Hi, another clue perhaps, I have the same problem with a 3 nodes cluster.
And another one : if I start Pacemaker so that DC is the second node node 2 , with the same stonith primitives configuration , and if I power off the first node node1, node1 is correctly fenced by the restofencenode1 from node4, without ipmi IP address mismatch. So the ipmi IP address mismatch seems to depend also of where is the DC node with regard to the stonith resource configuration like mine. If I start Pacemaker so that DC is on the third node and I power off the second node node2, the restofencenode2 which is expected to fence the node2 is in fact fencing the node 1 ... I can see again in the syslog : 1286953669 2010 Oct 13 09:07:49 node3 daemon info stonith-ng [11658]: info: log_data_element: process_remote_stonith_exec: ExecResult <st-reply st_origin="stonith_construct_async_reply" t="stonith-ng" st_op="st_notify" st_remote_op="6c2a04b2-b58f-4a9b-b3eb-3bba388a92fe" st_callid="0" st_callopt="0" st_rc="0" st_output="Rebooting machine @ IPMI:<BMC-IP-node1>...Done whereas it should have been <BMC-IP-node2> Does anybody uses fence_ipmilan with Pacemaker on RHEL6 on more than 2 nodes clusters ? And could perform such tests of fencing ? Thanks Alain _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
