VMware Solutions Discussions
VMware Solutions Discussions
This FAS2040 was originally setup for a VDI deployment and all disks were lumped into one aggregate and the second controller as a passive spare. The VDI project never happened and they are currently using the FAS2040 as storage for Server VMs.How can I break the aggregate apart? I would like to use the FAS2040 in Active/Active with splitting the disks 50/50. Currently the FAS2040 has 12 drives with a DS4243 with 24 drives (all 15k SAS 500Gb). Each controllers root aggregate has 2 disks in RAID4, the large data aggregate has 29 disks in RAID-DP, and then the last two disks are spares. Currently the VMWare hosts are being served over NFS shares from the FAS2040.
I know aggregates cannot be shrunk. This will a disruptive rebuild of the FAS2040. The main thing I need is to be able to preserve the servers that are already hosted on the FAS2040. I do not have another NetApp to SnapMirror to. Can I just vMotion the guest VMs to local storage, rebuild the aggregates, then vMotion the guest VMs back to NFS shares?
Solved! See The Solution
You probably have no choice except to get all the data off using Storage vMotion as you have described.
There is one slightly crazy option that might work if you haven't used much of the space on the big aggregate. You could disable the second controller and using the remaining disks to extend the RAID-4 volume being used for the root aggregate. You'd end up with a 6-disk root aggregate and one hot spare, all on one controller. If these are 15K SAS drives then they'll be 600GB which should give you an aggregate in the ballpark of 2.5TB which you could snapmirror your stuff onto. This obviously assumes you can have downtime on your server VMs.
You probably have no choice except to get all the data off using Storage vMotion as you have described.
There is one slightly crazy option that might work if you haven't used much of the space on the big aggregate. You could disable the second controller and using the remaining disks to extend the RAID-4 volume being used for the root aggregate. You'd end up with a 6-disk root aggregate and one hot spare, all on one controller. If these are 15K SAS drives then they'll be 600GB which should give you an aggregate in the ballpark of 2.5TB which you could snapmirror your stuff onto. This obviously assumes you can have downtime on your server VMs.
And now the question, is it worth the trouble?