UCCX HA CONFIGURATION COMPLETE WITH DBException ERROR

[charles-moore]

Hi All,

Need a little Help,

I have configured UCCX High Availability. However on completion of the Slave node, all components are completed / Done with the exception of the

CCX Config Datastore.

Any ideas ???

Details below:

Cisco Unified CCX Config Datastore Error Activating components.

com.cisco.cluster.ClusterException: Component Controller CRS Config Datastore on node 2 : Enable
com.cisco.database.util.DBException: syncConfigData
java.sql.SQLException: The operation could not be performed because the OLE DB provider 'SQLOLEDB' was unable to begin a distributed transaction., SQLException Details: [SQLException] (ErrorMsg=The operation could not be performed because the OLE DB provider 'SQLOLEDB' was unable to begin a distributed transaction. SQLState=S1000, SQLErr#=7391)java.sql.SQLException: The operation could not be performed because the OLE DB provider 'SQLOLEDB' was unable to begin a distributed transaction.

Cisco Unified CCX Historical Datastore Error Activating components.

com.cisco.cluster.ClusterException: Primary Controller for CRS Historical Datastore on node 2 : enable(): CRS Config Datastore is not enabled on this node 2

[Aaron Harrison]

Hi

Did the other 'datastore' components activate, or did you not enable them all?

Did you add the other servers' hostname/IP to the hosts/lmhosts file on each server?

Aaron

[charles-moore]

Hi Aaron,

Thanks for getting back.

Both contact centres are using the Host files and the Hostname are the following:

Server 1     ukgreipcc-crs71

Server 2     ukgreipcc-crs71

Server 1      Configured using the "First Node" Option all Services activated ok.

Server 2      Configured using the "Add to Cluster" Option. was also given the option to services to activate and completed.

                  apart from the two mention above.

The Servers are in a Lab environment running Win 2003 Ser Pack 2. (Not a Cisco OS 2003) would this make any difference.

The System works fine in our labs as a Standalone.

Cheers

Charles

[Aaron Harrison]

Hi

Yes, the OS could make a difference. It may be difficult to work out where the problem lies exactly though.

I would normally build a lab system in VMWare from the proper Cisco OS media. Builds find from one of the HP DVDs, and you can enter the license key from the sleeve to prevent it timing out after 30 days if you need to.

Regards

Aaron

[Pablo Leiva · Accepted solution]

Hello Charles,

This could happen if you create the servers based on a ghost image or it's a clone in VMware. Please compare the registry key “HKEY_CLASSES_ROOT\CID” on both the servers. If they are identical, then the Config Datastore Activation will fail.

In that case, you might need to re-install the Windows OS from the disc that will create it's own unique CID.

If you still want to use the ghost or VMware clone, you will need to re-create the CID before installing UCCX:

Step 1: Use Add Windows Components, and remove Network DTC.
Step 2: Go to the command line and run: MSDTC -uninstall
Step 3: Go to the registry and delete the following MSDTC keys in
  HKLM/Software/Microsoft/Software/MSDTC,
  HKLM/System/CurrentControlSet/Services/MSDTC, and
  HKEY_CLASSES_ROOT\CID (if they're still there).
Step 4: Reboot
Step 5: Go to the command line and run: MSDTC -install
Step 6: Use Add Windows Components, and add Network DTC.
Step 7: Go to the command line and run: net start msdtc

Step 8: Reboot again.

This procedure is described here:

http://support.microsoft.com/?kbid=873160

After this, the CID should be new and you can install UCCX and add to the cluster.

Hope that helps!

Pablo

[charles-moore]

Hi Pablo,

Not sure How you came across this but you are a Don / Guru.

It did the job, your Information was point on word for word.

What can I say but a Big Thanks.

BTW, what is the correct procedure when testing failover ?

Cheers Mate :-)

[Pablo Leiva]

No problem my friend.

I think you can test the failover, but just restarting the Node Manager on the primary server. From control center, you can see a small square next to each service with an M for master or S for standby. You should see the M on the second server after the failover. The CAD agents will be logged out and logged back in automatically.

Best regards,

Pablo