Could not power on VM: No swap file. Failed to power on VM

[bkelleman]

Hello --

A coworker did a simple "Power Off" on 2 live VM's. Now when they are powered on we are receiving the error " Could not power on VM: No swap file. Failed to power on VM "

The vswp file is present in the datastore.

Any guidance or help would be appreciated.

So far we have tried

--Cold migration

--Rename files

\-- Deleting the vswp (Invalid Argument error came back, file still present)

PLEASE Help!!

[sbeaver]

Unregister the VM and if you need to move the vmdk using vmkfstools and the vmx to a new directory and register the VM and then start

[bkelleman]

Thanks for your quick response. How do I unregister a VM safely?

[sbeaver]

Right click on the VM and remove from inventory but do not delete the files.

[bkelleman]

Ok, maybe this is the issue?

\[root@esxnj04 jes-mail]# vmkfstools -i /vmfs/volumes/45797ed5-bcb1531d-6716-00144f407010/jes-mail/jes-mail.vmdk /vmfs/volumes/45832dee-8f2fbae7-e9de-00144f407010/jesmailnew/jesmailnew.vmdk

Destination disk format: VMFS thick

Failed to open '/vmfs/volumes/45797ed5-bcb1531d-6716-00144f407010/jes-mail/jes-mail.vmdk': Invalid argument (1441801)[/b]

\[root@esxnj04 jes-mail]#

I am stumped

[sbeaver]

Check the vmdk header file on that one and one that works and see if anything jumps out to you that something is wrong

[sbeaver]

Did you move the vmdk to another folder and how did you do it? You might need to update the vmdk of the new file path or the vmx

[kitcolbert]

Are you sure the VM powered off properly before? 'cat /proc/vmware/sched/cpu' on the server where the VM was last running to see if it still is (or to see if it's in some sort of zombie state).

[ph0rse]

I have this same problem and under that proc file you mention the VM shows up. How can I manually bring it down now that its process has already been killed?

[danam]

For the record, below worked for me:

http://thelifecode.blogspot.com states:

[jpbutler]

If the VM shows up in the /proc/vmware/sched/cpu how can you bring it down after the process has already been killed?

Thanks.

[RParker]

**A coworker did a simple "Power Off" on 2 live VM's. Now when they are powered on we are receiving the error " Could not power on VM: No swap file. Failed to power on VM " **

Have him drawn and quartered immediately!

Take away his access rights and only give him "Look but don't touch" rights ..hehe..

[YeDai]

Hi

I got this problem early this week since I rebooted the ESX server without turning the VMs off. After the ESX server robooted, here is what I saw and what I did:

-

---

1. windowsxp (orphaned)

I right click the windowsxp (orphaned), then choose Remove from Inventory. Then I find the folder of windowsxp in the datastore, manually add the windowsxp into the inventory.

Then I trying to power on windowsxp, I got the same error message: "Could not power on VM: No swap file. Failed to power on VM ".

Since the windowsxp is off, what I did was that I try to migrate the windowsxp to another ESX server, you will get error message "A general system error occured: internal error". After that, the windowsxp will be automatically powered on properly

-

---

regards,

Levi.

[Park3r]

i had the same issue yesterday, was reconfiguring luns on SAN and cold migrating vm images here and there. 1 vm didn't make it back up though.

inspired by the earlier posts the following worked for me:

#1 cat /proc/vmware/sched/cpu | grep <machinename> to indentify the host running the zombie vm.

#2 let the host enter maintenance mode to migrate all vms off the host.

#3 step #2 fails (at least in my case) with 'operation timeout' even there are no more vms running on it and you keep clicking on enter maintenance mode.

#4 as there is only 1 (zombie) vm running on the host, which is not really running anyway, reboot the host. when it comes online, power on the vm

Message was edited by: Park3r

[skywalkr]

Be careful w/ using kill.. VMware support suggests several options, but the preferred method I've been handed a couple of times is as follows b/c it will clean up the vmworld setups etc.

Q: We have a VM which is in a hung state, what is the safest way to terminate or kill it?

A: For ESX, the safest way to terminate or kill a hung VM is to use the following VMware commands:

1. Login to the hosting system's service console as root

2. cd \tmp

3. Obtain the vmid of the VM you want to kill by typing vm-support -x

4. Kill the VM and generate core dumps and logs by typing vm-support -X XXX where XXX is the VMID you found in the last step.

1. You will be prompted for the following:

1. Do you want to include a screen shot of the VM? -> N

2. Do you want to send a NMI to the VM and send a ABORT to the VM? You must answer Yes to the ABORT question to kill the VM.

2. The entire process will take about 5-10 minutes to run.

3. The process will create a tar archive in the directory you run it in.

4. Delete the tar archive when finished if not needed.

Later,

GC Mobley, IBM

[Narkis]

Use the command vmware -l ( on all host to find the VM Need to know which Host currently has the VM )

if you indentified the host on which your target VM running, use the command ps -efl|grep <VMname> ( It will give the Process ID )

To get the Process ID

Use the command Kill -9 < PID >

Power on the VM in VC

Cheers!

[jcmoriaud]

@14: skywalkr@us.ibm.com

You saved my life Thanks and best regards, JCMoriaud