Bookmarks

Yahoo Gmail Google Facebook Delicious Twitter Reddit Stumpleupon Myspace Digg

Search queries

procmail modify subject, procmail alter subject, formail get body of message, www.xxxcon, www xxxcon, board.issociate.de, registrieren fehler dll server 0x8002801c, far2 vs near2 raid10, access abfragen aus linux, "Can't send mail: sendmail process failed with error code 69"

Links

XODOX
Impressum

#1: TCPv4 Counters "Connetions Active" and "Connections Passive" on W2K3IIS 6 Server

Posted on 2005-01-07 22:38:00 by themeanies

Our main internal IIS server has been experiencing a slowdown recently
due to load. This is an internal fully patched Win2K3 server running 7
asp sites. Only one site is heavily accessed and it is database
connection intensive. Both the IIS server and SQL server are dual
gigabit teamed connections. Bandwidth between the two servers is not a
problem.

Specifically my question is about the perfmon counters for TCPv4. The
following two counters are slowly increasing post re-boot. Is this
expected behavior?

Connection Active and Connection Passive at a rate of about 50 and 200
per minute respectively.

The code is very lean, but like I said very transactional.

Do you have any suggestions for working through growing pains like this?

tM

Report this message

#2: Re: TCPv4 Counters "Connetions Active" and "Connections Passive" on W2K3 IIS 6 S

Posted on 2005-01-10 17:42:16 by Joris Dobbelsteen

I believe the active and passive are similar to the FTP used names.
Active is the number of connections that are initiated by the system.
Passive is the number of connections that are received by the system.

They should indeed be increasing. For the assumption I guess you are using
monitoring the IIS Server. So Active is an indication for the connections to
the SQL server (or other services). Passive is the number of received
connections (200 client connections / minute).



Check IIS Server for high loads on the CPU.
Might be that the application is CPU very intensive.
How about the bandwidth to the clients?

For the SQL server some indications might be:
Monitor Disk for high loads, these may indicate problems here
Monitor CPU, high loads indicate problems in this area. Might be caused by
any number of problems probably you should look to optimize the queries or
indexes. Worse designed databases can easily eat a server.
SQLServer:SQL Statistics -> Batch Requests/sec is an indiction for the
workload.
SQLServer:Locks -> Lock Waits/sec or ... -> Lock Wait Time (ms). Increasing
values indicates that active queries are waiting for other queries. Heavy
locking will reduce performance.
Also monitor database size (is it growing?)

Also database size, queries and indexes can make huge differences. I had a
database that filled up and a query took 3-5 minutes to complete, some
tuning reduced the time to 30 seconds and I believe it also scales much
better for larger data sets.

If the problem is the SQL Server you can also use the SQL Server "Profiler"
to see what queries are costing most CPU/Disk reads or writes. Capturing
takes performance of the SQL server, but you can afterwards run the Index
Optimizer Wizard and let it suggest some changes (run the wizard when its
quite, it takes some performance too).
You can do the trace also on the server, storing it in a file. SQL Profiler
can create the script to do so. This reduces the performance hit. Later you
can load the file into the SQL profiler and analyze it.

Maybe the application can be optimized and reduce the load significantly...

But first find out what the problem is...

- Joris

"themeanies" <themeanies@nowhere.net> wrote in message
news:10tu09pe9pqkp75@corp.supernews.com...
> Our main internal IIS server has been experiencing a slowdown recently
> due to load. This is an internal fully patched Win2K3 server running 7
> asp sites. Only one site is heavily accessed and it is database
> connection intensive. Both the IIS server and SQL server are dual
> gigabit teamed connections. Bandwidth between the two servers is not a
> problem.
>
> Specifically my question is about the perfmon counters for TCPv4. The
> following two counters are slowly increasing post re-boot. Is this
> expected behavior?
>
> Connection Active and Connection Passive at a rate of about 50 and 200
> per minute respectively.
>
> The code is very lean, but like I said very transactional.
>
> Do you have any suggestions for working through growing pains like this?
>
> tM

Report this message