Getting "Invalid Fault" error when attempting to perform V2V conversion with Converter Standalone 5.5.3

[quantum_it]

I am attempting to migrate VMs from an older VMWare environment, using an ESXi 5.5 host (build 1331820), using the vcenter converter standalone utility (version 5.5.3), to a new ESXi 5.5 U2 (build 2068190). The new environment has a vCenter 5.5 server running on it, and the new ESXi 5.5U2 server has been registered with it.

The data store sizes are exactly the same on the old and new ESXi servers, just using separate SANs (the new environment is meant to be a 100% clean restart for this environment).

However, when I attempt to migrate a VM (version 😎 from the old environment to the new environment, and convert it to hardware version 10, I get an "Invalid fault" error from the standalone converter.

The only line I can find in the error log before the 'invalid fault' entry is:

2015-04-14T15:36:01.267-07:00 [08268 info 'wizardController'] FillPage()

2015-04-14T15:36:11.637-07:00 [00400 error 'wizardController'] Unable to submit job: vmodl.fault.SystemError

2015-04-14T15:36:11.637-07:00 [10620 error 'ConversionState'] Submit job failed: A general system error occurred: Invalid fault

2015-04-14T15:36:11.637-07:00 [10620 error 'ConverterWizardWindow'] Next/Finish failed: A general system error occurred: Invalid fault

2015-04-14T15:36:18.559-07:00 [09980 verbose 'HttpConnectionPool-000008'] [RemoveConnection] Connection removed; cnx: <SSL(<io_obj p:0x0390ff28, h:-1, <TCP '0.0.0.0:0'>, <TCP '192.168.0.10:443'>>)>; pooled: 0

I'm stumped as to what else I should be looking at, in regards to migrating and convering VMs from my old environment to the new environment.

[rcporto]

Your VM have any GPT disk ?

Anyway, if this VM are running from a SAN, why not just present the datastore storing this VM, remove the VM from the old host, register on the new host, and migrate the VM using Storage vMotion to the new datastore on the new SAN and then upgrade the VMware Tools and the Hardware Version.

[quantum_it]

I was hoping to have avoided going through all the trouble of setting up the old ESXi environment and doing vMotion, by using the stand alone converter.   It's only 8-10 VMs that I'm dealing with.

None of the VMs are GPT partitioned.  They are all Windows 2008 R2 server and Windows 7 hosts.   I'm suspect as to why this isn't working, since I was able to do the 'exact' same type of migration at another location between an older ESXi server, and a new ESXi 5.5U2 server, without incident.

[POCEH]

Could you upload log bundle?

[quantum_it]

Is there a way for me to PM to you, so that I don't potentially expose any of our internal details?

[quantum_it]

The target ESXi host is registered with a vCenter server (running as a virtual machine on the target host).   Does the target ESXi host need to be disconnected from vCenter in order to use the standalone converter?

[POCEH]

Yes, send me the logs as PM.

[POCEH]

No, this is not needed, moreover after VC version 5.x (not sure exactly) if the ESX host is managed from the VC you can't add a VM to the ESX directly but from the VC only.

HTH

[quantum_it]

I installed the converter on a 100% clean Windows 8 workstation and attempted to run the migration and received the exact same 'Invalid Fault error'.   However, this time I noticed the following lines in the converter-worker log:

2015-04-17T11:55:08.779-07:00 [04088 info 'Default'] TargetVmManagerImpl::CreateVM: Creating the target VM.

2015-04-17T11:55:08.779-07:00 [04088 info 'Default'] Reusing existing VIM connection to vcenter1-sj1.quantumscape.com

2015-04-17T11:55:08.779-07:00 [04088 info 'Default'] ComputeResource for vim.HostSystem:host-43 is vim.ComputeResource:domain-s41

2015-04-17T11:55:08.779-07:00 [04088 info 'Default'] ResourcePool for vim.ComputeResource:domain-s41 is vim.ResourcePool:resgroup-42

2015-04-17T11:55:08.779-07:00 [04088 info 'Default'] Creating VM OpsCenter-sj1 in folder vim.Folder:group-v22 on host vim.HostSystem:host-43 and resource pool vim.ResourcePool:resgroup-42 (isVirtualApp = false)...

2015-04-17T11:55:11.248-07:00 [04088 error 'Default'] CreateVm task failed: The operation is not allowed in the current state.

2015-04-17T11:55:11.248-07:00 [04088 error 'Default'] TargetVmManagerImpl::CreateVM: Vm creation failed with fault vim.fault.InvalidState

2015-04-17T11:55:11.248-07:00 [04088 info 'Default'] Scheduled timer canceled, StopKeepAlive succeeds

2015-04-17T11:55:11.263-07:00 [04088 info 'vmomi.soapStub[14]'] Resetting stub adapter for server <cs p:0168a5b8, TCP:vcenter1-sj1.mycompany.com:443> : Closed

2015-04-17T11:55:11.263-07:00 [04088 error 'Ufa'] Method converter.worker.ConverterWorker.createVM threw undeclared fault of type vim.fault.InvalidState

2015-04-17T11:56:02.627-07:00 [05272 info 'Default'] [Converter.Worker.DiagnosticManagerImpl] Generating LogBundle.

So it almost looks like the converter isn’t able to create the new VM on the target ESXi host, but I’m not sure why.  This is a brand new target host with nothing running on it except a clean build of ESXi 5.5U2, and a single VM running the latest version of vCenter 5.5.   I’ve created a new LUN from my storage and mounted it to the iSCSi storage adapter for the target VM that I want to migrate and everything there looks like it's prepped.

[quantum_it · Accepted solution]

So the issue was with the vCenter server (the Linux appliance version, which is a little disconcerning) not being in a 'ready state' to allow for migrations.

I restarted the vCenter server and everything started working as expected, immediately.