Conference amcucs::ms-sqlsvr

336.0. "SQL 6.5 going slower and eventually dies on 4100" by JOBURG::NICOP (Nico Pretorius) Mon Feb 10 1997 23:04

Hi,

I have a situation where a customer is running SQL 6.5 SP2 on NT3.51 SP5
on a 4100.  The AlphaBios is on 5.28 and we're using REV. C of the 
NT 3.51 HAL.

When the customer starts quering their database, the SQL query seems to 
run fine with the first query, but on subsequent queries SQL just gets 
slower and slower and eventually the SQL Service goes into a hung state 
and you cannot stop the Service. This happens within a minute or so after
they start with queries. At first I thought it might be something specific
to their application, but even if you just do a simple DBCC MEMCHECK you 
end up in the same situation. If you manage to do a query the SQL 
performance seems to be slow.

The rest of the O/S is fine but SQL is dead and the only 
way to get it going again is to restart the machine.

There are no errors in the event log.

The machine has 256MB of RAM, with 100MB of that allocated to SQL, a 
single 300Mhz cached processor, 26GB disk space (1 x RZ28B-VW and 
5 x RZ29-VW). The RZ29 disks are configured with NT software striping.

Does anybody have any ideas or seen this before?

Thanks for any replies.

Regards
Nico

    
    What queries are being done? 
    
    Are they through a client?  What kind of client?  
    
    If you are having a hard time finding what is being sent to 
    SQL, turn on SQL Trace and watch what is going on. 
    
    Have you watched anything through Performance Monitor? 
    
    Thanks. 
    
    Mary Jo