Hi,
I am trialling VMWare View 3 with vCenter 2.5 Update 4 and ESX Servev 3.5 Update 3. Everything was working fine with a linked clonse pool until we tried to reset all the VMs in the pool. All but one reset, but the final one just sat with a task running for hours and hours. Finally we had to restart all the services, delete the desktop pool and start again.
When we created a new linked clone desktop pool it all appeared ot be working fine, it creates the replica and source machine and created the first VM. In the vServer logs you can see the tasks all being completed with a final Reconfigure Virtual Machine entry. But that that point it just sits there. Doesn't power on the VM to do the quick prep, nothing.
I have tried re-creating the pool three time, and I get the same problem every time.
Any ideas as this is starting to drive me mad.
Thanks
Graham
Take a look in the event log. It sounds like it's having an issue adding the VM to the domain or something else. Also select the pool in the Pools & Desktop View to see the summary page for that pool. The error should show up at the top of the summary page.
Thanks for your reply.
Nothing obvious in the event log, and the machine entry in the desktop sources tab is just sitting with Task None and State Provisioning. No problem reported on the summary page.
It is as if it is waiting for something to complete that never does.
Graham
Hrmmm... Sounds like the Composer DB might be corrupted. What is the Composer DB? SQL? Oracle? SQL Express? Is the Composer service running?
It is running on SQL Server 2005 Workgroup Edition. Database looks ok, and I have been through and reset it all back to the default just to make sure.
I am getting some schannel errors from the VC machine (which is running VC and Composer) talking to our DCs, so it may be as you suggest a problem creating the user accounts, but there is no error from Composer itself.
Graham
Composer should be generating an error. Did you check the actual composer logs on VC besides the event log on View?
Also, last time I checked, SQL Workgroup wasn't support DB (unless that changed in 3.0.1).
Hi,
Ok, I have reset everything (cleared out the database, pools etc.) and started again with a clean log.
It all appears to progress ok until I get to the following section.
2009-03-30 12:40:57,586 | 5 | INFO | Sim.ServiceCore.SimServiceApiImpl - CreateSviClones was called for deployment group 02aebc3a-e941-47bd-8fa7-70c525d3ba6a, for 1 clones.
2009-03-30 12:40:57,586 | 5 | INFO | Sim.ServiceCore.SimServiceApiImpl - Received a request to create 1 clones for deployment group 02aebc3a-e941-47bd-8fa7-70c525d3ba6a. Filed as request id f2de3c04-de23-46ed-8f84-cb72553edf12
2009-03-30 12:40:57,586 | 5 | INFO | ServiceCore.LockManager.LockManager - A lock is requested for DeploymentGroup with id: 02aebc3a-e941-47bd-8fa7-70c525d3ba6a
2009-03-30 12:40:57,914 | 5 | INFO | ServiceCore.Action.UfaInitializeDiskAction - Created an UfaInitializeDiskAction with total disks 1.
2009-03-30 12:40:57,992 | 5 | INFO | ServiceCore.LockManager.LockManager - Multiple lock are required for SimClone with ids:
2009-03-30 12:40:58,476 | WFE thread 7 | INFO | CommonLib.VcSubsystem.PropertyCollectorUpdateTracker - Property collector now monitoring 1 tasks.
2009-03-30 12:41:00,773 | VC thread | INFO | ServiceCore.Action.CreateLinkedCloneVcAction - Linked Clone created with dsPath: temp-lc-c4f1a143-1dd3-43cd-8028-/temp-lc-c4f1a143-1dd3-43cd-8028-.vmx
2009-03-30 12:41:01,726 | VC thread | INFO | ServiceCore.Action.RegisterVmVcAction - Linked Clone created with VM MoId: vm-2654
2009-03-30 12:41:59,865 | 5 | INFO | Sim.ServiceCore.SimServiceApiVal - Call GetDeploymentGroup with parameter:
dgId : 02aebc3a-e941-47bd-8fa7-70c525d3ba6a
2009-03-30 12:42:00,068 | 5 | INFO | Sim.ServiceCore.SimServiceApiVal - Call EnumerateAdConfigEntries.
2009-03-30 12:42:01,865 | 7 | INFO | CommonLib.VcSubsystem.VcSubsystem25 - Vc operation failed.
System.Web.Services.Protocols.SoapException: The task was canceled by a user.
at System.Web.Services.Protocols.SoapHttpClientProtocol.ReadResponse(SoapClientMessage message, WebResponse response, Stream responseStream, Boolean asyncCall)
at System.Web.Services.Protocols.SoapHttpClientProtocol.Invoke(String methodName, Object[] parameters)
at VimApi.InternalVimService.WaitForUpdates(ManagedObjectReference _this, String version)
at VMware.Sim.CommonLib.VcSubsystem.PropertyCollectorUpdateTracker.b__4()
at VMware.Sim.CommonLib.VcSubsystem.VcSubsystem25.VcMethod`1.Invoke()
at VMware.Sim.CommonLib.VcSubsystem.VcSubsystem25.VcExceptionHandler`1.Invoke()
Hi,
Ok, got a bit further. I think it is hanging accessing the UFA service, whatever that is.
Graham
2009-03-30 12:49:20,385 | VC thread | INFO | ServiceCore.Action.CreateCloneAction - Clone created with VM MoId: vm-2656
2009-03-30 12:49:22,541 | VC thread | INFO | ServiceCore.Action.CreateSnapshotAction - Snapshot created with VM MoId: snapshot-2658
2009-03-30 12:49:25,041 | VC thread | INFO | ServiceCore.Action.CreateLinkedCloneVcAction - Linked Clone created with dsPath: source-lc-6ae2332a-fc7d-47e1-a99/source-lc-6ae2332a-fc7d-47e1-a99.vmx
2009-03-30 12:49:26,213 | VC thread | INFO | ServiceCore.Action.RegisterVmVcAction - Linked Clone created with VM MoId: vm-2659
2009-03-30 12:49:28,604 | VC thread | INFO | ServiceCore.Action.CreateLinkedCloneVcAction - Linked Clone created with dsPath: VDISTDP0001/VDISTDP0001.vmx
2009-03-30 12:49:29,729 | VC thread | INFO | ServiceCore.Action.RegisterVmVcAction - Linked Clone created with VM MoId: vm-2661
2009-03-30 12:49:30,416 | WFE thread 3 | INFO | ServiceCore.ActiveDirectory.AdUtil - Attempting to add a computer account using DC 1 of 4
2009-03-30 12:49:31,619 | WFE thread 3 | INFO | ServiceCore.ActiveDirectory.AdUtil - New computer account has been created successfully.
2009-03-30 12:49:34,229 | WFE thread 7 | INFO | CommonLib.UfaSubsystem.UfaSubsystem - UfaSubsystem: connecting to UFA service.
Hrmm... couple of questions about the parent:
1. Is the local firewall disable via services.msc or the appropriate ports opened?
2. did you create a new snapshot with it powered off?
3. Verified both forward and reverse lookup?
I know it sounds weird but I wonder if it got a partial success (fluke) on the first attempt but something has happened to DNS or some other component since then.
Is there any error after the attempt to connect to the UFA (it's part of the Quick Prep process)?
Hi,
No just stopped. I tried to restart the UFA service but that didn't work (just sat at stopping). I couldn't even kill it or end process, which would suggest it was in a bad way.
I have rebooted the server, and trying to provision a pool again.
Graham
Hi,
Right, been back through the logs, and on Thursday when I started to have the error I found these two lines next to each other. After that, it appears to stop at the 'connecting to UFA service' every time.
Graham
2009-03-26 14:38:09,754 | UFA thread 1 | INFO | CommonLib.UfaSubsystem.UfaSubsystem - UfaSubsystem: connecting to UFA service.
2009-03-26 14:38:09,754 | UFA thread 1 | INFO | CommonLib.UfaSubsystem.UfaSubsystem - UfaSubsystem: restarting UFA service.
Hi,
Hooray! It is now proceeding past were it was hanging before.
Thank you very much for your help.
Now, do you know if there is a log for the UFA service so I/We can troubleshoot why it crashed? Doing it now during trialling is one thing, doing during a reset of 200 machines is another.
Thanks
Graham
It seems to be the result of resetting while provisioning. Check in one of the VMs after it's fully provisioned to see if the agent recorded something (UFA would run in there I'd imagine -- otherwise, you'd have to increase the logging of the Composer component). See page 169 on how to collect full logs from Composer.
I believe UFA is the converter service.
did you create the guest system (parent vm) by converting it from a physical client? If you did, and if you didn't check "remove converter after conversion" or even if you did (it doesn't always remove in my experience) that would explain the ufa service in the clones;
It is however not needed and may even be a nuisance so remove it from the parent VM if you don't intend to use it anymore. Create a new snapshot and point your desktop-pool to use this new snapshot. If all goes well View Manager will start using this snapshot as soon as there's a request to deploy a new client.
kind regards
Jesse
Hi,
No, clean build VM with XP SP3 32bit, no converter ever installed. The UFA process that hung was on the VC server, not inside the VM.
Graham
I wonder if it was the quick prep process. The UFA was likely tied to Composer, not vCenter (since they are on the same server -- processexplorer could show you which component it was a child of).
Hi,
Having got it all functioning again (without the zombie UFA process) I have setup all the pools again and tried to re-compose a pool against a new snapshot... worked like a dream.
I hate non-reproducible bugs.
Graham
LOL.. I know how you feel. I suspect you may be able to reproduce it if you try to refresh it as it's deploying.
Anyways, if you do hit it, lemme know. I'm still curious as to the cause and the resultant error messages.