Problema borrando todos los snapshot con ESXi 3.5

[lkdml]

Estimados, estoy necesitando alguna sugerencia sobre el problema que tengo. Resulta que ante la necesidad de liberar espacio en el servidor, cometí el error de borrar todas las snap de 1 VM, sin calcular el espacio disponible inicialmente. Por lo cual, el resultado no fue el deseado.

Hoy me encuentro con una situación que no puedo resolver, por más que busqué información en la web.

El problema es el siguiente:

Situación inicial: 200GB libres en el servidor

VMx

Disco Base de 200GB

Snap 1 --> + 2gb

Sanp 2 --> + 10 GB

Snap 3 --> + 50 GB

Al borrar todas las Snapshot de la VMx, el ESXi, comenzó el proceso, pero en el camino se quedó sin espacio, por lo cual el resultado no fue el deseado. Y lo que tengo en mis manos ahora es algo que no comprendo.

VMx

Disco Base de 200GB

Snap 1 --> > a 16Kb

Sanp 2 --> + 100 Mb

Snap 3 --> > a 16Kb

Al arrancar la VM, me encontré con un clásico problema de CID, pero es solucionable. Lo malo es que al arrancar la VM, me encuentro que es la VM tal cual como se encontraba antes, sin cambio alguno. En otras palabras, la VM recién instalada sin info.

Alguien me puede dar una orientación de por qué sucedió esto, y si hay alguna forma de recuperar la info?

Por lo que leí si no había suficiente espacio en el servidor, se debería cancelar el proceso, y no borrar los archivos. (El único problema que hubiese tenido es el de la CID corrupta).

[a_p_]

I hope you don't mind me answering in English. My Spanish isn't good enough.

The issue is caused by the way "Delete All" worked in "older" builds of ESX(i) 3.5 and ESX(i) 4.0. The snapshots were committed into their parent, starting from the latest snapshot to the oldest one. Therefore the needed disk space would increase with each snapshot.

This issue has been resolved in one of the latest patches of ESX(i) 3.5 and ESX(i) 4.0

from http://kb.vmware.com/kb/1020052

Removing all snapshots from a virtual machine with the Delete All option can use large amounts of disk space.
When using the Delete All option in Snapshot Manager, the snapshot farthest from the base disk is committed to its parent, causing that parent snapshot to grow. When that commit is complete, that snapshot is removed and the process starts over on the newly updated snapshot to its parent. This continues until every snapshot has been committed. This method can be relatively slow since data farthest from the base disk might be copied several times. More importantly, this method can aggressively use disk space if the snapshots are large, which is especially problematic if a limited amount of space is available on the datastore. The space issue is troublesome in that you might choose to delete snapshots explicitly to free up storage.

This issue is resolved in this patch in that the order of snapshot consolidation has been modified to start with the snapshot closest to the base disk instead of the farthest. The end result is that copying data repeatedly is avoided

With that in mind, I see three solutions to your issue:

1. Upgrade your host to the latest patch

2. Clone of the virtual disk (including the snapshots) into a new vmdk

3. Recreate the vmsd file based on the CID chain in the header/descriptor vmdk files and delete the snapshots one-by-one, always selecting the one closest to the base disk.

If you want to go with solution 3 and would like some help with re-creating the vmsd file, please attach the current vmsd, vmx and all the header/descriptor vmdk files. A list of all the files in the VM's folder would also be very helpful.

Since you mentioned a CID issue, do I assume correctly that the VM is currently up and running again?

André

[lkdml]

André, thanks for the answer, my english is very poor to write, but not to read.

The information that you post, i know , becouse i read to much to understand how it works ESX Snapshot, but my problem is how to rollback.

in other words, my backup is too old, and i need recovery the last state.

Or to understand little more, why hapend that.

thanks.

Mariano

[a_p_]

Mariano,

I don't think there's a way to rollback.

What is the current state of your VM? Is it broken or does it work?

If it works, we should be able to merge the snapshots into the base disk. If you have another datastore with enough capacity to hold the 200 GB disk, I would prefer to clone the disk. This way the original disk is not touched in case the clone does not work.

André

[lkdml]

that´s the problem, I made to work it,  but the base_disk, is the original, (without information) and the snap they aren´t siginificant size.

I think that´s no solution, but i don´t understand why happend that. why ESXi erase the snap, without previously entered in the disk base.

[a_p_]

After reading your initial post again, I have to apologize. I misread the snapshot sizes (GB instead of MB).

You are right, this should not have happened and to me this looks like a serious issue in your build of ESXi when running out of disk space.

In this case only VMware could probably help. Sorry.

Maybe you can find any hints on what happened in the vmware*.log files in the VM's folder.

André