ESXi Storage Allocation

[CHiLL]

I've come across an issue where our VMs appear to be over allocated in terms of storage space. The OS disks for our VM servers are located on 15K SAS drives spread across our HP MSA. The data/storage drives are located on 7.2K SAS drives over the same two MSAs. The 15K SAS drives have limited space - and we have very nearly reached that limit (warnings have appeared within vSphere regarding storage limitations).

 

Upon inspecting the VMs, I've noticed that our servers drives/partitions are thick provisioned to a certain size, however the 'Maximum size' value is much higher. I thought thick provisioned pre-allocated the space as could not get any larger. Thus meaning that the 'provisioned space' appears much larger than what we've actually allocated.

 

Maybe my understanding is a bit off, but could someone explain this situation and what I can do to free up unused (if any) space? I've attached a screenshot of an example of one of the over-allocated servers.

 

[ATTACH=CONFIG]47956[/ATTACH]

 

Thanks.

Edited March 5, 2018 by CHiLL

[Norphy]

Has that VM got any snapshots? If so, it might be worth removing and consolidating them.

[john]

Maximum size is the biggest it could be made and is based on the space on the LUN that disk resides on

 

For example:

 

 

That disk is on a 1TB LUN but it shared with others so I can only expand it to that size shown, if I need to go bigger I need to make more space EG Move disks off the LUN, expand the LUN etc

 

Does that help?

 

Things to look for are Snapshots they grow so if you took some ensure you threw them away, if you've disposed of VMs did you remember to tell it to delete the disks?

 

A useful tool is RVTools to help look through and find Zombie VMDKs, issues with Snapshots etc...

[CHiLL]

Thanks for the quick replies.

Has that VM got any snapshots? If so, it might be worth removing and consolidating them.

I have just checked, and there is only one snapshot, which is the 'VEEAM BACKUP TEMPORARY SNAPSHOT'. All the other servers are the same.

 

Maximum size is the biggest it could be made and is based on the space on the LUN that disk resides on

 

For example:

 

[ATTACH=CONFIG]47957[/ATTACH]

 

That disk is on a 1TB LUN but it shared with others so I can only expand it to that size shown, if I need to go bigger I need to make more space EG Move disks off the LUN, expand the LUN etc

 

Does that help?

 

Things to look for are Snapshots they grow so if you took some ensure you threw them away, if you've disposed of VMs did you remember to tell it to delete the disks?

 

A useful tool is RVTools to help look through and find Zombie VMDKs, issues with Snapshots etc...

Cheers, I understand that now. I have noticed that there are multiple .vmdk files per server when browsing the datastore, so I'm thinking that they are taking up the space. How do I remove these old/zombie /.vmdk files?

[john]

Each disk makes an individual VMDK file in the LUN so it is possible you have multiple vmdk's in a LUN. Snapshot also make files in the LUN to track the changes.

 

For hunting Zombies, I strongly suggest you download RVTools - https://www.robware.net/rvtools/ and connect to your VCentre or host if you just have 1 and look at the vHealth tab in that and it shows suspected Zombies as follows:

 

 

Be warned if the LUNS are shared with other hosts that RVtools cannot see, for example if you use VMware without a vCentre server so connect to an individual host, or you use VVOls you may find more marked than you expect so double check before you go deleting anything, should your LUNS be exclusive for the host(s) and if you have vCentre and go via it you should be safe to delete what it says are Zombies.

 

 

RVtools is free so anyone using VMware should have it on your PC its a great thing and shows more than Zombies.