> I'm not very good at reading these logs yet, but I notice it had to kill > off a bunch of processes to get node1 to unmount. Could this have caused > it to be viewed as unstable and hence not failback? I've got the same
No, the unmount in the end succeded. > ipfail[20237]: 2010/05/19_19:09:07 info: Ping node count is balanced. > ipfail[20237]: 2010/05/19_19:09:07 info: Giving up foreign resources > (auto_failback). > ipfail[20237]: 2010/05/19_19:09:07 info: Delayed giveup in 4 seconds. > ipfail[20237]: 2010/05/19_19:09:07 info: Aborted delayed giveup (7) I am sorry, I don't know what does it mean. You could check the logs from peer node. Or it could be caused by use of ipfail instead of pingd. I don't know. > >> If we say Heartbeat V1 config style is obsolete, use of ipfail is >> obsolete absolutely. If you don't want to move to Pacemaker, try to >> use pingd at least. >> > > Should I be using Pacemaker? It seemed rather more complex to me, so I > went with heartbeat. Yes, Pacemaker is your choice. It could be seen as a bit complex for the first sight. But as heartbeat V1 is not supported anymore, you don't have any other option if you will have to deal with some issues (like this one could be). The sooner the better. Documentation for first-timers is briliant now, do not worry. Check "Cluster from scratch" Tino _______________________________________________ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
