Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
i want to say that drbd can detect changes to the underlying block devices and do the needful. but i don't know for sure. to be safe, one should consider that any changes to the underlying block device, such as the those that happen with volume being mounted RW, are not replicated through drbd automatically. before the secondary server can be used, one would need to re-do the mounts on the master, force a sync, and wait for the sync to complete. drbdadm invalidate on the secondary or invalidate_remote on the primary can be used for force a full resync. mike Dan Barker wrote: > Won't the disk changes between the drbd sync and switching the mount to > /dev/drbdX fail to be replicated? I think the mount point switch needs to > happen before the secondary node is brought online for the replica to be > valid. My understanding was that the sync doesn't need to be complete, but > the access does need to be via the drbd device rather than the underlying > block device. > > Dan Barker > > -----Original Message----- > From: Mike Lovell > Sent: Thursday, July 08, 2010 4:54 PM > <snip> > > you should adjust your mount points so that the primary node is mounting the > /dev/drbdX device instead of mounting the raw block device. that can be done > with just a few seconds of downtime after drbd is set up and it doesn't have > to happen immediately. > > </snip> > > _______________________________________________ > drbd-user mailing list > drbd-user at lists.linbit.com > http://lists.linbit.com/mailman/listinfo/drbd-user >