We have a cluster with SLES 10 SP4 / OES2 SP3, and last week we had a power outage that took both nodes to an unclean reset.

After that, several users (not all) complain that access to novell shares is too slow.

Next time I tried to switch cluster nodes, it got about 4 hours to do it. This message was flooded to the syslog:

"Feb 2 21:51:24 oes2-server kernel: Waiting for 2927353 Purging beasts to unlink"

The same messages were also sent to dmesg; each time the message was sent, the number was decremented. To process repeated for each clustered volume, starting with different numbers.

After that, the system is not as slow as before, but it's still slower than we remember of it.

Oh, and google sends me to TID 7009350 (NSS pools taking a long time to deactivate), which is not the same issue, and not solved my problem.