On 06/07/2013, at 1:28 AM, Andrii Moiseiev <amoise...@gmail.com> wrote:
> Any clues? Not with only fragments of the logs Do you have a log file (not syslog) configured? I'll need that complete file If you could also set PCMK_trace_function=apply_xml_diff (see http://blog.clusterlabs.org/blog/2013/pacemaker-logging/ for details), that would be very helpful. > Also, it broke 'default' fencing. Previously, even with stonith-enabled: > false, pacemaker was trying to kill cman / corosync if connection is lost or > split brain occurs, Technically its cman asking for the node to be fenced, Pacemaker is just following orders. > but now it's not happening: I don't see how you can say that. This > Jul 5 09:54:25 devpacemaker01 crmd[20840]: notice: tengine_stonith_notify: > Peer devpacemaker03_eth1 was not terminated (reboot) by devpacemaker02_eth1 > for devpacemaker02_eth1: No such device > (ref=1fc11b87-529d-4f6c-b4e6-ffaa82c06bd8) by client stonith_admin.cman.8832 very clearly shows Pacemaker trying to shoot the node. Granted it does not succeed, but it is trying to. _______________________________________________ Pacemaker mailing list: Pacemaker@oss.clusterlabs.org http://oss.clusterlabs.org/mailman/listinfo/pacemaker Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
