Cold Migrate Failed - Need Help

[DanReynolds]

I have vSphere Essentials and recently upgrade to 5. I have one host that is vSphere 4 and one that is vSphere 5 (both are HP DL160 servers, older one is Xeon E5520 and newer one is Xeion E5606 CPU).

I just put the newer host into production and attempted to cold migrate some machines across. This is essentials so I don't have vMotion.

The took forever and then failed. And sort of failed back - the VM's are grayed out and say Invalid. Error was "A general system error occured: The system returned an error. Communication with the virtual machinen might have been interrupted."

I tried removing the machine from inventory, navigating to the datastore and the VM's .vmx file and "adding to inventory" manually. But I still get the grayed out INVALID.

It appears the VMX file is empty.

What is causing this?

What can I do to fix it.

As for the VM's I can just create new VM's and attach the old VMDK files... But that is sure messy.

I don't have support on my Essentials bundle. This kind of sucks.

[Dave_Mishchenko]

A couple of links to look at.  It may be that one of the hosts has a lock on the file.

kb.vmware.com/kb/1002367

kb.vmware.com/kb/1003742

[DanReynolds]

Dave thanks. I'll take a look at those plus I saw some other's referenced when I looked those up that might give me a clue as to what happened.

And the file locking sounds like a plausable reason - but two of the VM's I migrated weren't even powered on when I started...

A little more background. I just built this new host to balance the load a little better and give me N+1 for patching, downtimes, etc. I was assured that the Xeon 5520 and 5606 were compatible. But I don't think the CPU has anything to do with this.

I'm going to check the other VMX files but it looks like in the process of migrating them it got to the 10% mark, hung and somehow wiped out the VMX files. That's not a good thing in my book.

I've "restored" one VM but I did that by removing one of my 'invalid' VM's from inventory (not disk of course), creating a new "dummy" VM and adding the old VMDK file as a 'new' disk to the dummy VM. The machine is just fine but what happened to the VMX?

I'm cross posting this to the Spiceworks forums as well, BTW.

This is just weird. That will be real unhandy if I can't even cold migrate a VM!!!

[DanReynolds]

Yep, all the VMX files are empty spaces (where the characters used to be).

This really sux.

[continuum]

wiped vmx files ?

if you still have vmware.log files use this tool I made
http://sanbarrow.com/files/log2vmx.zip

[Virtualinfra]

Ulli Hankeln wrote:

wiped vmx files ?

if you still have vmware.log files use this tool I made
http://sanbarrow.com/files/log2vmx.zip

Great tool and will be very helpfull.. just tried to see how it work and its awesome..

Thanks for sharing..If you have some tools developed for vmware crash dump analysis please share that too..

[DanReynolds]

Thanks I will definitely give that a try tomorrow. I was trying to remember if a couple of these VM's had two CPU's and how much memory, etc.

This should help me out. And yes, I saw log files, at least with some of the VM's.

I appreciate this.

It appears that maybe this new vSphere 5 host is "bad"... I tried to recreate one of the messed up VM's on it and it didn't work.

Rather than mess with it I think I'll just rebuild it. At this point I'm not out anything.

Is there anything peculiar about building a vSphere 5 host?

I've been building ESX and ESXi hosts for years and years. Never, every had a problem like this.

[Dave_Mishchenko]

Nothing particular with ESXi 5 to worry about.  Have you tried with a fresh VM?  Perhaps something is screwy with the storage.

[DanReynolds]

Yes, I tried creating a dummy VM to bring back one of my "invalid" VMs. Maybe I just screwed up building it or something.

The storage should be fine. It's all the same storage. I will be a little more cautious rebuilding it. I'll tell you that for sure.

It's a proliant DL 160. 32 GB RAM, (2) Xeon 5606 CPUs, added a dual NIC card for the iSCSI storage. I setup the multi-pathing in vCenter, now that is 5 you can set that up in the GUI.

I never even anticipated having any problems. This caught me totally off guard.

[continuum]

Today I hear that the tool I made triggers a false alarm if you let it check by  virus-total.com.
The Antivirus tools from nProtect flag it as Trojan/W32.Agent.735887
I am not sure what to do about it  - maybe I should remove the download ...

[DanReynolds]

If it helps my Kaspersky didn't flag it.

I guess you send that vendor a note telling them to re-evaluate the 'ruling'?