Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Mon, Jun 04, 2012 at 06:39:12PM +0200, Sebastian Kaps wrote: > Hi, > > I'm having a two-node Corosync/Pacemaker cluster (active/passive) with XFS on two DRBD-volumes (SLES11 SP1 + HAE). > The active node has both volumes mounted and if this node fails, the passive node should take over and mount these volumes. > I'm aware that I can't mount XFS simultaneously on both nodes (which is not a problem). > > The setup runs fine so far, with both DRBD-volumes in dual-primary mode (and STONITH configured). > But I'm currently not sure, if the dual-primary mode is actually needed, Not at all. > or if there's a way to run in safer primary/secondary mode. Yes. Just do it. > I think, in an earlier stage of this cluster and the disks being > configured as primary/secondary, I've had problems to get the > secondary disk promoted to primary after the old primary failed (i.e. > was shut off). > I remember a "refusing to be primary while peer is not outdated" message in that case, but that may have had other causes. > > What's the best/safest approach in my case? -- : Lars Ellenberg : LINBIT | Your Way to High Availability : DRBD/HA support and consulting http://www.linbit.com