Today we encountered some performance problems at our customer site. After checking with the customer it seemed that especially the CIFS shares were having problems. At first CIFS was getting slower until the shares were not even accessible anymore.
Restarting the CIFS on the filer did solve the problem for a few minutes, but within half an hour the the problems were back again.
Checking the /etc/messages file on the filer showed the following error
Tue Sep 9 16:00:08 CEST [XXX0011:cifs.stats.pBlkExhaust:info]: CIFS: All CIFS control blocks for the STANDARD pool are in use. The request for a new control block can not be granted.
We solved this problem by breaking the oplocks using the lock break -p cifs command.
After that the CIFS shares became available again for the customer.