Moving VM definitions and storage files on backend storage

[BDS__Vince]

The ESXi setup I inherited has all of the VMs and storage defined in a single folder structure (\\root\vms)

We'd like to take advantage of the synchronization in the storage system to do some replication and would like to move the VMs into sub folders based on the replication frequencies determined by our business needs.  For example, the new folder structure would be

\\root\vms\30

\\root\vms\60

\\root\vms\120

\\root\vms\240

Is there an Easy way to do this?  Can it be done on running machines?

Thanks!

[vfk]

What storage system are you using? i.e vendor, make & model, and how is your storage presented to the esxi hypervisors?

[JPM300]

it almost looks like openfiler/freenas or something with the \\root\vms\  file path.  Like vfk said, which storage system are you using?

[BDS__Vince]

We're using NFS storage provided by an Isilon stack.

Creating the folders on the back end isn't the challenge. 

The challenge is migrating the existing VM inventory from their old folders to the new folders on back end storage.

Thanks for the replies. 

[vfk]

You can use storage vmotion to move the VM from one datastore to another.

[BDS__Vince]

Storage vMotion will move a machine from one folder on a storage system to another folder on the same storage system?  I didn't think it was that granular....

[vfk]

That is why I asked how is your storage presented to the esxi hypervisors?  If a folder is mounted on esxi through NFS, how is the VM working?

[BDS__Vince]

The volume is presented to ESXi as \\root\vms and all of the guest systems are defined in individual folders at that level.

I'd like to organize the guest systems into sub folders.  Lets assume we have two systems foo and bar

They are currently in \\root\vms\foo and \\root\vms\bar

I want them in \\root\vms\30\foo and \\root\vms\60\bar

Is that possible or do I need to re-think the approach?

[vfk]

In that case you will have mount \\root\vms\30 as a datastore and use storage vmotion.  Any LUN/storage holding VM data has to be accessible to the hypervisors.

The Approach: This all depends how you perform your replication, do you replicate everything or only selected VMs?  And also this depends on your opperational process, the administrative over of making sure that administrator provision VM in the correct for replication or someone accidentally placing VMs in the wrong location because that datastore had enough space.

I think you are probably better off using storage profiles, and probably configuration manual storage capabilities, if VASA is available, and make use of Storage DRS cluster.