High CPU time per user as soon as they connect?

On a couple of our servers, users are currently scoring very
high on "CPU Time" as reported by Enterprise Manager and by
sp_who2, both as soon as they log in, and afterwards -
5,000,000 at login and 47,000,000 later on are the top figures.
Most of the fleet is on SQL Server 2000 Service Pack 2, with
between 2 and 8 hard-working processors. The hardest working
server is on Service Pack 3(a), clustered, and its top 5
current values in CPU time are between 382,847 and 705,879.
The servers are all on clients' sites and networks, with client's
choice of firewall. Usual application is a Java interface using
jTDS, but CPU time is also reported high for a new Query Analyzer
connection. SQL Server, not Windows, accounts are most often used.
Clients aren't complaining of poor performance, apparently, but
maybe they already gave up phoning for that.
So, should we worry? Reboot? Reindex? Patch? Disinfect?
Or ask around, "Who left SQL Profiler running"...i would run profiler to find out whats going on.
i would also run perfmon to find out what average cpu is
over the entire day
otherwise, you have nothing from which to make reasonable
assessments
>--Original Message--
>On a couple of our servers, users are currently scoring
very
>high on "CPU Time" as reported by Enterprise Manager and
by
>sp_who2, both as soon as they log in, and afterwards -
>5,000,000 at login and 47,000,000 later on are the top
figures.
>Most of the fleet is on SQL Server 2000 Service Pack 2,
with
>between 2 and 8 hard-working processors. The hardest
working
>server is on Service Pack 3(a), clustered, and its top 5
>current values in CPU time are between 382,847 and
705,879.
>The servers are all on clients' sites and networks, with
client's
>choice of firewall. Usual application is a Java
interface using
>jTDS, but CPU time is also reported high for a new Query
Analyzer
>connection. SQL Server, not Windows, accounts are most
often used.
>Clients aren't complaining of poor performance,
apparently, but
>maybe they already gave up phoning for that.
>So, should we worry? Reboot? Reindex? Patch?
Disinfect?
>Or ask around, "Who left SQL Profiler running"...
>.
>