Problems with UCCX 7.0(1)SR05_Build504

Unanswered Question
Jun 15th, 2010
User Badges:

Hi there,


After a power down of servers at the weekend the UCCX is giving us errors doing anything within subsystem>rmcm.  I.e if we try and assign skills or take skills off users we get the following error attached.



Any ideas? We have tried restarting the UCCX server and also restarted node manager.

Attachment: 
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Aaron Harrison Tue, 06/15/2010 - 06:05
User Badges:
  • Super Bronze, 10000 points or more
  • Community Spotlight Award,

    Member's Choice, May 2015

Hi


If you take a look in System/Control Centre, does the server appear fully up? Do you have HA?


Aaron

barrymcnally Tue, 06/15/2010 - 06:44
User Badges:

Hi


Looking at the Control Center the only thing that isn't started is the Cisco Unified CCX Editor. If you start this up from the server then only Manager Manager and Bootstrap Manager are partial service.


I'm not sure what you mean by HA ( sorry bit of a newbie).  I have attached our server event logs in case that gives you any clues.

Attachment: 
Aaron Harrison Tue, 06/15/2010 - 07:24
User Badges:
  • Super Bronze, 10000 points or more
  • Community Spotlight Award,

    Member's Choice, May 2015

Hi


HA = High Availability, two servers in failover config.


It's normal for editor to appear stopped.


Everything else should appear as in service, including manage manager and bootstrap manager.


In the event logs I see an unexpected shutdown, nothing else too telling.


The UCCX logs are in c:\program files\wfavvid\log\, perhaps post back the latest files from the MCVD folder in there?


Regards


Aaron

barrymcnally Wed, 06/16/2010 - 01:46
User Badges:

Hi,


Things have got a lot worse since we last spoke, we carried out the following action that we found online


Solution
Complete these steps in order to resolve this issue:
1. Go to the C:\program files\wfavvid\ClusterData\Default\ folder on the CRS server.
Rename the com.cisco.crs.cluster.config.LockConfig folder to
com.cisco.crs.cluster.config.LockConfig.bak.
2.
3. Restart Node Manager.
If you do not wish to restart Node Manager, here is another way of clearing MutexLocks:
1. Click Start and type CET.
2. Choose No on popup message.
3. Find and click com.cisco.crs.cluster.config.LockConfig in the list located on the left.
4. Double−click the one record located on the right.
5. Select the com.cisco.crs.cluster.config.LockConfig tab located at the top.
6. Clear any fields that are not empty.



After carrying this out the UCCX started rebooting. We took off reboot on error to stop the reboot but system has stopped taking calls etc.



Attached are our last 3 logs from the MCVD folder.

Attachment: 
Aaron Harrison Wed, 06/16/2010 - 02:31
User Badges:
  • Super Bronze, 10000 points or more
  • Community Spotlight Award,

    Member's Choice, May 2015

Hi


Yep, looks like you broke it :-)


Exception=com.cisco.manager.ManagerException: Cannot load manager class; nested exception is:
com.cisco.manager.ManagerException: Exception while finding local nodeid; nested exception is:
com.cisco.config.ConfigNotFoundException: Query returned zero results


Of those steps you listed, which did you do?


Did you edit via CET, or did you rename the lock config folder? Or did you rename the whol clusterdata/default folder?


Can you screen grab and post the clusterdata/default folder, and the files from the com.blah.blah.NodeConfig folder underneath it?


Aaron

barrymcnally Wed, 06/16/2010 - 03:12
User Badges:

Hi,


Yip well truly broke. :-(


It was edited through CET and this didn't work. Then the lock config folder was renamed. Attached is a screendump of the clusterdata /default folder. There is nothing under the nodeconfig folder which i guess may be a problem :-)

Attachment: 
Aaron Harrison Wed, 06/16/2010 - 03:43
User Badges:
  • Super Bronze, 10000 points or more
  • Community Spotlight Award,

    Member's Choice, May 2015

Yep, no nodeconfig is a problem indeed. Are other folders also empty (e.g. appadminsetupconfig, ccmauthenticationconfig etc)?


Do you have a backup of the system?


Aaron

barrymcnally Wed, 06/16/2010 - 04:54
User Badges:

Two other folders are empty Archiveidconfig and archiveleaseconfig. All other folders contain files.

Unfortunately it looks like there was no back up set up by engineer on install.

Aaron Harrison Wed, 06/16/2010 - 05:34
User Badges:
  • Super Bronze, 10000 points or more
  • Community Spotlight Award,

    Member's Choice, May 2015

OK, just the NodeConfig is the problem I think.


Just had a go at my lab server, and you can rebuild the object.


1) Go into CET, and then click the NodeConfig item on the left

2) On the right hand side, right click and click 'Create'

3) On the dialog that appears, copy the data in the 'time' field. Click on the second tab and paste it into the install and upgrade times.

4) Set node id to 1, deployment type to 1, and make up a GUID e.g. "{01010101-1234-9876-5656-123456789012}" (including quotes)

5) Set the install path to "c:\program files\wfavvid\" (including quotes)

6) On the next tab, set model to a string representing your server type e.g. "MCS7845", set OS Arc to "X86-based PC", set OS name = "Microsoft Windows Server 2003", OS ver "502" Vendor "HP" or "IBM", compy name "YOURHOSTNAME" < using your server hostname, Harddisk size - 70000. Finally click 'add' and add the server IP in quotes e.g. "192.168.0.71".


This (if done correctly, things may be case sensitive as well) allows my lab server to start OK after I deleted my nodeconfig.


Some disclaimers:

- Would TAC like this solution? Probably not

- Will the server work properly? Perhaps

- Will you get problems when upgrading the server or patching the OS? Possibly

- Would I like to do this on a production server? Not unless I was very desperate


Obviously the ideal solution would be to have a backup of the clusterdata files to restore, but hey...


Good luck.


Aaron

Actions

This Discussion