[Linux-HA] crmd core dumping, is this normal?

Robert Lindgren robert.lindgren at gmail.com
Mon Sep 24 08:17:00 MDT 2007


Hi All,

Running the latest interim packages for Ubuntu, I get this error:

crmd[5918]: 2007/09/24_15:29:44 ERROR: crmd_ccm_msg_callback:
Membership instance ID went backwards! 8->2crmd[5918]:
2007/09/24_15:29:44 ERROR: crm_abort: crmd_ccm_msg_callback: Triggered
fatal assert at callbacks.c:520 : current_ccm_membership_id <=
membership->m_instance
heartbeat[4325]: 2007/09/24_15:29:44 WARN: Exiting
/usr/lib/heartbeat/crmd process 5918 killed by signal 6 [SIGABRT -
Abort].
mgmtd[5917]: 2007/09/24_15:29:44 ERROR: crm_log_message_adv:
#========= cib:cmd message start ==========#
heartbeat[4325]: 2007/09/24_15:29:44 ERROR: Exiting
/usr/lib/heartbeat/crmd process 5918 dumped core
heartbeat[4325]: 2007/09/24_15:29:44 ERROR: Respawning client
"/usr/lib/heartbeat/crmd":
heartbeat[4325]: 2007/09/24_15:29:44 info: Starting child client
"/usr/lib/heartbeat/crmd" (104,110)
ccm[4421]: 2007/09/24_15:29:44 info: client (pid=5918) removed from ccm
tengine[6746]: 2007/09/24_15:29:44 ERROR: subsystem_msg_dispatch: The
server 5918 has left us: Shutting down...NOW
pengine[6747]: 2007/09/24_15:29:44 ERROR: subsystem_msg_dispatch: The
server 5918 has left us: Shutting down...NOW


The scenario to which i got this error was: removing eth0,eth1 and
serial cable from node 2. When cables are unplugged on node2, it for
some reason starts resources (even though pingnodes are dead, this
might be my problem I guess). When reconnected HA finds out that both
nodes are running the services and the error occurs. Eventually the
service is started on the correct node (node1), and drbd is messed up
on node2.

Any hints? And tell me if more logs are needed.

BR
Robert Lindgren



More information about the Linux-HA mailing list