Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
I would have 2 virtual IPs, one per 2-node cluster. I would then have the servers in each location mount the virtual IP. Should one cluster go down both IPs would migrate to the remaining running cluster. I think this would be a massive headache to configure in pacemaker but I want to see if this scenario is even possible, and what problems it could cause before I start looking into pacemaker. On 22/03/12 13:51, Doros Eracledes wrote: > I'm a new drdb user but as far as I can tell you'll need a cluster > filesystem like GFS or OCFS2 running on both nodes and then you "share" > with NFS. Note that that would still make it an active-standby solution > since the alias IP can only be on one of the two machines > > Theo > > On Thu, Mar 22, 2012 at 12:16:39PM +0000, David Whiteman wrote: >> Hello, >> >> I am currently working on a highly available NFS service. I would >> like to set this up with 2 servers, running RAID 6 and drbd running >> on top of the raid device. I can then provide a virtual IP address >> for the clients to mount the NFS directories. >> >> This will later be expanded upon where we would need to write to the >> NFS storage from another location. Is it possible to have 2 2 node >> clusters, replicated by a stacked resource? Is the stacked resource >> able to work in dual primary mode? >> >> Here is what I want to setup. >> >> [ Primary ]<-------------->[ Primary ] >> [Primary]--->[Secondary] [Primary]--->[Secondary] >> >> Is this possible? > _______________________________________________ > drbd-user mailing list > drbd-user at lists.linbit.com > http://lists.linbit.com/mailman/listinfo/drbd-user