Description of problem:
In a two node cluster setup, when trying to stop clustering 
using "service clumanager stop" or when shutting down the server, 
clumemd hangs pretty frequently and never times out. As a result, 
clumanager does not stop on the system on which this happens. The 
services failover to the passive system, but on the current system we 
have to kill the process manually to allow the system to shutdown or 
stop the clumanager.



Version-Release number of selected component (if applicable):
clumanager-1.2.9-1


How reproducible:
very frequently - not always.

Steps to Reproduce:
1.Setup clustering with two nodes - one active and one passive.
2.Stop clustering on the active system using "service clumanager 
stop" or through the redhat-config-cluster GUI interface.
3.Run ps -ef.
4.Three clumemd threads are waiting on something and never stop or 
timeout - the cumanager stop consequently hangs.
    
  
Actual results:
root      9876     1  0 10:38 ?        00:00:00 /usr/sbin/clumembd
root      9877  9876  0 10:38 ?        00:00:00 /usr/sbin/clumembd
root      9878  9877  0 10:38 ?        00:00:00 /usr/sbin/clumembd
root     29812 14862  0 11:28 pts/4    00:00:00 /bin/sh /sbin/service 
clumanager stop
root     29815 29812  0 11:28 pts/4  
00:00:00 /bin/sh /etc/init.d/clumanager stop



Expected results:
clumemd process stops cleanly and shuts down clustering

Additional info:

strace outputs for the clumemd processes:

#strace -p 9878
Process 9878 attached - interrupt to quit


# strace -p 9877
Process 9877 attached - interrupt to quit
waitpid(9878, 

# strace -p 9876
Process 9876 attached - interrupt to quit