I'm a bit new to this type of vsphere maintenance, but not so much to the concept of raids, disks, etc.
The situation is this: A guest server named "example-FS-01" has a disk on 'esx-datastore-02' This disk is 1 TB in size, the datastore itself has a total size available of 1.5TB There are 2 other unused datastores, each with 1.5TB of usable space.
Each datastore isn't a single disk but a system of disks in a redundant raid.
For example, using vmotion to shift data around is it possible, safe, or close to/far from best practice to expand this guest disk to a size of 2.5TB spanning across datastore 1 and 2?
Thanks
As @joeqwerty has said in the comments, you can't expand a single VMDK across datastores.
You pretty much have two options to get where you want to go:
Create second VMDK
You can create a second VMDK on one of the empty datastores, add it to the same VM, then use the in-guest tools (LVM for Linux, Dynamic Disks for Windows) to extend the first disk with the space from the second.
Add an extent to the
esx-datastore-02
datastoreThis would entail deleting one of the other two datastores, then adding its capacity as a second extent to the first. You can then extend the VMDK in question to use the space you want.
Both of these approaches have the same issue: losing either of the LUNs/extents will cause the volume in the guest to do funny things, but the errors will be reported in different ways (in the guest for option 1, in vSphere for option 2).
The ideal solution would be to combine the two datastores at the array level, and then have this VM sit on its own, appropriately sized, datastore.