Disappearing storage and VM sizes

[amitair]

Hello,

I am trying to figure out where all our Storage is disappearing to .

I couldn't find any documentation of how exactly are the vm's kept on the datastore/LUN

we have a 1.8 TB LUN

Using vsphere datastore browser, I counted around 240 folders. (vsphere only shows 23 computers on that datastore)

Looking at the files ( vmx,vmdk,nvram...) I assume each one is a seperate VM

I then searched *.vmx on each LUN (assuming it represents a specific vm)

Some machines are duplicated more than twice.

One example is computer17

there are five separate folders with almost identical almost identical computer17 files

The main difference being only holds the a 2GB vswp file

All the rest hold 2GB vmss files

which brings me to my questions:

Does each folder represent a distinct computer?

why does one computer have 5 folders?

How do I prevent a computer from taking up too much storage? can I erase any redundant folders?

Thanks in advance

[dburgess]

The vmss files are for the memory state of a VM if you call the snapshot function while the VM is powered on. I suspect this is causing your space to be consumed. Somebody/thing is calling the snapshot function alot, could be a user or something like a backup product. Normally speaking you would expect one folder per VM and in that folder the vmx file represents the configuration of that VM so that should be 1:1 with the VM's you see from the mgmt interface.

dB

[amitair]

Thank you for the quick reply,

There are no snapshots at all on this VM.

Could these be clones with the same name that were created somehow?

How do I know if it is safe to delete from the datastorebrowser those folders that don't correlate to a specific machine?

[a_p_]

To be honest, I'd recommend you hire a VMware consultant for a day or two to go through your configuration.

It seems there are some things which need to be corrected or explained to the people working with this environment. Deleting the wrong files could end in a disaster.

André

[dburgess]

They could be clones - the vmss files are definitely snapshots so you need to be very careful about deleting these.

My initial thoughts are use the VI client to id all the known servers and their related folders and list of suspect ones (all the others). Then attempt to register the suspect clones etc to either VC or a specific ESX server if you don’t have VC. If this then registers cleanly delete it using the management tool that way you should ensure you do not mess with any of your known 'valuable' VM's.

dB

[amitair]

Thanks,

I'll look into it.

[smokey71]

I would strongly recommend downloading and using RVtools. It's a great app for digging through your infrastructure. The tool will search for snapshots which can save you time as opposed to running a linux command to search for delta.vmdk's or using powershell. The last tab of the tool also has file/folder info that can indicate inconsistent folder names and also identify orphaned vmdk's with the statement: "Possibly a Zombie vmdk file! Please check." It has a great many uses and could help you in your endeavors. As previous posters have stated, please be careful if you plan on removing objects.

Hope this helps.

[amitair]

Thanks for all the help,

RVtools is a very nice app. The problem originated in the fact that the datastore is also used by labmanager, and that led to all the folder duplication. for every snapshot (even if erased) it created a folder with delta disk.