2008/4/30 Andrew Beekhof <[EMAIL PROTECTED]>: > On Wed, Apr 30, 2008 at 11:28 AM, Dennis Kern <[EMAIL PROTECTED]> wrote: > > Hi, > > > > I use Heartbeat 2.1.3 with DRBD 8.0.11 > > > > The configuration for a two node NFS file server works very well, > > except split brain handling. > > > > I have no power device for STONITH and I didn't found any hint in the > > web for handle split brain with a 2 Node Cluster without STONITH. > > > > What is the "best" solution for me to avoid split brain? Is there no > > mechanism like "ping demon". I mean if ping ist not possible (either > > to local LAN, crossover connection to the other node), than the node > > can go down? > > no. nodes dont shut themselves down based on pingd or anything else. > > > > Or did I missunderstood something? > > > > Would it be possible: If split brain is detected and the node who not > > run the resources actual shut down? > > not really. the only correct and reliable solution is STONITH. > things like pingd can help but can still be fooled if you're > particularly unlucky. > > unless you have shared storage. in which case there are a couple of > tricks you can do.
And what kind of tricks are these? DRBD is not a "must have". > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > Dennis _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
