3

I have installed an Exchange 2013 Mail server on our network following the Microsoft guide on migrating from an Exchange 2010 to 2013 server. I am at the point where I have both servers running side by side with the 2013 server proxying the 2010, and am preparing to move mailboxes.

I moved one mailbox that was created before the AD Schema was updated by the 2013 install, and the user could not log into Outlook after the mailbox migration had been completed (and syncing active directory). I ran the Microsoft Connectivity Analyzer and discovered that this user was getting a 401 error when accessing the AutoDiscover URL.

At this point, I ran a backup of the server before proceeding to make some test changes that might break the system. After the backup completed, the user could log in without any error. With the error fixed and not wanting to disturb any other users, I created 3 more test accounts which I then tested migrating, and they could log in fine immediately.

As suggest in a post I migrated another user (failed to AutoDiscover) and restarted the Information Store which had no effect. I then restarted another service which failed to restart, forcing me to reboot the server. After the server finished rebooting, the user was able to connect fine to the Exchange Server.

I have also applied the following commands to try and force RPC over HTTPS:

 Set-OutlookProvider EXPR -OutlookProviderFlags:ServerExclusiveConnect
 Set-OutlookProvider EXCH -OutlookProviderFlags:ServerExclusiveConnect

as specified in https://techcommunity.microsoft.com/t5/exchange-team-blog/ambiguous-urls-and-their-effect-on-exchange-2010-to-exchange/ba-p/593809 with no apparent effect.

I have also tried manually specifying the connection details, but when checking the name of the user, I receive an error 'Outlook cannot log on...' followed by 'The name cannot be resolved' when querying against the 2013 server. If I change the server name to the 2010, it finds the user and the server name changes back to the 2013 with both underlined. However when starting Outlook, I receive the error 'The server is unavailable'.

Has anyone experienced this before? Any idea where this permission update needs to happen without a backup or rebooting the server?

Please let me know if there are any details about the servers that would help with solving this error.

Community
  • 1
Zone12
  • 121
  • 2
  • 3
  • 7
  • Having the same exact issue. Did either of you end up finding a fix? I've been doing an iisreset on the Ex2013 CAS after each mailbox is moved as a workaround. – yeahimagek Nov 06 '14 at 03:32

3 Answers3

1

UPDATE: There is now fix for this. The workaround is to set the autodiscover application pool to recycle every 30 minutes, and do your user migrations to Exchange 2013 to off working hours.

Im having almost the same issue. Im still trying to find out what is causing it, but you don't need to do a complete restart of the server, you can just restart the autodiscover pool in IIS on the frontend server (in my case the CAS and MBX role are on the same server).

The issue seems to be that the Ex2013 FrontEnd server (where the user was connected while he was migrated to Ex2013) still proxies the request to the Ex2010 servers event though the user is migrated. Im still no sure what is the cause, and how to fix it. If I find a solution to the root cause I'll let you know.

Zarko
  • 21
  • 2
  • I am still having this issue in CU11 even though many places reference that the issue will be resolved in CUnn. – pat o. Mar 22 '16 at 17:19
1

This issue occurs because the cache on the server didn't get update. Generally, recycle the Autodiscover application pool on Exchange 2013 server would be enough:

Restart-WebAppPool MSExchangeAutodiscoverAppPool

Here is the Microsoft KB: https://support.microsoft.com/en-hk/help/3097392/outlook-logon-fails-after-mailbox-moves-from-exchange-2010-to-exchange-2013-or-exchange-2016

Sue.J
  • 366
  • 3
  • 12
0

Note that Exchange 2013 exclusively uses HTTP for Outlook connections, while Exchange 2010 could do either direct (RPC/TCP) or HTTP connections.

If your outlook clients are not using Outlook Anywhere before migration, this can be part of the issue as well. Outlook, after the mailbox moves to 2013 needs to be updated to use either Outlook Anywhere or MapiHTTP, however Outlook can struggle sometimes with having this set correctly.

i would compare the difference of migrating and creating a new profile to reusing the same one.

Eriq
  • 11
  • 2