Failed to start the virtual machine error 18

We had an problem where we had actually to manually rebegin our VM server due to a hardware problem. Once the organize was earlier up, I attempted to power on among our VMs and also received the complying with error:

An error arisen while creating temporary file for /vmfs/volumes/5a4cf63c-6357d2b8-5276-b49691125884/WIN2012TEMPLATE/WIN2012TEMPLATE.vmx: The file already exists

Cannot open the configuration file /vmfs/volumes/5a4cf63c-6357d2b8-5276-b49691125884/WIN2012TEMPLATE/WIN2012TEMPLATE.vmx.

You watching: Failed to start the virtual machine error 18

Fairesulted in begin the online machine (error -18).

Other posts I have actually watched states concern deserve to be either not enough space on the datasave, or that the temp file itself (the .vmx~ file) is stuck and cannot be deleted. I've tried SSH to the volume and perform an rm to remove it yet it says the file does not exist.

Below is a screenshot from the SSH session reflecting the directory:


Any ideas on trying to get this addressed is appreciated.

1 comment
67% Upvoted
This threview is archived
New comments cannot be posted and also votes cannot be cast
Sort by


level 1
1 year ago · edited 1 year ago
That "No such file or directory" popping up from doing a catalog listing is saying that you likely have actually filemechanism corruption. You have the right to usage VOMA to perdevelop an analysis and attempt repairs, but the datastore most likely requirements to be quiesced. See what the docs and also tool states around it as soon as you give it a go, though:

On the bbest side, the most important files look to still be accounted for here (both flat and descriptor VMDK files, and second, your VMX) and you can most likely gain the VM to power on, if you can't wait to perform an evaluation and also repair, leaving the VMs off for longer.

See more: Microsoft Office Error Code 30038-26, Microsoft Office Error Code 30038

Run vim-cmd vmsvc/getallvms to display the VMs registered to the organize. I intend this WIN2012TEMPLATE or CADDB01 VM must present up in the list somewhere. If not, then look from the organize that has this VM registered to it, according to vCenter Server. Anyhow, if it shows, note its VMID on the left-many column.

Edit the .vmx file (vi, nano, etc) and also situate the swap line. It's most likely precisely this: sched.swap.derivedName = "/vmfs/volumes/5a4cf63c-6357d2b8-5276-b49691125884/WIN2012TEMPLATE/WIN2012TEMPLATE-80498f61.vswp"

Edit this line and provide the swapfile name a new, unique name. For circumstances, change it to WIN2012TEMPLATE-80498f62.vswp, or rename it to CADDB-swap.vswp, whatever. Save the changes (:wq in vi, or ctrl+x and save, leave in Nano).

Then re-register the VM to the host/VC because we modified its VMX file, with vim-cmd vmsvc/refill wbelow is the VMID you caught from column 1 in an earlier step.

It must power on after this. But I think after that's all sassist and also done, you should look at repairing the datastore; it'd most likely settle this absent file issue also, and let the next power-on attempt actually discover and also usage that missing vswp file. That means you're not stuck editing the VM config and stuff.

See more: Fix: You Need Permission To Perform This Action Network Share

Also a valid time to reach out to Support, otherwise, but this is around what will be said. Perhaps beginning through VOMA, though.