Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
In a two-node, primary/primary configuration with drbd-0.8pre6 when both nodes have rebooted I have a split-brain situation for some reason. Yet, I cannot seem to remedy it! Here is what I have tried (and many permutations thereof in frustration ;): node0: drbdsetup /dev/drbd0 primary node1: drbdsetup /dev/drbd0 secondary node1: drbdsetup /dev/drbd0 outdate Both still in standalone mode now obviously not knowing of each others disk state. node0: drbdadm disconnect all node1: drbdadm disconnect all node0: drbdadm connect all node1: drbdadm connect all node0: drbdsetup /dev/drbd0 invalidate-remote node1: drbdsetup /dev/drbd0 invalidate As you can see - I've tried to keep node0 primary. The states remain the same - even after a node1 reboot again. According to my configuration file these should fix themselves up? after-sb-0pri discard-younger-primary; after-sb-1pri discard-secondary; after-sb-2pri disconnect; The latter is of course (I think!) what has happened. How can I now force one to be primary and the other secondary (or even primary but just re-sync and enter a connected state) and begin a re-sync? I don't really want to invalidate another one because this means a copy of every block from primary to secondary - and they're not (or shouldn't!) be that out-of-date! Any help appreciated. Jim Vanns -- James Vanns Systems Programmer Framestore CFC Ltd.