I would like to know if its possible to find out which user or users might be doing a lot of MAPI transactions against the mailbox server that we have in our ORG. Exchange 2003 with SP2. General performance issues since Saturday and no known configuration changes to the environment. All of a sudden though my performon stats have started spiking and my clients have high latency in Outlook that they didn't see before. Are there any tools that let me dig into transactions to see where the offenders might be? I do run several third party products (Cisco Unity, Captaris Rightfax, and Blackberry BESX, all of which dig into Exchange's insides, so it could be one or more of those, but again, nothing changed that I know of before and my disk queue lenths have spiked as well as read activity on my store's. Any ideas?Anything worth doing is worth doing right.

Exmon

Hi TowedJumper, Any updates? Please remember to mark the replies as answers if they help and unmark them if they provide no help. If you have feedback for TechNet Subscriber Support, contact tnmff@microsoft.com.Frank Wang TechNet Community Support

Thanks. I actually found that and was looking at the page for the download when the email letting me know about your post hit. As with most problems, this one might be like an onion (multiple layers to the issue). We have some kind of problem between our Brocade Silkworm 300 and our netapp filer. One of the FC channels was flapping so it was disabled. Another product looks like it was fixed and its LUN's live on the same filer and used that same channel to get to the LUN's, but the problem was not resolved on the exchange side. My suspicions still point to some kind of FC issue as usage hasn't really increased in our small env. and no other changes were made prior to the issue on Saturday. I did download ExMon and ran it to find mostly low activity and sporadic at that. Anything worth doing is worth doing right.

Fiber channel to filer (filer to FC switch) was disabled but I think MPIO is broken on the exchange server or not installed correctly. We disabled the connection from that FC switch to the exchange server so its forced to use the other path and now its working much better. Thinking about it though, that really doesn't sound right. If MPIO was broken entirely then I would think even disabling of that connection wouldn't have helped (ie it shouldn't be working from the start if MPIO was broken. Dunno). Regardless, queue lengths have dropped and general performance returned to normal (for a 2003 machine with mailboxes that are way too large). Anyways, marking as resolved with credit to A_D_ for answering the question I asked, even if it wasn't the solution :) Anything worth doing is worth doing right.

I wanted to throw out one more suggestion in case anybody has perf problems and stumbles across this thread later. In addition to ExMon, I would recommend the ExTra tool (Exchange Troubleshooting Assistant) by Microsoft. In addition to including the ExMon scans as well (to get it to do ExMON I would assume you have to install that tool as well first, then select the ExMON scan options in the Advanced Options section of scan setup), it goes through MUCH more and seems to be a combination of ExBPA and the ExMon functionality. Went through my RPC statistics (higher than expected, generated an error), checked NICs for bad packet counts (expected 0, found 10 in the 5 minutes I had its scan, so something more to look into now), checked LDAP latencies and much more. It really is a very handy tool. Thanks Microsoft! ExTRA tool from Microsoft link Anything worth doing is worth doing right.