This bug has been copied from bug #599654 and has been proposed to be backported to 5.3 z-stream (EUS).
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: Previously, the Closed Process Group (CPG) interface returned the wrong result which could lead to incorrect behavior in some situations. With this update, the CPG interface behaves now as expected.
killing (-9) groupd now correctly shows as PROCDOWN (and triggers node removal and fencing): Nov 1 10:21:47 z3 groupd[30921]: kill node 2 - groupd PROCDOWN Nov 1 10:21:57 z3 openais[30908]: [TOTEM] The token was lost in the OPERATIONAL state. Nov 1 10:21:57 z3 openais[30908]: [TOTEM] Receive multicast socket recv buffer size (320000 bytes). Nov 1 10:21:57 z3 openais[30908]: [TOTEM] Transmit multicast socket send buffer size (262142 bytes). Nov 1 10:21:57 z3 openais[30908]: [TOTEM] entering GATHER state from 2. Nov 1 10:22:17 z3 openais[30908]: [TOTEM] entering GATHER state from 0. Nov 1 10:22:17 z3 openais[30908]: [TOTEM] Saving state aru 8774 high seq received 8774 Nov 1 10:22:17 z3 openais[30908]: [TOTEM] Storing new sequence id for ring 94 Nov 1 10:22:17 z3 openais[30908]: [TOTEM] entering COMMIT state. [...] Nov 1 10:22:18 z3 fenced[30927]: fencing deferred to z1
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2010-0828.html