vmwarerules2011
Contributor
Contributor

Making sense out of datastore info

Hi,

Sorry for the newbie question. I have a VM that is full on storage and refuses to boot up. From the screenshot below, could you advice whether there is a way to grow the datastore further so the bootup can start? I am hoping there is a way to increase datastore size so 373GB-227GB=146GB can be partially allocated to the VM thus the VM can boot up again for cleanup?

1792431.jpg

Thanks!

0 Kudos
12 Replies
bilalhashmi
Expert
Expert

This might be happening cuz your swap file can't be created. Try moving the swap file to a different datastore. You can also look at extents to extend your datastore so that you can fix the issue. Not sure if moving the VM to a diff DS is an option for you or not.

Follow me @ Cloud-Buddy.com

Blog: www.Cloud-Buddy.com | Follow me @hashmibilal
0 Kudos
AndreTheGiant
Immortal
Immortal

You must keep at least the 10% of free space in your datastore (more suggest 20%).

This bacause thin disks, snapshots, VM swap files may fill the datastore and create several issues.

Andre

Andre | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
0 Kudos
a_p_
Leadership
Leadership

To give you the appropriate advices, we first need to understand your current setup.

  • Is this a standalone host or is it running in a cluster (vCenter Server)?
  • Which version/build of ESXi are you running?
  • How much memory do you have in the Host?
  • How many virtual machines do you run on this host and how much memory did you assign to each of these VMs?
  • Do you have snapshots on the virtual machines?
... thus the VM can boot up again for cleanup?

This sounds as if you are using thin provisioning!? Deleting files/cleaning up the virtual machine's guest OS does not result in more free disk space on the VMFS datatstore!

It may be possible to prevent the VM from creating a swap file by configuring a memory reservation for it. However, IMO with only 12 MB of free disk space you are running a high risk and immediate action/cleanup is required.

André

0 Kudos
vmwarerules2011
Contributor
Contributor

Hi,

Thanks for the replies. To answer the questions:

- This is the standalone setup

- This is ESXi 4.0.0 Build 208167

- The host has physical RAM 4GB and 3GB is used for the VM

- There is only one VM on this host and 3GB is assigned to it

- It has 4 snapshots since it's been up a year ago.

When you said "immediate action/cleanup is required," do you have pointers or URLs I can look up?

I think I really need to understand what's under the hood better.

If you don't mind, I have trouble finding to a way to reply to my thread on the website. Is it by design?

Thank you again!

0 Kudos
bilalhashmi
Expert
Expert

vmwarerules wrote:

Hi,

Thanks for the replies. To answer the questions:

- This is the standalone setup

- This is ESXi 4.0.0 Build 208167

- The host has physical RAM 4GB and 3GB is used for the VM

- There is only one VM on this host and 3GB is assigned to it

- It has 4 snapshots since it's been up a year ago.

When you said "immediate action/cleanup is required," do you have pointers or URLs I can look up?

I think I really need to understand what's under the hood better.

If you don't mind, I have trouble finding to a way to reply to my thread on the website. Is it by design?

Thank you again!

Do you have a different DS where swap file can be moved to? That will give you 3GB of the room the swap file will need. One thing you can try is reserve 3GB for this VM and then try turnig on. This will not rquire the swap file to be created which might be the reason why you can't turn it on.

Snapshots are a big space takers.. plus its highly recommended not to keep snapshots for ever. They are not meant to serve as backups by any means. If you can, try commiting some snapshots, that should also give you more breathing room..

But all in all, you will have to ensure that you have enough DS free space 15-20% is whats reommended. Or else you will encounter some  unexpected behaviour. Do you have any other DS connected to this host. From your post it sounds like you dont, But I figured I will ask..

Not sure about your query on finding hard to reply to the thread..

Follow me @ Cloud-Buddy.com

Blog: www.Cloud-Buddy.com | Follow me @hashmibilal
0 Kudos
a_p_
Leadership
Leadership

When you said "immediate action/cleanup is required," do you have pointers or URLs I can look up?

Sorry, but there's no rule of thumb for issues like this. It always depends on the situation.

Now that you described your setup, please post a screen shot of the datastore browser window which shows all the VM's files. Please make sure the screen shot shows the full names, sizes as well as time/date of all the files. In addition to this screen shot, please attach the vmsd file, the vmx file and the log files (please archive/zip the files and attach the zip file).

André

If you don't mind, I have trouble finding to a way to reply to my thread on the website. Is it by design?

To find your own threads or threads on which you posted, select "My Discussions" under your Login name (upper right).

0 Kudos
vmwarerules2011
Contributor
Contributor

Hi,

Attached are two files:

- a screenshot of the data store files.

- the zip file with logs, vmx, and vmsd files.

Thanks!

0 Kudos
a_p_
Leadership
Leadership

There's a good chance to be able to delete the snapshots without needing more disk space.

One quick question in advance. Do you have the chance to update the host to at least ESX 4.0 Update 2 (build 261974)? This would make it very easy to resolve the issue without any manual editing.

In any case, what you should do at this point is to see whether the VM is still ok after the out-of-disk-space shutdown. To be able to power on the VM you may delete the four vmsn files (~3GB each). I assume you do not intend to revert to any of the old snapshots!?

Please do not continue to work with the VM, this would most likely fill up the datastore again. Only make sure everything is ok, then power it off again.

If you cannot update the host for whichever reason, attach the vmware.log file, which will be created after powering on (and off) the VM. A sreenshot of the VM's virtual HDD configuration (Edit Settings -> Hard Disk) which shows the virtual HDD size would also help. This is for me to double check the virtual HDD is configured as a thick provisioned disk with a size of 150 GB.

André

0 Kudos
vmwarerules2011
Contributor
Contributor

There's a good chance to be able to delete the snapshots without needing more disk space.

I've deleted a snapshot from Snapshot manager but no space is freed up for some reason.

Do you have the chance to update the host to at least ESX 4.0 Update 2 (build 261974)? This would make it very easy to resolve the issue without any manual editing.

Could you elaborate what I can do to fix if I upgrade to the build you recommended?

In any case, what you should do at this point is to see whether the VM is still ok after the out-of-disk-space shutdown. To be able to power on the VM you may delete the four vmsn files (~3GB each). I assume you do not intend to revert to any of the old snapshots!?

I do not need to revert to any of the old snapshots. I've deleted an old snapshot for testing and confirmed the .vmsn file corresponding to it is gone but no space is freed up - it still says

provisioned space: 373.43GB

not-shared storage: 227.18GB

used storage: 227.18GB

free: 12MB

Attached is the virtual HDD size screenshot.

---

ESXi Config Guide:

"Used Storage – Shows datastore space actually occupied by virtual machine files, including

configuration and log files, snapshots, virtual disks, and so on. When the virtual machine is running,

the used storage space also includes swap files."

Could you explain why there are only <100GB when I add up the files in the datastore but it complains there is no space?

Also, is it true deleting snapshots do not free up space in datastore? Based on the above excerpt from ESXi Config guide, it sounds like it should free up some space.

Thanks as always - I find this community is the best around because people are fast to respond and are polite to provide useful helps.

0 Kudos
vmwarerules2011
Contributor
Contributor

Regarding the file sizes do not add up, my bad, they do add up to 227GB. I just don't understand what happened to 373GB-227GB=147GB and removing snapshots do not seem to help.

Thanks,

0 Kudos
john23
Commander
Commander

You have deleted the snapshot from the snapshot manager. but seeing the screen shot still delta and vmsn is in datastore.

Try this steps mention in the kb (may solve the issue):

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=100504...

Thanks

-A

Thanks -A Read my blogs: www.openwriteup.com
0 Kudos
a_p_
Leadership
Leadership

The reason you cannot delete individual snapshots to free up disk space is, that the vmsd file is corrupted. So what we need to do is either to "repair" this file or to update the host to ESXi 4.0 Update 2 or later. With Update 2, VMware modified the the way "Delete All" works. With this - IMO major - change you don't need (a lot of) additional disk space when running "Delete All" with a thick provisioned base disk.

André

0 Kudos