Database Replication "Link has failed" after no communications


  • I have a secondary server / distribution point SCCMSECBBG1 that was working successfully until this week.  I use this site primarily for OSD, and five weeks ago I was able to successfully distribute an image and driver packages from our primary site SCCMLU1.  This week I am unsuccessful in attempting to distribute another image to it.  After the Content Status stayed yellow for a couple of days, I checked SENDER.LOG.  I see these entries every couple of minutes.

    There is no existing connection, Win32 error = 64

    Error during connection to \\\SMS_SITE (64).

    We looked in Monitoring / Database Replication and found that while the 'Child to Parent' link is active, the 'Parent to Child' link has failed.  Running the Replication Link Analyzer returned this:

    8/23/2013 2:16:54 PM: ReplicationLinkAnalysis Information: 1 : Initializing replication link analysis engine...
    8/23/2013 2:16:54 PM: ReplicationLinkAnalysis Information: 1 : Checking registry access to
    8/23/2013 2:16:54 PM: ReplicationLinkAnalysis Information: 1 : Checking registry access to
    8/23/2013 2:17:57 PM: ReplicationLinkAnalysis Error: 1 : GetRegistryKey: IOException - server not found.
    8/23/2013 2:17:57 PM: ReplicationLinkAnalysis Error: 1 : Link analysis engine could not access the registry on target site
    8/23/2013 2:17:58 PM: ReplicationLinkAnalysis Error: 1 : Failed to initialize the link analysis engine.
    8/23/2013 2:17:58 PM: ReplicationLinkAnalysis Information: 1 : Completed replication link analysis thread.
    8/23/2013 2:19:27 PM: ReplicationLinkAnalysis Information: 1 : Uninitialized the link analysis engine.

    I don't know where to go from here.  I don't know how significant this is, but three weeks ago the circuit connecting our site with the primary was damaged and off line for 54 hours.  I notice several logs that haven't change since either the time the failure started or the time the circuit was repaired.  I've tried rebooting the server but it made no difference.  I don't have direct access to the primary site and have to work through another person.

    Charlie Spencer

    Friday, August 23, 2013 3:04 PM


All replies

  • 64 = The specified network name is no longer available. Have you already checked basic network connectivity?

    Torsten Meringer |

    Friday, August 23, 2013 3:23 PM
  • Well, we can ping each server from the other one.  I can copy files between the two and view logs on the primary from the secondary.  I'm running the console on the secondary successfully.

    Charlie Spencer

    Friday, August 23, 2013 3:29 PM
  • I don't know what happened, but the OS image was distributed over the weekend.  SENDER.LOG no longer shows the connection errors.   Monitoring / Database Replication still shows 'Parent to Child' down.

    Charlie Spencer

    Monday, August 26, 2013 1:09 PM
  • Blatant, unrepentant bump.

    I guess this can be summed up as "What could fail when a secondary remains disconnected from its primary for too long?"  I suspect the problem is at the SQL level, but I don't have the skills to troubleshoot that.

    Charlie Spencer

    Thursday, September 05, 2013 7:36 PM
  • Did you ever get an answer to this or figure it out?  My lab servers are giving me fits with the SQL link.  Bashing my head on the desk trying to resolve this.  Not only that, the link analyzer only works on my primary site and not the CAS site.

    Keeps failing to launch with an IO error: not found.  When, clearly, I can ping the DNS name and reverse nslookup using the IP.  

    Carlton Lee

    Wednesday, October 09, 2013 2:25 PM
  • No.  Last week we concluded there's something going on with out WAN independent of SCCM and SQL.  For example, multiple devices at the same location as the primary server cannot connect to file or print shares at my location, although I can reach theirs and other locations can reach mine.  Unfortunately, if the primary server can't reach my shares, it can't recreate them either, so I can't even try rebuilding the local secondary server until we get that problem fixed.  It's in the hands of our LAN team now, and I get to learn more about SCCM's stand-alone media options :-(


    Charlie Spencer

    Wednesday, October 09, 2013 2:40 PM