issues with our web servers

We deployed some new code and we dont know whats causing them to behave odd
after a few hours.

The site stops to function and then a reset of the IIS Service fixes it for
a while and then it goes back to misbehaving.

We dont wish to rollback, but at the same time, we want to know the root
cause behind the unresponsiveness after a while. We dont want to speculate..
Nothing really conclusive in the error log.

So the question is where else can I look and what else can I do to track it
down ? What perfmom counters should I look at ? The CPU looks fine..i.e
under 10%.

We are using IIS 6.0

Re: issues with our web servers

On Nov 28, 10:32 pm, "Hassan" wrote:
> We deployed some new code and we dont know whats causing them to behave odd
> after a few hours.
>
> The site stops to function and then a reset of the IIS Service fixes it for
> a while and then it goes back to misbehaving.
>
> We dont wish to rollback, but at the same time, we want to know the root
> cause behind the unresponsiveness after a while. We dont want to speculate..
> Nothing really conclusive in the error log.
>
> So the question is where else can I look and what else can I do to track it
> down ? What perfmom counters should I look at ? The CPU looks fine..i.e
> under 10%.
>
> We are using IIS 6.0


I would first figure out whether "stop functioning" is a crash or
hang. Use DebugDiag for this.

Armed with the stack backtrace of the crash/hang, look at the code
identified as the culprit. And go from there.

Just monitoring perf counters is not going to work. There's a million
things to monitor, and one or none of those things may indicate the
actual issue. You want to use DebugDiag to get a smoking gun culprit
to narrow down what to look at, and then go to perf counters depending
on situation.


//David
http://w3-4u.blogspot.com
http://blogs.msdn.com/David.Wang
//

Re: issues with our web servers

So this tool will tell me if its a crash or a hang ? I believe its more of a
hang as for some reason, the requests in the queue start climbing and we see
some requests being processed but very slowly and some being timed out..

Also from a memory perspective, would there be memory issues if say the w3wp
service is using around 1GB even if the server which has 2GB in total may
have some memory left.

This is an x64 bit windows 2003 server as well. Do 64 bits use more
memory,etc. ?


"David Wang" wrote in message
news:679bf031-fc21-4a8e-b09b-6ad9bfd8c980@s36g2000prg.google groups.com...
> On Nov 28, 10:32 pm, "Hassan" wrote:
>> We deployed some new code and we dont know whats causing them to behave
>> odd
>> after a few hours.
>>
>> The site stops to function and then a reset of the IIS Service fixes it
>> for
>> a while and then it goes back to misbehaving.
>>
>> We dont wish to rollback, but at the same time, we want to know the root
>> cause behind the unresponsiveness after a while. We dont want to
>> speculate..
>> Nothing really conclusive in the error log.
>>
>> So the question is where else can I look and what else can I do to track
>> it
>> down ? What perfmom counters should I look at ? The CPU looks fine..i.e
>> under 10%.
>>
>> We are using IIS 6.0
>
>
> I would first figure out whether "stop functioning" is a crash or
> hang. Use DebugDiag for this.
>
> Armed with the stack backtrace of the crash/hang, look at the code
> identified as the culprit. And go from there.
>
> Just monitoring perf counters is not going to work. There's a million
> things to monitor, and one or none of those things may indicate the
> actual issue. You want to use DebugDiag to get a smoking gun culprit
> to narrow down what to look at, and then go to perf counters depending
> on situation.
>
>
> //David
> http://w3-4u.blogspot.com
> http://blogs.msdn.com/David.Wang
> //