migration-and-stale-handles
-on-clients
I'm not so sure it is possible to use NFSv4 and having a smooth failover in
a pacemaker environment.
Thanks in advance!
Tim
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
