After we made a change to the syncronization connection (added an exclusion for userAccountControl BitOn Equals 2), the service is trying to do a DS_FULLSYNC.  This fails on step 2 of 6 (Six domains in the forest, one step to import each) with the error stopped-database-connection-lost. 

We have the same issue in our test environment when trying to do a DS_FULLSYNC on two specific domains.  It appears to be related to the size of the domain since the two that fail are the largest by far (80000 and 40000 objects, I think).

No errors reported on the SQL server, and the only information in the event logs is on the SP admin server are:

Event ID 6322 FIMSynchronizationService

The server encountered an error because the connection to SQL Server failed.

Event ID 6075

The management agent "MOSSAD-***** Profile Synch" failed on run profile "DS_FULLSYNC" because the connection to the server database was lost.
 
 Additional Information
 Discovery Errors       : "0"
 Synchronization Errors : "0"
 Metaverse Retry Errors : "0"
 Export Errors          : "0"
 Warnings               : "0"
 
 User Action
 Verify that SQL Server is running.

Event ID 2004

The FIM Synchronization Service failed to update the timestamp. Verify that SQL Server is running.
 Error Code: 0x80230621
 Error Message: (A connection to SQL Server could not be established)
 

First thought was a timeout on the connection, however on test I have set the max time on queries for the SQL server to unlimited, and the connection timeout for DS_FULLSYNC to 20 minutes, neither change made a difference.

In Production, Step 1 importing the first domain runs for around 25 minutes and completes successfully.

Step 2 importing the next domain runs for about 30 minutes and then fails: stopped-database-connection-lost.  If I try to 'resume run profile' it will run for about 5 minutes on step two and I get the same error.  (note that looking at the "Number of user profiles" in SharePoint, I can see it increment by a few hundred/thousdand each time I do a resume)

Environment:  SP 2010 ServicePack 1 CU for July 2011 KB2536599, OS Server 2008 R2, Dell PowerEdge R710 12GB RAM 2x QuadCore Intel E5530 (SP Upgraded from existing MOSS 2007)

SQL 2005 ServicePack 4,OS Server 2008 R2, Dell PowerEdge R710 32GB RAM 2x QuadCore Intel E5530

Oh, and we removed the exclusion filter in test and still have the same isuses.  If Anyone has any ideas or knows of another Timeout/Connection settings, it would be gretaly appreciated.

As this is regarding the UPS inside Sharepoint 2010 (which uses a "black box" version of FIM built in that isn't meant to be messed with except through the SharePoint management interface(s)), you might be better served checking out the forum for that product.

That said, the first thing that jumps out at me is that you have this on SQL Server 2005.  FIM 2010 requires SQL Server 2008 SP1 or later.  I haven't seen anything that indicates the SP2010-integrated FIM can work on SQL 2005, but I'm always willing to be proven wrong.  Maybe it would work, since you seemed to get some results.  But maybe that's also the source of your trouble.

Other than that, all I can recommend is removing the UPS service/application and re-adding it to the farm, making sure that your service account(s) have the correct permissions (log on locally, local administrator at least for the setup, etc.)  We fought with that thing for awhile.

We did a clean install of SP2010 rather than upgrading from MOSS 2007, so your experiences may vary.

Chris

Hi Chris

Thanks for the suggestions.  I removed the service app and readded it (in test env), I even deleted the synchronization connection and recreated it, and it still fails on the larger domains.  Smaller domains in our forects sync fine, 583 obejects, 3229 objects, 6376 objects, but the larger two fail, 41516 objects and 92145 objects.

DS_FULLIMPORT (staging) succeeds, DS_FULLSYNC fails when it hits the larger domain (runs for about five minutes then stops, smaller domains that succeed actually ran for thirty or more minutes)

No errors are showing in SharePoint Central admin, it shows the sync importing information and its log looks ok.  The reason I opened the FIM client was we noticed that information for some users wasn't updating, thats where we saw the errors.  As for SQL version, SharePoint Server 2010 software reqiuirements states it supports SQL 2005 with SP 3 CU 3, so hopefully that includes all components that come along with it...... (fingers crossed)

I'll post this on the SP forum as well, but please, if you have any other ideas or suggestions, let me know.

James

 

Hi James, I am running into the same issue and errors.  The only difference is we am running SQL 2012 beta, so I'm wondering if thats an issue.  The OU its failing on is the largest one we have, about 22,000 objects.

 

We also have a sharepoint 2010 enviornment with SQL 2008 backend, same settings, that syns no problem.

James, do you have a link to your post in the SP forums so I can follow that one?  Thank you

http://social.msdn.microsoft.com/Forums/en-US/sharepoint2010setup/thread/fec1bf86-63aa-4d99-9d2f-77f0c5f0a49e

Still haven't found the issue, however in TEST I have been fiddling with the forbidden MIIS client.  Actually, just manually kicking off jobs, with some success. I disabled SP profile synch job so it wouldn't interfere, then ran the following jobs.  I looked at what SP was running and used the same order.

1. DS_FULLIMPORT (completes successfully)
2. MOSS_DELTAIMPORT (completes successfully)
3. MOSS_FULLIMPORT (completes successfully)
4. DS_FULLSYNC (fails, however using MIIS to resume job it will finally succeed, after 26 resumes)
5. MOSS_FULLEXPORT (fails, I've 'resumed' this job 63 times, and each time it adds a few hundred to a few thousand profles, but its not done yet)

If I get the last step to complete, I'll let SP do an incremental sync to see if it works.  Question I have, am I missing a crucial step (there is a MOSS_FULLSYNC, does this need to be run?), and if it does end up working, what are the negative impacts this may have (before I decide to do this in prod)?

 

I see that the MOSS Export will resume on its own up to three times if it fails during a run, is there a setting for the DS_FULLSYNC so that it also will retry if it fails?

OK, gonna volley this one back to the FIM side.  On the thread for this on the SP side, someone suggested creating multiple connections for subsets of OU's for the domains that I was having trouble syncing.  This actually works (had to create three connections for one domain), but a follow-up post stated that "It is not supported to have multiple sync connections to the same forest (or domain)." and that it could corrupt the sync database.

So, I know this would just be a workaround and not a fix for the issue I'm having, but it's been a couple weeks and thinks are really getting out of sync.

Anyone have any input?  I would also like to find out if there is a way to:

A) Have DS_FULLSYNC resume automatically if it fails (at least a configurable number of times)

B) Is there any logging available for the DS_FULLSYNC that might give me more information (tried verbose logging in SharePoint, but that didn't display any errors)

Don't know if this applies; but a customer of mine resolved some connection issues with this http://www.techsupportforum.com/forums/f137/wireshark-question-tcp-checksum-offload-248812.html#post3619034

That worked!  The actual setting on our NICS was named differently, but the process worked.

Note: We disabled "TCP Connection Offload (IPvX)" on the sync server AND the SQL server.

Update: I tested each of the NIC (Broadcom BCM5708C NetExtreme II GigE) Offload setting combinations on all servers and have found that all I need to do is disable "TCP Connection Offload (IPvX)" on the database server only.

I have found the solution and have resolved for me!!!
I have got two events - one with the same text (EventID: 6075) and the second with similar event

The management agent "MOSSAD-xxxxx" failed on run profile "DS_FULLIMPORT" because of connectivity issues.
 
 Additional Information
 Discovery Errors       : "0"
 Synchronization Errors : "0"
 Metaverse Retry Errors : "0"
 Export Errors          : "0"
 Warnings               : "0"
 
 User Action
 View the management agent run history for details.

EventID: 6050

I have deleted the synchronization connection, than created a new connection under the account which I had my SharePoint installed. After this new connection was created, I have edited this connection and changed the account which has granted Active Directory Domain Services permissions for profile synchronization.

Voila - it works :-)

HI,

Disabled "MySite Cleanup job"

Please check one small OU Add and Run Test All The Process get success. 

then Edit Connection Add next OU and then run incremental , one by one add all the OU

 Enable "MySite Cleanup job"

----------------------

http://blogs.msdn.com/b/russmax/archive/2010/03/20/sharepoint-2010-provisioning-user-profile-synchronization.aspx

http://technet.microsoft.com/en-us/library/gg750257(v=office.14).aspx
http://sharepointgeorge.com/2010/configuring-the-user-profile-service-in-sharepoint-2010/
http://blogs.msdn.com/b/spses/archive/2010/12/02/guide-to-user-profile-service-application-upa-part-2-setting-up-the-user-profile-service-application.aspx
http://blogs.msdn.com/b/spses/archive/2010/11/28/guide-to-user-profile-service-application.aspx

------------------------------------------------------