Good morning,
I've got 85MB of space available on the datastore where my virtual machine is located.
I can not make backups with Veeam Backup because I have no space for snapshots.
Anyway, I made a consolidation (has been executed without errors), but in the snapshot manager snapshot still appears and I have not managed to free space on the datastore.
The virtual machine disks are pointing to -000001.vmdk
I attach images of the configuration of the virtual machine, snapshot manager, datastore files and .vmx file settings
What can I do to free up space on the datastore?
Which files can I safely delete?
Besides all this, as you see in the datastore.jpg file, I have two files with the same name that the 2 disks but ending in _working.vmdk. These files have date of March 2012. I have copied them to another datastore and i have tried to delete the originals, because I have not seen any reference to them in the VMX file , but I can't delete them. In addition, when I have copied them to another datastore, the modified date has changed.
I hope someone can help me because I can't do backups or replicate the virtual machine and I am quite concerned ....
Thanks and best regards,
Leire
Hi there,
Does it allow you to Delete All in Snapshot manager to get rid of the snapshot? Typically consolidate is only needed once the snapshot has been deleted and the consolidation has failed during the delete action.
If you need to clear some space you can move or delete the vmware-xx.log files safely which might give you some breathing room, the current log is vmware.log but all else can go.
Hope this helps.
vM
-----------------------
VCAP-DCD / VCAP-DCA / VCP-CLOUD / VCP-DT / VCP5 / VCP4
-----------------------
vMustard.com
Hi martinriley,
I have deleted some logs but I need more space
What would be happend if I delete the snapshot? Do I lose data? And, can I do it with the virtual machine working?
Thanks for your answer
Leire
So if you choose 'Delete All' from the Snapshot manager essentially what will happen is that it will merge all changes made to the VM since the snapshot was taken from the delta file (the xxxxx-00001.vmdk file) into the original VMDK file of the VM, you should not delete the snapshot file directly.
The reason you are running out of space on the datastore is that since the snapshot was taken all disk changes etc. to the VM is being written to the new file instead of changing the data in the original file, if left long enough the snapshot delta will grow and grow consuming more space, in theory growing up to the size of the original VMDK. So deleting the snapshot via Snapshot Manager will 'commit' all of these changes and revert back to the original VMDK with all the changes incorporated. I hope that makes sense!
You can delete the snapshot at any time whilst the VM is running, you may see a slight dip in performance whilst the snapshot is consolidated due to the additional IO overhead but yes you can do do this with the VM up and running. You need a certain amount of free space on the datastore to delete the snapshot, hopefully you have enough if not you may need to clear some more room somehow. Also I know in Snapshot Manager the snapshot name states 'DO NOT DELETE' or similar, safe to ignore this now as I'm assuming the VM is not in the process of being backed up? Unless it is, in which case Veeam will delete the snap once it's done! ![]()
Hope this helps
vM
-----------------------
VCAP-DCD / VCAP-DCA / VCP-CLOUD / VCP-DT / VCP5 / VCP4
-----------------------
vMustard.com
There is a file on the datastore, in most cases, equal in size to the RAM of the VM. It is the *.vswp file.
One potential way to free up space on a datastore is to set the VM Memory reservation (VM Settings > Resources) equal to the allocated RAM. This causes the *.vswp to go to a size of 0, after a full power-cycle of the VM. If you have more than one VM on the datastore, you can try this on several VMs at once and potentially recover just enough datastore space to delete to delete your remaining snapshot.
As the other posters have indicated, "delete all" will consolidate all of the "redo" logs (A.K.A. *00000*.vmdk snapshots) into a single disk. If it works, you will move forward in a continuum.
OK. This night I'll try to delete snapshot with the snapshot manager.
I'll stop before the vm to delete the vswp file. I think that I'll have 4GB of free space more or less.
Do you think the snapshot manager will have enough free space to delete the snapshot?
What would happen if there is not enough space?
And, what can I do with working.vmdk files?
Thanks
Leire
Do you think the snapshot manager will have enough free space to delete the snapshot?
There may not be...
Keep in mind that you're running with thin provisioned disks. The primary .vmdk where approx. 32.4GB is used but 41.9GB is allocated might try to blow up in full size when merging the content of the delta...
/Rubeck
So, what can I do?
How can I delete working.vmdk files? I have a copy of them in other location.
If I delete them, I should have enough space to delete the snapshot
With that small amount of free space that you have available I would not risk to use any snapshotmanager operation.
A VMFS-volume that gets filled to the max in a snapshot-operation can damage one of the vmdks - and even worse.
If this case were my task I would allow a safe procedure only.
I would split the existing single VM with two vmdks into 2 dummy VMs with one vmdk each.
Then I would copy one of the 2 basedisks to a temporary location - a Windows-fileshare, NFS-server or even to a USB-disk of your admin host.
Once that is done there is enough free space to consolidate dummy VM 1.
When you are finished with the first - upload the file you copied out of the way before and consolidate the second dummy VM.
Once that is done you can delete the 2 new vmx-file you created for the 2 dummy VMs.
When that is done you can start using the original VM again.
Along the way you have to edit vmx-files and descriptorfiles .... if that is new stuff for you it is probably better to ask someone for assistance.
Probably you ask yourself how I can call this procedure safe ?
I admit that the risk to make a mistake while doing all the small modifications that are required can not be neglected.
But I promise that if you think about it a while it starts to make sense.
This approach has a risk that a user mistake results in bad vmx-files, bad vmdk-descriptorfiles and other helper files.
But you can easily backup all the small configfiles if you want extra safety.
This scary looking aproach has the big advantage that it splits the consolidation job that at the moment requires luck and a lot of crossing fingers, into small tasks that do not require luck at all.
Actually its quite straight forward ....
I see that you used Datastorebrowser to create the screenshots ... do yourself a favour and never use Datstorebrowser when you face a task like this.
In my recovery work one of the most important steps I do at every environment I visit is to install WinSCP and Putty - or go home at once if I am not allowed to install those two.
I have connected a NAS to the server so now I have a new datastore with 1.7GB of free space.
What is the best, clonate or copy the virtual machine to the new datastore?
