none
Object Replication Manager failed to process Object changes

    Question

  • Hello,

    we receive error "Object Replication Manager failed to process Object changes. These changes will be retried on next processing cycle." constantly on our sccm primary site server. the error keeps popping up for a long time now.

    i looked through objreplmgr.log, but i cannot find any error messages. the error seems to happen when "processing changed GCA objects":

    +++Begin processing changed GCA objects	16.01.2013 18:44:39	4704 (0x1260)
    STATMSG: ID=6004 SEV=E LEV=M SOURCE="SMS Server" COMP="SMS_OBJECT_REPLICATION_MANAGER" SYS=sccm01.domain.com SITE=xxx PID=2324 TID=4704 GMTDATE=Mi Jän 16 17:44:39.794 2013 ISTR0="" ISTR1="" ISTR2="" ISTR3="" ISTR4="" ISTR5="" ISTR6="" ISTR7="" ISTR8="" ISTR9="" NUMATTRS=0	16.01.2013 18:44:39	4704 (0x1260)
    +++Completed processing changed GCA objects	16.01.2013 18:44:39	4704 (0x1260)

    any idea where i can look further? what are GCA objects?


    Wednesday, January 16, 2013 5:56 PM

Answers

  • no more errors today for me, the following solved the problem:

    according to MS, the problem happens because of an assignment to a collection that no longer exist.

    to get all assignments: SELECT * FROM vClientSettingsAssignments
    to get all collections: SELECT * FROM collections

    so i built the following query, to look for assignments without a valid collection - as you can see it found one entry for me:

    i removed the entry in row 12 (delete ... where UniqueID = xxxx), and the error went away.

    PLEASE DO NOT MODIFY YOUR SQL DATABASE WITHOUT A BACKUP, I WON'T GUARANTEE ANYTHING :)

    • Marked as answer by Robert Rostek Wednesday, February 13, 2013 8:56 AM
    Tuesday, February 12, 2013 7:20 AM

All replies

  • Did you install SP1 by chance?  I had this same problem and I just opened a case with MS.
    Wednesday, January 16, 2013 6:02 PM
  • You have already taken a look at the objreplmgr.log. Error may be related to any network, DNS etc issue and the change would usually apply in the next processing cycle.

    • Proposed as answer by Adil Rathore Wednesday, January 16, 2013 6:03 PM
    • Unproposed as answer by Robert Rostek Wednesday, January 16, 2013 6:32 PM
    Wednesday, January 16, 2013 6:03 PM
  • yes.

    SP1 broke quite a lot of things (our pxe boot image for example), but i got it all fixed over the last days. this seems to be the last remaining error for now. by the way, i also patched to SP1-CU0 today, to fix the certificate error in the client installer.

    if you get any feedback from MS, i would be thankful if you post here.


    edit: some of our packages no longer go from yellow to green status in the console / content distribution, after a package update. maybe this is related?
    Wednesday, January 16, 2013 6:04 PM
  • @adil: this error appears for nearly a week now, with several reboots of the corresponding servers in between.

    it did not fix itself.

    Wednesday, January 16, 2013 6:32 PM
  • I had the exact same issue seen in objreplmgr.log, beginning after installing SP1 and a subsequent reboot.  We only have a single site/server, and the dozen reboots I attempted made no difference.  And while I can't say for sure how it was resolved, I can say that the errors stopped appearing after 2 changes/discoveries:

    1) I uninstalled the Management Point, rebooted, reinstalled, and I believe rebooted again.  This was done on 1/11/13 around 4:00 PM and the error last occurred on 1/12/13 at 11:48 AM.  I was not seeing errors with any other logs, but this was done in desperation, as I was hoping to solve a possibly unrelated issue with clients not reporting their deployment status.

    2) I found a firewall policy that appears to be corrupt or locked.  In the console, I cannot edit it or open properties, and it still exists.  Perhaps coincidental, but I did discover this sometime in the AM on 1/12/13.

    Wednesday, January 16, 2013 9:45 PM
  • I too have a fire policy I can not edit.  Did you ever find a solution for this problem?
    Sunday, January 20, 2013 1:04 AM
  • i still have this error message every 30 minutes. no firewall policy in sccm.

    i will contact MS support now.

    Tuesday, January 22, 2013 7:25 AM
  • Anyone find a solution to this.

    I am getting this error every 30 mins since SP1 was installed. Everything appears to be working ok. We have a stand alone primary site.

    Message ID 6004

    "Object Replication Manager failed to process Object changes. These changes will be retried on next processing cycle."

    Thursday, January 24, 2013 2:44 PM
  • Not yet.  Been to busy fixing other issues I am having. :)

    Shawn

    Thursday, January 24, 2013 2:53 PM
  • i am confident i just solved this on my system. microsoft support sent me some helpful hints - i will monitor it until tomorrow and will post here should the issue be resolved for good.
    Monday, February 11, 2013 4:02 PM
  • Cool.  I would be very interested to know what it was and if it works.

    Shawn

    Monday, February 11, 2013 4:05 PM
  • Me too, post SP1 upgrade has left one of our primaries having this issue from our CAS.


    www.bighatgroup.com

    Monday, February 11, 2013 9:00 PM
  • Hi Niall,

    Turn on Verbose SQL Logging in the registry on your server.

    H_K_L_M\Software\Microsoft\SMS\Tracing\Sqlenabled=1

    Restart SMS_Executive and your objreplmgr.log should be more easy to analyse.

    Find the row where it reports the Status message ID and you will find your query that is failing.

    It's probably a collection that's gone missing that still has some kind of a assignment.

    And you might se some similar queries.

    select ID,UniqueID,ClientSettingsID,CollectionID,IsTombstoned,SourceSite,DateCreated,DateModified,rowversion,PolicyCRC from vClientSettingsAssignments where rowversion > 0x0000000000C604E1  order by rowversion

    Sucessfull

    select CollectionID from Collections where SiteID = 'XXXYYYY'

    Failed

    I think there are two options how to resolve this i cannot guarantee it will work and I'm sure it't not supported unless a MS engineer tells you to do so during a support call.

    you will find you Clientsettings assignments with the first query.

    1. Drop the row of your missing assignment.

    or

    2. Set the assignment to IsTombstoned = 1 where they are binded to your missing collection.

    update vClientSettingsAssignments set IsTombstoned = 1 where CollectionID = 'XXXYYYYY'

    Regards,

    Magnus

    Monday, February 11, 2013 11:23 PM
  • no more errors today for me, the following solved the problem:

    according to MS, the problem happens because of an assignment to a collection that no longer exist.

    to get all assignments: SELECT * FROM vClientSettingsAssignments
    to get all collections: SELECT * FROM collections

    so i built the following query, to look for assignments without a valid collection - as you can see it found one entry for me:

    i removed the entry in row 12 (delete ... where UniqueID = xxxx), and the error went away.

    PLEASE DO NOT MODIFY YOUR SQL DATABASE WITHOUT A BACKUP, I WON'T GUARANTEE ANYTHING :)

    • Marked as answer by Robert Rostek Wednesday, February 13, 2013 8:56 AM
    Tuesday, February 12, 2013 7:20 AM
  • Hi,

    That worked for me like a charm!

    Just delete the entry with Null in CollectionName.

    Thanks!

    Tuesday, February 12, 2013 10:41 AM
  • Excellent.

    I have run the query and located the null entry. Do I delete the entire row or just the Unique ID section.

    Cheers

    Tuesday, February 12, 2013 10:21 PM
  • the entire row, with a query like this:

    delete from vClientSettingsAssignments where UniqueID = 'your-guid-here'

    Wednesday, February 13, 2013 8:56 AM
  • the entire row, with a query like this:

    delete from vClientSettingsAssignments where UniqueID = 'your-guid-here'

    Is there a special format for the guid?

    When I run this I receive an error message:

    Msg 207, Level 16, State 1, Line 1
    Invalid column name '7FCEC793-70E3-4FEA-955D-8E0DA086DCE6'.

    Friday, June 20, 2014 11:00 AM
  • Don't modify anything in the database manually. That's not supported and might make things worse. Call Microsoft support (CSS) instead.

    Torsten Meringer | http://www.mssccmfaq.de

    Friday, June 20, 2014 11:49 AM
    Moderator
  • This solution worked for me as well.

    Same problems:

    SMS_OBJECT_REPLICATION_MANAGER

    "Object Replication Manager failed to process Object changes. These changes will be retried on next processing cycle.

    Solution: Review the previous status messages and logs for further clarification about this problem."

    This problem began after upgrading from SCCM 2012 RTM to SCCM 2012 SP1. The error message would reoccur every 30 minutes.

    I read this very Forum article that Robert started, ran the "join" query and found my environment had five different NULL values under "collection name." I wasn't comfortable deleting the rows so I worked with microsoft tech support. Without any mention of the solution that Robert provides, the tech came to the same conclusion. The tech and I deleted the rows and then restarted the "SMS_Executive" service. That immediately resolved the problem; no more error messages in "SMS_OBJECT_REPLICATION_MANAGER."

    To be specific, and to reiterate what Robert Rostek wrote earlier, the join statement the microsoft tech ran against the SQL server was...

    select ca.*,c.[CollectionName] from vClientSettingsAssignments as ca LEFT JOIN collections AS c ON ca.CollectionID=c.SiteID

    This exposes the NULL values under the "CollectionName" column.

    For each of the rows that contained a NULL value, the tech and I ran the following delete command...

    delete from ClientSettingsAssignments where ID=xxxxxxxx

    Wednesday, July 23, 2014 12:15 AM