ESXI snapshot running for one year

[mikrotech]

Hello all,

I have a ESXI host with two virtual windows machines. One of the machines runs MS exchange and a snapshot was taken 1 year ago.

The host has run out of space on the datastore and i'm looking to remove the snapshot which is 100GB. I realize that the snapshot needs to commit it's changes to the VM. This is a production server.

What is the best way to go about this?

Regards,

Andy

[Troy_Clavell]

You may find your best bet to commit that snapshot is to use VMware Converter to move the guest to another LUN. Once it has been moved, power it back on and ensure everything is functional. If not, alteast you have the original to roll back to.

Commiting a snapshot that has been active for a year, would more than likely cause the server to go down due to heave Disk I/0 during the commit process. Besides it may take a VERY long time to commit it.

[mikrotech]

Hello Troy,

Thanks for the response, this is a single server with an internal disk array so I believe it only has one LUN if I am understanding it correctly.

Is there another way?

Thanks

Andy

[Troy_Clavell]

that's a tough one. You can try to commit the snapshot, but with something that old, I wouldn't be to sure of the outcome.

[mikrotech]

I could create an image of the VM using a third party software like Acronis delete the vm, create a new VM and apply the image.

I could also setup a second temporary ESXI server and convert the VM to a new VM on the new ESXI host as you originally suggested. Will this work? I've never used VMware converter.

[Troy_Clavell]

I could also setup a second temporary ESXI server and convert the VM to a new VM on the new ESXI host as you originally suggested. Will this work? I've never used VMware converter.

I like this option!! But it would require downtime for the guest.

Here's some information on Converter

http://www.vmware.com/pdf/converter_standalone_guide401.pdf

...and the download

http://downloads.vmware.com/d/details/converter401/dGhiZHRwaGJ0QGg=

[a_p_]

Andy,

did the host run out of space while the VM was powered on or did you do a clean shutdown of the system when you found out that space was low?

If the VM crashed, there's a high chance of data corruption in the VM. Hopefully you have an up to date backup.

The only way to merge the snapshot to the base VM will be a migration as Troy already recommended.

Committing the snapshot will most likely fail due to insufficient disk space, since the snapshot file itself will exist until the commit operation has finished.

The Acronis image you mentioned will also be a possible option.

If you need free space on your datastore, you may set the memory reservation for a VM to the full amount of the assigned RAM.

André

EDIT:

Please forget my comment on insufficient disk space while commiting the snapshot if you are not using thin provisioned disks. ))

[mikrotech]

The clone timed out.

I was running it as a "powered on machine"

I added a NAS and moved the other virtual machine so I do have more space but the server is still running with the snapshot.

Any other suggestions?

Thanks

[HannaL]

You could try deleting the snapshot again now that you have more space it's actually a good sign the VM is still running. If you don't have any backups then before doing that if you have sufficient free space to make a hot conversion of the VM as if it is a physical server that should at least make sure you can save what you have now since it is up and running. A year old snapshot is not good, If it hasn't crashed already it probably will so either delete the snapshots or make a copy image of the running vm as people have suggested. Now that you have more space If you do a delete all operation be very very patient. A snapshot running that long and depending how busy the VM is still could take many hours. This might be a good time to contact support if you don't have any backups of this production VM.

Hope this helps

Hanna