Archive Task on new server in a lab causes the Exchange IS to crash

[Beystor]

Hi everyone

we have just installed a new EV 2007 SP3 onto a Windows Server 2003 SP2 server into our lab environment. The Lab consists of a Domain Controller (Server 2008), two Exch2007 Mailbox servers (used for SCC) and two CAS/HUB servers in a NLB configuration, exactly as in production. All Exchange servers run on Windows Server 2008 boxes.

After installing the EV server and finishing the configuration, I try to start the Archiving Task. About 3 seconds later, the Information Store on the Exchange MBX server crashes and the Archiving Task fails.

I am running SP1 for Exchange and Update Rollup 8 and there is a direct connection between Exchange and EV.

I cannot seem to pinpoint the issue and would appreciate any help that someone can offer. I also included a trace file hereto (removed some priveledged info). Some of the common errors in the log are:

• CMailboxHelper::GetMailboxState - Failed to Open EV Hidden Message [0x8004010F]
• CAgentExchSynch::SynchronizeMailboxEntryEx() - Exit hrx=[0xC0040D1E]
• CAgentExchSynch::GetExchangeMailboxEntry - Caught COM exception [0xc0040d1e] - Unknown error 0xC0040D1E

On the Exchange server, three events are created at 12:28:33 for the store crash. The event details are:

Log Name: Application
Source: MSExchangeIS
Date: 2009/07/27 12:28:33 PM
Event ID: 9659
Task Category: (6)
Level: Error
Keywords: Classic
User: N/A
Computer: <MyExchServer>.<MyDomain>.local
Description:
The Microsoft Exchange Information Store encountered an unexpected exception 0xC0000005 at address 6ABE5C81 while processing a request for user EVServiceAcc.



Log Name: Application
Source: MSExchange Common
Date: 2009/07/27 12:28:33 PM
Event ID: 4999
Task Category: General
Level: Error
Keywords: Classic
User: N/A
Computer: <MyExchServer>.<MyDomain>.local
Description:
Watson report about to be sent to dw20.exe for process id: 2916, with parameters: E12N, c-rtl-x86, 08.01.0375.001, store.exe, mdbrest.dll, 45c81, c0000005, 8191, 08.01.0375.001. ErrorReportingEnabled: False



Log Name: Application
Source: Application Error
Date: 2009/07/27 12:28:33 PM
Event ID: 1000
Task Category: (100)
Level: Error
Keywords: Classic
User: N/A
Computer: <MyExchServer>.<MyDomain>.local
Description:
Faulting application store.exe, version 8.1.375.1, time stamp 0x49f67601, faulting module mdbrest.dll, version 8.1.375.1, time stamp 0x49f67622, exception code 0xc0000005, fault offset 0x00045c81, process id 0xb64, application start time 0x01ca0ea40e092ccd.

Any help will really be appreciated!

Thanks
Beystor

[Maverik]

I think to be honest this is an Exchange issue rather than an EV one. All EV is trying to do at this point is use MAPI (the same as Outlook) to view the mailbox. Exchange then seems to fall over and so then will the task. 

It looks similar to these MS KB article but this was fixed in rollup  8 which you say has been applied.

http://support.microsoft.com/kb/960495/

I would recommend in first instance discussing with MS. 

[Beystor]

The article is very close to what I am experiencing here, except that the faulting module is mdbrest.dll. I will download Update Rollup 9 and see if this helps at all.

Funny thing is, all other remote operations such as opening mailboxes and sending messages via Outlook or OWA works without any problems. I am wondering whether this is not something to do with a particular message. The crash happens exactly 3 seconds after I start it. I will do a few traces and see ifd the crash happens while the task is busy with the same mailbox.

Thanks for your ideas so far

[Beystor]

Strangest thing happened just now. In my attempt to create at least two trace files, something unexpected happened.

I started the first trace and then started the Archiving Task. It failed after about 3 seconds or so. This was expected. I then started the IS service again and mounted the databases. As planned, the next trace was configured but, when I then started the Archiving Task, it did not fail as expected. Instead, it kept on running and allowed me to synchronise manually without any issues. I then ran the Provisioning task and reran the synch. At this point it failed again.

This got me thinking that I did something different to the previous times. All consequent tests and attempts failed in the same way and I could not get the task to start again.

Eventually, I started changing the Logon account details in the task and the max connections etc. But the same thing happened. Lastly, I deleted the task and recreated it to do exactly what the Default one does.

The task started and has been running for about 30 minutes. I was able to provision, synchronise, enable mailboxes and everyting I expected to do up to this level of my deployment.

So, I hope this is a permanent situation.