Link existing link clone pool to fresh parent image

[Bluemoon404]

I’m using VM View 4.6 and my VM’s are all Windows 7 dedicated with persistent disks.

I need help with link clone pools.

I wanted to patch and re-compose one of my pools, unfortunately the master image had not been turned on for a long while and when I powered it on it was rejected by AD. This turns out to be due of the length of time the VM was powered off (expired password) coupled with a demotion of the DC that the master image had last contacted. All linked VM’s are performing as normal. In normal circumstances rejoining a physical PC to the domain would resolve the problem, so having no other option I decided to go ahead and disconnect the master image from AD and then rejoined it. I then patched the parent VM image and took a fresh snapshot and then tried to recompose the pool. Unfortunately the re-composer now fails, other groups can be re-composed.

Could someone please help with?

• Would rejoining the master image to AD have this adverse affect?
  
• Can I create a brand new master image and recompose the existing VM’s of this pool to link to the new image?
  
• What other options if any do I have at rescuing these linked VM’s about 100?
  

[mittim12]

Bluemoon404 wrote:

Could someone please help with?

• Would rejoining the master image to AD have this adverse affect?
  
• Can I create a brand new master image and recompose the existing VM’s of this pool to link to the new image?
  
• What other options if any do I have at rescuing these linked VM’s about 100?
  

  1: No, I don't think it would cause composer to fail.

  2:  Yes, that shouldn't be a problem as long as the OS is the same.   Persistent disk can only be attached on the same OS they were created with.

What kind of error are you getting with composer?

[Bluemoon404]

Internal composer error: contact administrator.

This gets a little more strange. To test linking an existing pool of vms to a new image I created a new pool and then performed a recompose but I choose the latest snapshot from my dodgy master image and it worked fine, why for then does the vms of the initial pool not also recompose with the dodgy master image?

Internal composer error.

Are there any logs where I can find out more of what is going on?

[mittim12]

Here is a KB on VMware View log locations.  There are logs for composer, agent, connection broker, and client.

http://kb.vmware.com/kb/1027744

[Bluemoon404]

Ok having looked at the log files it turns out to be an issue with a recent password change which I made. I don’t fully understand it though.

I recently changed the local admin account on the Virtual Centre which is where the composer also sits, I then changed the password for connecting to the VC from within the view manager – view connection – servers – vcenter servers. From this point I could create new pools and recompose these new pools but was unable to recompose the older pools which were created when the local admin account was different to the new password. Changing the local admin and password back to the original password has allowed me to recompose the older pools.

This has confused me, does anyone know why this would happen?

[mittim12 · Accepted solution]

I have actually seen something similar before.   The end result was that there were two vCenter accounts in the View Composer database and the pools that are not working are somehow linked to the older account.    I can give you a high level overview of how it was fixed or if you have VMware support I can give you the SR number so that you can open a ticket and references.  

[Bluemoon404]

I logged a call with VM support and as mittim12 correctly diagnosed there were multiple entries vc_config ID's in the SQL database.

[mittim12]

Great news.