Beren Gamble wrote:
A little advise if you don't mind:
What would be the best configuration for a two node cluster with shared storage
running internet and intranet.
Currently we have seperated the instances of mysql and apache and php required
for internet and intranet. All data is on NFS shares.
Where should I keep the configuration files? I tried keeping them on the NFS
share, but if the NFS connection is broken, the service tries to go down, but
it can't because it needs to read the config file to go down.
Currently, unplugging the network makes it fail over fine, but unplugging NFS
causes issues.
What's generally the best way to approach this setup?
Beren
Hi Beren,
I would suggest you keep your configuration files on a DRBD disk sync'ed
between the nodes, or alternatively keep them locally on the nodes as you would
for a non-cluster setup, but synchronized with something like rsync. I'm using
320GB drives in my cluster nodes right now, carved up into LVM slices for each
service that should have independent failover capabilities and then turning
that LV into a DRBD disk managed by heartbeat. Seems to work great so far and
because it's based on LVM, I have the ability to grow the volumes if necessary
(however unlikely).
--Ryan
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
