Jump to content

Recommended Posts

Posted

Ok, after our mail storm the other day we've got a new exchange 2010 server running in co-existance with our 2003 server and connections are horrendously slow.

 

The new server itself seems fine, it sat using about 90% ram at all times but that is understood to be a normal situation. It is on a VMware ESXi box

 

The settings are as follows:

4096mb RAM

2vsockets/4cores/8totalcores

C: system drive on a 10k array 35gb

E: database drive on a 15k array 300gb

1 network card

 

the users complain of delays of up to 5 minutes when performing operations within outlook, at no point does anything actually time out/error but the informational "outlook is trying to request data from the exchange server" pops up occasionally. The actual operation doesn't seem to make a difference - for example sometimes swapping between delegated mail boxes takes 5 minutes, sometimes it is instant, sometimes opening a message takes 5 minutes, sometimes it is instant, etc. The average responses in the connection status box for directory operations to the exchange server seem quite slow, around 5k-15kms, operations to the DCs take 0-60ms.

 

I've run the performance analyser. I'm getting errors as follows:

 

! top 6 users account for 99% of MAPI cpu usage

 

X Potential issue with RPC activity found

X The Exchange Server User Monitor (ExMon) RPC data indicates that clients are performing online operations. Online operations for MAPI operation "QueryRows" account for 50.62% of the processor usage devoted to processing RPC requests. When a single MAPI operation type is responsible for a large percentage of the load, this may be an indication that the user is running an application that is behaving poorly.

 

X Slow Extensible Storage Engine Calls

X Slow Calls to Jet: Long wait time in function call to Jet: A function call to 'JetOpenTable' took longer than the threshold value of 2.5 seconds to complete. The call took 5.21 seconds. This delay occurred at 08/28/2013 10:00:45. The calling thread ID was 6692.

 

Ideas?

Posted
Hmmm, that does seem a little short on RAM IMO, but I'm fairly new to Exchange so no expert. For 1000 mailboxes I'm seeing 10GB usage (that obviously includes caching) and very few are in use currently. Our Exchange server has 32GB of RAM - I know it will use all you can throw at it.
  • Thanks 1
Posted

Ok, i've given it 16, it's using 10 and it is still pretty bloody slow. :(

 

I've also got an notice coming up on the performance troubleshooter to say that the temp and tmp folders are on the system drive. Now i've followed the instructions and set the environment variables and rebooted but it is still telling me they are on the system partition. Now, I've had a look in the registry and only the sid for the administrator are saying that temp and tmp are on E:\, is it safe for me to manually change all the temp/tmp settings in regedit or am I going to break stuff?

Posted (edited)

I meant SP3. There is a fix in there that solves a deadlock problem for mail delay. Long delay in receiving email messages by using Outlook in an Exchange Server 2010 environment

 

Also Outlook SP3 because it contains this fix: "Outlook is retrieving data from the Microsoft Exchange Server" message after you upgrade to Outlook 2010

 

Also check that your users running in cached mode? Running in 'online' mode is another cause of performance problems for Outlook/Exchange 2010.

 

Similarly, if you run in cached mode, but your users move between machines, or you have mandatory profiles, every time they logon the cache has to be rebuilt, and a few tens of users with large mailboxes can really harm overall performance in this scenario.

 

What performance do they see when using OWA?

Edited by psydii
  • Thanks 1
Posted

OWA is sluggish nut not painful.

 

we run online. We tried cached mode but the additional transfer of info as the vast majority of our users move around a lot was unsustainable

 

We've got outlook 2010 with all the latest wsus supplied updates installed.

 

We will see what happens after sp3

Posted

If the problem remains after taking Exchange to the latest SP and CU, and Office to the latest SP and CU it might be worth looking into the issues around online mode.

 

I wonder whether the performance would be better overall if you ran everyone in cached mode, even if it does cause additional traffic? You would have to not roam the ost. For us, in general rebuilding a new ost at each logon for users who roam is a mild inconvenience, except for those who have 3000+ messages or 1Gbyte+ mailboxes, for whom it is a significant pain. Our recommendation for roaming users is to use OWA.

 

For 1500 active/simultaneous online users you probably need 24Gb+ Ram.

  • Thanks 1
Posted

How many real cpu's are on the host? What cpu does it run?

Cache mode is not needed for internal exchange normally tbh and causes issues like you found.

I would keep the 16GB of ram as well as exchange should speed up as it consumes more ram.

What is the base os of the vm running exchange?

Have you done a manual windows update to see what updates are available (after enabling microsoft update)?

  • Thanks 1
Posted
How many real cpu's are on the host? What cpu does it run?

Cache mode is not needed for internal exchange normally tbh and causes issues like you found.

I would keep the 16GB of ram as well as exchange should speed up as it consumes more ram.

What is the base os of the vm running exchange?

Have you done a manual windows update to see what updates are available (after enabling microsoft update)?

 

The physical host has 2x 12 core AMD Opteron

The VM running exchange is 2008R2.

 

We haven't run windows update manually yet.

 

The install of SP3 is still going and looks like taking a looooong time!

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...