In one of our SQL Server's cluster i faced a problem on a named instance.
That particular instance (nr. 5) it was normally working and having already allocated the "normal" memory (about 1, 7 Gb) for it. Suddenly without any particular problem/error it starts to hang, and the users start to have dreadful response times!!!
We have done allot of debugging in the front-ends, in the networking, and the results pointed for the SQL cluster. At the cluster level we start to have for that particular instance, 40%-50% of occupation in the 8 CPUs!!!
It wasn’t supposed to work like that, but strangely only after a take off-line fallowed by bring on-line we solve the problem.
Now 24h have passed it have already reached the maximum memory and everything’s look fine.
The configuration is:
SO: Win2003 EE
DBE: SQL2000 SP3a
Boot.ini: /PAE (for using AWE on other named instance)
It seems that SQL Server Engine wasn’t reallocating the memory needed for the new processes. I would appreciate if any one can give me some information, on this subject and how to detect and identify it.