outlooks 2003 users disconnecting from exchange server

infrastructureadmin
infrastructureadmin used Ask the Experts™
on
We have approximately 300 users connected to a MS Exchange 2003 Enterprise server to several different information stores. We are randomly getting disconnected throughout the day. Users with larger inboxes (greater than 1 GB) are more noticable.
Email is slower than usual (sending an email stays in the outbox for approximately 60 seconds)
The D Drive, located on a san, is heavily fragmented (currently 99% ) and has 50% of its available space free  (400 gb used of 800 GB total)
There are no errors in the exchange server or client machines
Is it possible that the fragmentation on the drive is causing the slowness?

Recent changes to the server
Added OMA access 2 weeks ago
Added 10 users with email storage of approximately 1.5 GB each.
The largest information store is 121 GB
There are no networking issues/ping times are normal
We notice when users remove "rules" from outlook the disconnects are less frequent
When users are in cached mode (only 50% have outlook 2003) the problem does not appear

Comment
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
looks like you answered your own questions, limit mailboxes to 1 GB on most groups, make an exception group, make a cached mode policy to force as many users to use it as possible, and defrag.

I'd probably defrag the message stores and remove white space also.

Sorry for being captain obvious, but I'd resolve those obvious Issues before I got too worked up finding alternative answers,

and while I'm demonstrating my amazing command of the Obvious, backup backup backup.
Jon BrelieSystem Architect

Commented:
I disagree.  We have a few thousand users and average mailbox size is about 2-3 gb.  Some are huge, some not so much. With 2003 it really depends more on the number of disks you have supporting your mail stores.

Are you noticing the issue with users on one specific store more than another?

Author

Commented:
The larger stores do seem to have the issues, but most users are in the larger ones too so it is hard to tell.

The odd thing seems to be is the users with a lot of rules seem to really choke up outlook and disconnect them.
Ive done exchange for 10 years or so and never had to do a physical defrag, but this is the largest server I have ever managed. I did a white space offline ESEUTIL 6 months ago.

Starting with Angular 5

Learn the essential features and functions of the popular JavaScript framework for building mobile, desktop and web applications.

Top Expert 2010

Commented:
hi
Can you download E 2003 BPA and run a health scan.
See what comes-up ?

http://www.microsoft.com/downloads/details.aspx?FamilyID=dbab201f-4bee-4943-ac22-e2ddbd258df3&displaylang=en

thanks

Author

Commented:
I ran a health scan and everything comes up normal. Im starting to think there is a physical problem.
I did a basic file copy of a 1 gb file to the data drive and it did well, and a network copy to the server and it did well also.

Author

Commented:
The mailbox store 'UserStore' on server 23EXG01 is 147 GB in size. It is recommended that the maximum database size does not exceed 100 GB.

If I have enterprise exschange 2003 version can I exceed 100 gb?

Author

Commented:
Looks like enterprise is unlimited. Checked the San and the throughput is ok

Last week we added roughly 20 users connecting from London (147 ms ping time) over a site to site VPN. They are using cached mode with outlook. Could this cause an issue?
Top Expert 2010

Commented:
Hey
Sorry i couldnt get to this earlier. EE was freezing up on me and didnt show a submit button. Tried for last couple of hours.

Back to the qn.
The users who are getting disconnected - are they from London ?
AD/Exchange over VPN - we get to a lot of MTU issues.

We were working on this case and tried to check this from every angle and in the end it came down to MTU size of VPN packets / packets defragmenting.

See this and the accepted soln from ChiefIT
http://www.experts-exchange.com/Software/Server_Software/Email_Servers/Exchange/Q_26284889.html?cid=748#a33091465

Will be worth exploring if the users slowing down are the one's from london.

Author

Commented:
We have an HP Storage Works 500. When we get the disconnects I am seeing really high disk writes in perfmon. I disconnected the London office and the disk writes/disconnects didn't stop.
So I am certain it's a disk problem at this point.. The drive appears to be in good shape. Not sure what to do next.
Top Expert 2010

Commented:
Did you check your RAID performance ?
do you have any tools to monitor DISK IO ?

Check Exchange JetStress
http://www.microsoft.com/downloads/details.aspx?familyid=94b9810b-670e-433a-b5ef-b47054595e9c&displaylang=en

and this one
http://www.microsoft.com/downloads/details.aspx?familyid=7F944850-945F-4E60-B6D6-CF7341D7F9C3&displaylang=en

--
Can you fire-up perfmon and get some basic disk counters.

what process is causing the high disk writes in perfmon ?
Jon BrelieSystem Architect

Commented:
How many disks/spindles do you have supporting your mail store?  What is the RAID configuration?
Actually I ran the ExTra (Exchange Troubleshooter) and found the antivirus software was hammering the RPC queue. After disabling the AV services performance was at its best.
I am working on reinstalling GroupSheild at the moment.

Thanks for all your input guys!
Top Expert 2010

Commented:
Great !!

Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial