M
MSDN
Hi,
We are running a webfarm of server 2003 web edition with SQL2000 enterprise
running on 2003 in a failover cluster.
Periodically we suffer from performance problems where the websites slow to
the point of not being available.
The websites are running .Net 1.1 - we are not in a position to upgrade to
..Net 2 for all sites yet.
Today we experienced a period of about 1 hour when our SQL Server was
severely affected by a repeated calling of the stored procedure
TempGetStateItemExclusive2. Having monitored the server there doesn't appear
to be any explanation for why the frequency of calls to this procedure
increased so drastically over the period of time involved. The number of
users, transactions/sec etc where within normal ranges. No other performance
bottlenecks exist except when TempGetStateItemExclusive2 is called
excessively.
Having looked around various forums all I can find is people reporting the
same problem but with no replies indicating a possible solution.
Is there a fix to this? or a workaround? or even an alternative?
TIA
Phil
We are running a webfarm of server 2003 web edition with SQL2000 enterprise
running on 2003 in a failover cluster.
Periodically we suffer from performance problems where the websites slow to
the point of not being available.
The websites are running .Net 1.1 - we are not in a position to upgrade to
..Net 2 for all sites yet.
Today we experienced a period of about 1 hour when our SQL Server was
severely affected by a repeated calling of the stored procedure
TempGetStateItemExclusive2. Having monitored the server there doesn't appear
to be any explanation for why the frequency of calls to this procedure
increased so drastically over the period of time involved. The number of
users, transactions/sec etc where within normal ranges. No other performance
bottlenecks exist except when TempGetStateItemExclusive2 is called
excessively.
Having looked around various forums all I can find is people reporting the
same problem but with no replies indicating a possible solution.
Is there a fix to this? or a workaround? or even an alternative?
TIA
Phil