Rebuilding Aggr0 and Vol0
2014-04-14 06:28 AM
I did something incredibly stupid the other day, as soon as I did I knew it was too late! I have a FAS2220 that had 24 disks split 12 and 12, dual controller. I also had a empty shelf, I installed an additional 12 disks for more storage and decided to also split these so I would have 18 and 18. I installed the drives upgraded the firmware and then assigned the disks to each controller. I then went to the AGGR0 on both controllers and assigned the disks. I looked at the storage size and noticed it wasn't what I had expected I went back in to look at the AGGR0 and noticed my Raid group size was 14. So now I have a raid group with 14 disks and a raid group with 3 disks (1 spare, RAID DP). Not only is this incredibly wasteful but now I fear that I may have performance issues with this small Raid group as it really only has 1 usable disk.
My plan was to just VM Motion my VMs to my secondary controller delete the AGGR0 and rebuild it. However I wasn't thinking about Vol0 which from my understanding contains all of the configuration. I don't see any way to export the configuration and everything I am seeing online suggest moving the Vol0 to another Aggr which I also do not have since all of my disks are attached to this aggr0, unless I can move it to the secondary controller while I make the change?
So my question is given my problem, what is the easiest way to fix this. I will have to unfortunately do this for both controllers as I was an idiot and moved both sets of disks before I realized what I had done. I was planing on doing 1 at a time so I can at least keep my environment online during the changes. I also do have a third seperate single controller however there is not enough storage on that device to hold all of my volumes. I planned on using that to hold some of my storage though during this change.
Thanks in advance,
1 REPLY 1
Re: Rebuilding Aggr0 and Vol0
2014-04-15 02:59 PM
- evacuating one controller
- ndmpcopy'ing the /etc directory off to the partner or the third controller,
- rebuilding the aggregate
- boot options 4... (might take a while)
- Setting the correct RGSize and add remaining disks
- moving (ndmpcopy'ing) the old /etc back to an empty volume in the new aggregate, let's say 'volroot'
- vol options volroot root
- recreate the other volume(s) for your datastore(s)
and do the same with the other controller.
The only space you'll need is room for the content of /etc (and all subdirectories...)
You will loose your old snapshots, though!
Hope that helps...