I've got an a rare problem with ldirectord. Every now and again the perl
process just quits. I see nothing in the ldirectord logs nor system logs. It
seems to occur if I am taking real servers on and off of line frequently.
The net effect is that the cluster just stops. heartbeat is still running but
it hasn't realised that its ldirectord resource has vaporised - hence I get
no failover.
I'm on RH9 w/ all the latest heartbeat, ldirectord and kernel patches.
What can I do to debug / address this issue. Should I be looking into DJB's
daemontools or similar ...
Any help / pointers greatly appreciated.
Cheers,
James
|