We have three servers (beyond my development box) in our organization.
The first is a testing server that has IIS and SQL Server on it. The
second is another testing server that also has IIS and SQL Server. The
final is the production box that only has IIS. I develop on my own
machine, copy to the first testing server, the code is tested, copy the
code from testing 1 to testing 2, the code is tested again, and then
copy testing 2 code to production. Right now I'm in the process of
moving my revised code to the second testing server. The code is
running very slow on that server, even though it ran just fine on the
first testing server, and I cannot figure out why.
All hits (first and subsequent) are taking up to 20 seconds or longer.
Then sometimes the pages start to time out. In fact, pages that do not
access the database are just as slow. Even if I copy just an html page
(with an aspx extension) it runs slow. But if I remove all aspx and
dll files and leave just my test html page the page responds very
quickly. If I leave the code and rename my test page with an html
extension that page renders quickly.
I decided copy old code from the production back to the second testing
server and that code is running slow, too (even though production
currently runs very quickly). Of course, when someone tested that code
last month the second testing server was running just fine. To the
best of my knowledge nothing major was changed on the testing 2 box
beyond Microsoft updates.
I turned on tracing and viewed all ASP.NET/ASP.NET Applications
counters in perfmon. Tracing times were below a second but the request
execution time was slow. Total number of CLR Exceptions seems to be a
little high, which is unusual since I don't see anything on screen or
in the Event Log where I am logging all thrown exceptions. Nothing
else in perfmon stood out. The processor is not even spiking.
We even have an old ASP application on the server and it runs as fast
as usual.
I'm starting to run out of ideas for ways to diagnose this. Can anyone
offer any ideas to help? Should I try looking at different perfmon
counters? If so, what should I look for?
