VCSMM kernel module had not been reloaded after patch installation: modinfo |grep vcsmm revealed that, on the problem node, VCSMM kernel module version 5.1 SP1 RP2 was still loaded (compared to the working nodes, where 5.1 SP1 RP2 P1 had been loaded).
1. Unload VCSMM / LMX kernel modules:
# modunload -i
( where
2. Check the file /etc/vcsmmtab and make sure it is identical across all cluster nodes:
# cat /etc/vcsmmtab
/sbin/vcsmmconfig -c > /var/VRTSvcs/log/vcsmmconfig.log 2> &1 &
3. Restart VCSMM / LMX to reload relevant kernel modules:
( on Solaris 9 )
# /etc/init.d/vcsmm start
# /etc/init.d/lmx start
( on Solaris 10 )
# svcadm enable vcsmm
# svcadm enable lmx
Applies To
Solaris systems running Storage Foundation for Oracle RAC
Mar 23 00:15:14 node1 gab: GAB INFO V-15-1-20036 Port o gen 1172b06 membership 0123
Mar 23 00:15:14 node1 vcsmm: Warning: RFSM: received (VRFSM_TK_CONFIG_STATE) from node 3. Local RFSM state: 8
Mar 23 00:15:14 node1 vcsmm: Warning: RFSM: message RFSM version 1, local RFSM version 1
Mar 23 00:15:14 node1 vcsmm: Warning: RFSM: message CLIENT version 10, local CLIENT version 20
Mar 23 00:15:14 node1 vcsmm: Warning: RFSM: Dropping the message.
Mar 23 00:15:14 node1 vcsmm: Warning: RFSM: received (VRFSM_TK_STOPPED) from node 3. Local RFSM state: 8
Mar 23 00:15:14 node1 vcsmm: Warning: RFSM: message RFSM version 1, local RFSM version 1
Mar 23 00:15:14 node1 vcsmm: Warning: RFSM: message CLIENT version 10, local CLIENT version 20
Mar 23 00:15:14 node1 vcsmm: Warning: RFSM: Dropping the message.
Mar 23 00:15:19 node1 gab: GAB INFO V-15-1-20036 Port o gen 1172b07 membership 012
Also, /sbin/vcsmmconfig -W shows discrepancies among the nodes:
( on working nodes )
# vcsmmconfig -WVCSMM Protocol Version Information:
Supported Protocol Versions(s) : 10, 20
Current Protocol Version : 20
( on problem nodes )
# vcsmmconfig -WVCSMM Protocol Version Information:
Supported Protocol Versions(s) : 10
Current Protocol Version : 10