So, we've inherited down some badly configured NetApp FAS2040's and I'm in the process of upgrading them to Data ONTAP 8.0.5 from 7.3.4 (7-mode). We're contractors so these things change hands quite a bit and just getting access to NetApp support was something of a task.
Anyways, I've been working on this issue since no one feels comfrtable at it and I've run into some snags. I want to maximize space since we have 4 of these 2040's (12 600g 15k SAS drives), a 2020, and a FAS2040 with the DS4xx something shelf unit. Working with one aggregate is fine as we don't think we'll need the added protection from having a root aggregate and aggregate for data volumes.
What is best way to setup these volumes and LUN's? I've looked at it from both angles of having 1 volume and 2-3 LUNs, or by having multiple volumes with one LUN each. I think the single volume works good enough and we could take advantage of deduplication more efficiently. However, I'm confused on something. If I make 2x 1.75 TB or 2x 1.5 TB LUNs does that give me enough space for snapshots?
I don't want to make snapshots, but I'm confused on how these things operate as I was under the impression that I should have Fractional Reserve @ 100% and snapshots at 0, but I was getting out of space alerts (everything is thick provisioned). I then proceeded to changing the FR to 0 (unchecking it in OnCommand) and the alert went away in vCenter plugin.
I say I don't want snapshots because I thought that taking advantage of SnapVault would be a great idea since we are running an old 4.1 vmware infrastructure, can't use VDP, and have no ways of buying a backup/dr software. I figured allocating a 2040 to backups with SnapVault or using the netapp plugin in vcenter to do scheduled backups would be decent enough for now.
So... I'm incredibly confused on how to design the volumes in respect to not running out of space with this fractional reserver and snapshots. Am I making my luns to large for my volumes? Is there some mathmatical formula I should use?