[Linux-cluster] Running MySQL under cluster-suite

[Matthew Geier]

  I'm trying to set up an active/passive failover for MySQL and have a 
problem with getting cluster suite to start mysql. It just keeps telling 
me to 'look in the logs' for a problem that isn't there.

  The main issue seems to be that it decides that MySQL has failed and I 
can't 'disable' it to 'enable' it. The disable command fails 'cause 
MySQL isn't actually running anymore.


Oct 17 17:00:40 Aristotle1 clurgmgrd[7495]: <notice> Stopping service MySQL
Oct 17 17:00:40 Aristotle1 clurgmgrd: [7495]: <info> Executing 
/etc/init.d/mysqld stop
Oct 17 17:00:40 Aristotle1 mysqld: Stopping MySQL:  failed
Oct 17 17:00:40 Aristotle1 clurgmgrd[7495]: <notice> stop on script 
"MySQL" returned 1 (generic error)
Oct 17 17:00:40 Aristotle1 clurgmgrd: [7495]: <info> /dev/dm-9 is not 
mounted
Oct 17 17:00:45 Aristotle1 clurgmgrd[7495]: <crit> #12: RG MySQL failed 
to stop; intervention required
Oct 17 17:00:45 Aristotle1 clurgmgrd[7495]: <notice> Service MySQL is failed


  The only way I can get past this is manually mount the database file 
system, manually create the IP address and start MySQL manually - then 
tell the cluster manager to disable MySQL. Only even then it gets in a 
wierd state and one let me re-enable MySQL again - only this time it 
doesn't write anything useful in the logs.

  Any one have any pointers ?. I'm not trying to load balance or 
anything, just having a standby machine to take over if the first fails.