[Edit: I did figure out a little more of the issue. I have figured out that what is causing this is the IIS Application Poll that The Exchange 2013 installer creates by default when installing called "MSExchangeRpcProxyAppPool". The server is recycling the RPC Application Pool every 20 - 30 minutes or so. I have recycling completely turned off  on this pool yet it still restarts which leads me to believe it may be a code problem since it only should recycle if the poll crashes. Short of a bug in the code somewhere, I for the life of me can not find why the Application Pool keeps restarting. This is the only pool out of all the ones Exchange created during installation that is recycling constantly. It is frustrating because our Outlook Anywhere clients keep getting disconnection messages several times an hour because of this. The event logs I listed below are the only event logs I can find that trigger when the pool recycles, so as you can see, they give me no useful information in debugging the problem.]

This problem only applies to outlook clients using RPC over HTTP. ActiveSync, OWA and EWS are working fine.

There is a problem with the RPC over HTTP tunnel on our server. Every 20-30 minutes, sometimes more sometimes less, but usually with in that 20-30 minute timeframe, the RPC over HTTP tunnel drops and all Outlook Anywhere clients get immediately disconnected. IT comes back up very fast and all clients reconnect with in a matter of seconds. Not a huge deal for people using cached mode outlook but all the non-cached mode people (Thin clients are an example are not using cached mode to save diskspace). People using non-cached mode would get a couple errors pop up and outlook stopped working.

The only relevant entries I have found in the exchange server event log are here. These same grouping of events show up everytime the RPC over HTTP tunnel shuts down and restarts. I just can't figure out why its doing it:

1-23-2013 9:21:18 AM RPC Proxy 4 (1)
-- RPC Proxy successfully loaded in Internet Information Services (IIS).

1/23/2013 9:22:41 AM WAS 5138 None
--- A worker process '75832' serving application pool 'MSExchangeRpcProxyAppPool' failed to stop a listener channel for protocol 'http' in the allotted time.  The data field contains the error number.

1/23/2013 9:22:41 AM WAS 5013 None
--- A process serving application pool 'MSExchangeRpcProxyAppPool' exceeded time limits during shut down. The process id was '75832'.

I have been searching for weeks on the problem and all I really have found is that the problem can sometimes be caused by RPC timeouts so I set the RPC timeout on the server to 120 seconds via the registry. That actually helped for about 5 days and then the problem mysteriously came back and I can not get it to go away even though the registry value is still set to 120 seconds on that RPC time out.

Does anyone else know of any thing else that causes this? Its Exchange 2013 on Server 2012 and is a pretty fresh install. Was setup in mid-December.


--Trent W.

• Edited by Talmen Wednesday, January 23, 2013 11:16 PM Added some information to the post.

try the below :)

a long term work around by adding a registry key that extends the timeout with the following key

HKEY_LOCAL_MACHINE/Software/Policies/Microsoft/Windows NT/MinimumConnectionTimeout

With a DWORD VALUE of 120

• Proposed as answer by James Dyke Thursday, January 31, 2013 9:20 AM

try the below :)

a long term work around by adding a registry key that extends the timeout with the following key

HKEY_LOCAL_MACHINE/Software/Policies/Microsoft/Windows NT/MinimumConnectionTimeout

With a DWORD VALUE of 120

See, I found that possible solution through some google searches. I tried that a few weeks ago and it seemed to work for a few days but then started doing it again.

After I posted this thread, the RPC Tunnel started behaving for a couple days and didn't drop connections, through no action from me. I checked again before I posted this reply and noticed the exchange server started dropping all RPC over HTTP connections again every 20-20 minutes this morning right after midnight and has been doing it ever since.

That DWORD value has been set on the server since the middle of December, and I did double check to make sure it was still set. I've rebooted the server several times since I set that value as well.

• Edited by Talmen Thursday, January 31, 2013 3:19 PM
• Marked as answer by Simon_WuMicrosoft contingent staff, Moderator Friday, February 01, 2013 8:14 AM
• Unmarked as answer by Talmen Tuesday, February 05, 2013 3:17 PM

Still no news regarding this issue... I have also tried to set the MinimumConnectionTimeout without any luck

Martin

• Edited by Martin_157 Tuesday, February 05, 2013 9:45 AM

I have been experiencing this issue since we started the migration/co-existence with exchange 2007 and 2013. I have also been noticing some rpc service stops and restarts, as well as watching connections drop in outlook. I have opened a case with Microsoft and they have reported back to stop the Microsoft Exchange Health Manager Service on all your 2013 cas and mailbox servers. I did this as well as put the service to manual start-up so hopefully it doesn't get turned back on. After this, restart the Microsoft Exchange RPC Client Access service on all 2013 exchange servers. Since i have done this i haven't seen the rpc crashing nor have i seen any connections drop in outlook. This has been this way for about 4 hours. I will update later today but i think this might be the ticket. Now Microsoft will need to figure out the bug with that Health Manager Service.  A side note, i have been noticing my databases failing over to different servers with out any real reason that i can tell, and i think this health manager might be the culprit as well. As in the 4 hours today none of my database have failed to another server yet.

• Proposed as answer by MC Wee Friday, July 12, 2013 12:40 AM

Hi Justin, were you able to find out if the MS Health Manager service is really responsible for your databases falling over to different servers? Thanks.

Maaaarj, I have been monitoring this since July 3 and have had no database fail overs like i was having before. I still have the MS Health manager service still stopped. I also have no longer seen the dropped connections in outlook. I just responded back to MS Support today giving them my results. I have asked what are the ill-effects with having this service stopped and when a fix maybe released. Ill try and keep this posted with what i find. So from what can i tell that was the ticket.

• Proposed as answer by Amir Al Friday, July 12, 2013 11:56 AM
• Unproposed as answer by Amir Al Friday, July 12, 2013 11:56 AM

Hi,

CU2 seem to be fix for this. Everthing work as it should.

Actually what has lead me to this forum is that after installing CU2 MS Health manager doesn't start by it self any more.

If I start it manually it starts without any complains.

And I didn't notice any disconnections mentioned above, but this might be because server hasn't let out to the real world nature yet.

• Edited by Aurimantas Kulisauskas Thursday, July 18, 2013 2:16 PM spelling

Don't stop "MS Health Manager service", You will end up with serious problems.

Hi All,

I have the same exact problem... every 5-10 min.  Is there any updates or hotfixes out for this yet??

Cheers

Chris

Having experienced a whole host of Exchange Server 2013 RPC issues, I would as you and anyone else seeing this to check if they use a Quest Agent called Change Auditor.

This has known RPC Thread issues.

Latest version is 6.0.217 released on Friday. 

• Proposed as answer by Rob Derbyshire Tuesday, July 22, 2014 12:28 PM

I am also experiencing the same issue on Exch. 2013 when i enable ChangeAuditor agent. 

Its the same issue even for v6.5 as well.

Any one was able to resolve issue?