I have a new HP server, G3 ML350 /w 4GB RAM, which I have loaded with Exchange 2k. Our outlook clients our experiencing noticable delays when retrieving data, at time up to 45 seconds.

Do you have a virus scanner scanning the machine?  That can kill it.

What is the host? Windows 2000 Standard?

If so, you have too much RAM. Exchange cannot cope with more than 2gb of RAM without making a change to the configuration. The change is only valid on Windows 2000 Enterprise. Pull out 2gb of RAM and you will have a much happier machine.

Simon.

how many domain controllers to you have?
how many users do you have?
check the event logs on the exchange server to see if they reveal anything.

If Exchange 2k on Win2k server then you may consider using the /3GB switch in your Boot.ini file. Look this up and follow the instructions.

My OS is W2k Standard and I am using Exchange 2k standard as well which means I cannot use the /3GB switch for the extra 2gb of RAM. Last night I took out that extra 2GB of RAM and left 2GB in the system as suggested by Simon. I won't know until later today whether or not this is the solution.

Here's more information about my network
I have 5 domain controllers.

NU-FS1 (W2k Standard) is the infrastructure master and also doubles as a file server and is in the 10.2.1.x network segment

NU-PDC (W2k3 Standard) has all of the other multi-master roles. It's also a DNS, WINS and DHCP server as well as a Global catalog and is on the 10.2.1.x network segment.

NU-Exchange2k (W2k Standard) Exchange 2k standard along with being a Global catalog and is on a 10.254.1.x network segment.

NU-FS (W2k3 Standard) Future file server with RAID 5 on 10.2.x.x network segment.

NU-Library (W2k Standard) on 10.50.x.x network segment.

do the event logs indicate anything?
have you done any network tests to see if it is a networking issue and not an OS issue?

Not sure if you checked to see if another anti-virus was running other that groupshield...if yes it should be turned to just scan on a schedule and not as mcafee calls it on-access.  Another thing that sometimes happens with HP servers is that you forget to turn on or disable the raid accelerator.

It is 8:30 am Wednesday. Last night I took out 2GB and left it with 2GB. I didn't do anything else with it. I was expecting it to behave at least until 10:00am today, however, this morning I am already experiencing delays at my outlook client. Do I need to Reinstall Exchange 2k with the extra memory out in order for it to optimize to the current 2GB total?

nope, I'd check the page file size though to make sure it's consistant with the new amount of RAM

i just thought about something else.... what version of outlook do your clients have? And how are they connecting (mapi, imap, pop etc)?  Also,, what kind of delays are you experiencing,, when outlook starts, when they sent, when they recieve etc?   Also,  do you use outlook web at all and if so is that slow or just the outlook client software??

Yes I have Mcaffee VirusScan V.7.1 and I have the Exchange directory and groupshield directories excluded along with the M: drive and the inetsrv directory.

I have the Raid accelerator enabled. Should this be disabled and if so can this be done on the fly?

Raid accelerator enables is how you want it, disable the VirusScan on the machine (not groupshield) and see if it speeds up.  We had the same problem and once we disabled the computers VirusScan it sped up.

For outlook clients we use outlook 2k, xp and 2003 and OWA with Plus Pack. The delays are happening without any recongnizable pattern. I usually have my client open and either doing other things or not on the computer at all when I get the delays. It can be up to 30 seconds. If I try to do something in outlook during this time it can hang it up. I usually just wait for it to come back to life. I'm not aware of OWA having a problem.
 

I just now disabled my on access Viruscan and left the Groupshield on access running with out any changes. I'll let you know if this takes care of it.

What is the downside of not having an on-access virus scan running and leaving Groupshield to fend for itsself? Our we at any more risk in doing it this way?

Very little risk as long as you keep it updated with windows update and don't have anyone browsing the internet with the machine.  As long as you set a machine scan to run every night you should be fine.

After disabling the on access McAfee virusscan 7.10, the alerts seem to have quieted down a little. In the last 1 1/2 hours I might have only seen the alert once or twice briefly. The question that remains is even with the on-access scan disabled and still having delays, what else could be causing this on a brand new HP ML350 G3 server w/Raid "1+0" (Raid accelerator enabled) and 2GB of RAM?

I'd like to ask someone about the pagefile settings on my new exchange server. I looked on the old server that wasn't having any of these delays at the client and found that the page setting was set to an initial size of 1536mb and a maximum size of 3072mb. The old server always had 2gb of RAM while the new server had 2gb of RAM pulled out last night and has been left with 2gb just like the old server. The pagefile settings on the new server are currently set to a minimum of 2046mb and a maximum of 4092mb. Should I change this to what the old server was set to? If so can I do this with out resetting the server or should I wait until I can reset the box?

I just read an article that said I should set the initial pagefile to 1.5xRAM (3GB) and the maximum should be set to 3xRAM (6GB). My current settings are 2GB for initial and 4GB for maximum. Could this cause pagefiles to get fragmented and could this cause the system to be sluggish at times. I down loaded two utilities that were recomended for troubleshooting this problem. One is PageDefrag to check to see if the Page file is fragmented the other is Filemon which is suppose to monitor I/O activity. What road do you think I should go down to get this troubleshooted. I can easily change the pagefile settings tonight to the recommended and then reboot the server just to see what it does. I also read an article that mentioned I should have the pagefile on a seperate disk other that the boot disk. It also recommended that I avoid a mirrored partition. My configuration is using Hardware RAID "1+0" which is basically a mirrored RAID. Should I add another drive and put the pagefile on it or is that a waiste of time?

Waste of time at this point, we have 350 users on one machine and the page file resides on the boot disk and things run fine.  I read that after the fact as well.

Definitly test the pagefile, but I doubt it's the issue.

I think I might try the recommended settings of 3GB and 6GB then reboot the system. It's the only thing I can think of doing at this time. I can change it back if it makes it worse.

Exactly...