Hi guys, we are currently migrating the last of our users to exchange 2010 and seem to be causing our unity servers to be unhappy. MWI's are not syncing up and Malex ,CiscoUnity_AvWm warnings A MAPI function call to the Microsoft Exchange server took XXXX milliseconds, which exceeds the system threshold of 5000 milliseconds. We have tried rebooting, re syncin UTIM and the problem continues to get worse. All of the users are unified but they still miss the red light o the phones. If I open a users account in unity and resync the MWI it works fine. We are thinking there is some sort of config issue between exchange and unity.We have also re run the message store config wizzard on multiple occasions. I would say today the delay is about 3 hrs on the light syncing up. I think I have opened about 6 support cases with cisco and the problem may go away for a day or so then return. Does anyone have an ideas on what may be causing this distress?
Take a look at this post, specifically the area discussing ExchangeMapiCdo and the version of mapi32.dll in the c:\microsoft\system32 directory - https://supportforums.cisco.com/message/3238293#3238293. I just performed an integration this past weekend and made sure I renamed the existing mapi32.dll before installing ExchangeMapiCdo. I know your version is 7.0(2) but could be the same problem for you. I am assuming you have the ES38 applied to your server for Exchange 2010 support.
Hi Ginger, thanks for the reply. We have ES35 applied and also ran ExchangeMapiCdo.EXE and have MAPI Provider : CDO Version - 6.5.8190.0.
This morning lights were synced and seemed to be working fine, then the CiscoUnity_AvWm: Technical Details:
Method IMAPITable::QueryRows (returned 0x00000000) warnings in event viewer on our primary unity box started along with CiscoUnity_ExchangeMonitor: Completed resyncing mailboxes on the following Information Store: for all of our information stores. Now MWI's are delayed again. I am watching our MWI only ports and it is working but very very slow.
Nixon - you have pretty much summed up our Unity issues perfectly, same event logs and symptoms. I can manually sync either in the SA or via the manual resync under the UTIM and the light goes to the correct state. After a MALEX or MAPI disconnect is logged the MWI sync is intermittent. We have 3 physical Exchange 2010 servers, 10 MBDB's, and a Cisco ACE load balancing the connections. We are on Unity 7.0.2 ES40 and have renamed the mapi.dll and reinstalled the MAPICDO which we thought fixed the problem. Most of our MAPI calls that take xxxx milliseconds happen during the day indicating this is load related. At the same time we also see an Exchange message stating that MAPI Session "639959-33aa-3434-395399a9" has exceeded 250 connections of type "objtMessage". We set the MAPI thresholds to Null for the UnityMsgStoreSvc during the 2010 partner server integration. I'm not sure if the two logs are related but they happen within about 1 minute of one another.
Curious if you'd made any progrss on this one?
We did make good progress on this. Basically we had to get cisco on the phone and re run permisions wizzard and correct issues within the message store configurator on the unity server. Along with that, using the GUSI tool we made some registry changes that the tool suggested and version of mapi is MAPI Provider : CDO Version - 6.5.8190.0, I think the big one was making sure firewalls were off on unity and exchange, ours was off on exchange but it seems an ES re enabled it on unity at some point. There was one entry in the registry for the cas that needed to be changed to 0, which after a reboot seemed to kick this in the pants. In our environment we were working with top level exchange 2010 people from Microsoft and come to find out the recommended amount of RAM for those exchange boxes was low, we bumped up to 36 GB per server and along with everything we did to unity it works fine now.
Here is the MS # 110120573647780
and here is the cisco one with the Resolution Summary: Disabled the firewall on the Unity servers. Created a new throttling policy on exchange for Unitymsgstoresvc account. Restarted the RPC client access service on exchange.
Hope this helps!
I'm having the exact same problem with my deployment. Added the Exchange 2010 servers to Unity 7.0(2) about a month ago and everything was working great - until more and more users were migrated over. Then the MWI lights for Exchange 2010 users started to break. Been through the MAPI reinstall, resetting permissions, disabling firewall, nothing has worked. Can you share what registry settings you applied to the Exchange CAS servers and Unity? All my Exchange 2007 users have no issues - it's only the 2010 users. Thanks!!!
We had same issue, however after applying all fixes, including the mapi32.dll, we still had issue with MWI for subscribers on Exchange 2010. Then today, I found out the issue was with Hardware Loadbalancer. I did have persistence configured for by source-ip, however the timeout was 20 minutes. It is too short for Unity. Unity does not like to change CAS server. I changed it to 1440 minutes (the max allowed) today, and I also have scheduled MWI daily sync configured. I figure we should be good now. I will monitor for a few days to make sure.
I considered that fact in our deployment because we to are using load balancers. Specifically, Barracuda 340 Load Balancers. However, Barracuda does not recommend turning on persistence in the configuration. I will have to contact them and see if it can be done. Oddly enough our MWIs will work fine for a week and then break. We do resyncs every morning which doesn't seem to have any impact.
I'm having a similar issue, but our Unity has the latest ES41. What registry setting in the CAS server that need to be set to "0". Please share the info.