Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
/ 2004-07-16 14:57:40 +1200
\ James Doherty:
> Hi guys,
>
> I've been asked to set up DRBD for our database and fileserver. We have
> MySQL running on the database server and NFS running on the fileserver.
> Tools used for this are: mysql, nfs, drbd and heartbeat. The way it
> needs to be set up is as follows:
>
> The drbd device on the database server is to be sync'd to the
> fileserver. If the database server fails, then the fileserver will
> failover the ip of the database server and take over as the primary drbd
> device for the database resource. The fileserver will now be running
> both mysql and nfs.
>
> If the opposite happens and the fileserver fails, then the database
> server will failover the ip of the fileserver and it will be serving nfs
> in addition to mysql.
>
> I've got failover working fine for one drbd resource between these
> machines. However adding a resource for the other machine adds a whole
> new level of complexity. I've had trouble getting the individual
> machines to come up as the primary machine for their resource and I'm
> also not sure how I'd migrate a resource back in the case of one machine
> failing and it being brought back up again.
>
> Has anyone had to do this before? Any help would be much appreciated.
/etc/init.d/drbd
USAGE: drbd [resource] start|stop|status|reconnect|checkconfig
OPTIONS: --dry-run --info --parse --dummy --config=arg --dontwait --dontkill
USAGE: datadisk [resource] start|stop|status
OPTIONS: --dry-run --info --parse --dummy --config=arg --dontwait --dontkill
so, you can just do
# configure and connect all resources
on both# drbd start
# make database resource primary on this node
on data# datadisk database start
# make fileserv resource primary on that node
on file# datadisk fileserv start
hth,
Lars Ellenberg