7 Posts
0
1892
Migrate all VNX File storage to a new set of disks and reclaim the old ones
We've had a change of customer performance requirements on our VNX5600 that means we want to move away from our current (RAID 1/0 NL-SAS) File Storage Pool to a new one (RAID 1/0 SAS) and release the old disks. I believe that the Data Mover LUNs (~filestorage HLID 0-4) are on our NL-SAS pool, and so they will need moving, but there's no obvious way to do that. My procedure is currently looking like:
- Create target Storage Pool
- Create 10x equally-sized LUNs to fill the new Pool (as per best practice)
- Allocate those LUNs to File Storage/~filestorage (will the DMs handle renaming these or should I do it?)
- Create new File Storage Pool and get it to use those LUNs
- Migrate the File Systems (not sure what the best option is here)
- # somehow move the root_disks to the new pool or the vault RG
- Remove any volumes from the old Pool
- Remove the LUNs from ~filestorage and delete
- Destroy the Storage Pool
How do I do step 6 and are there other things do I need to do to clear the old pool out?
Thanks for any guidance!
DELL-Josh Cr
Moderator
Moderator
•
8.6K Posts
0
March 8th, 2023 12:00
I don't see a way to move the root drives, all of the user data could be moved to the faster drives, but the current drives would need to stay as root.
DELL-Josh Cr
Moderator
Moderator
•
8.6K Posts
0
March 9th, 2023 07:00
Since you are NFS only yes that should work, EMCopy is only for CIFS. There may be something in our solve tool that may be helpful. https://dell.to/3kY11Mw;
DELL-Sam L
Moderator
Moderator
•
7K Posts
0
March 8th, 2023 08:00
Hello A-G,
Do you have a DAE connected to your vnx5600? Are the new drives already installed in your vnx5600? What is your current OE?
A-G
7 Posts
0
March 8th, 2023 09:00
Hi Sam, thanks for replying.
The drives are already in place and currently unallocated. We are running 05.33.021.5.322 / 8.1.21-303. The current layout is 4x 3 TB NL-SAS in RAID-1/0, and we are moving to 10x 600 GB SAS in RAID-1/0. It's a step down in capacity, but we need the performance.
DELL-Josh Cr
Moderator
Moderator
•
8.6K Posts
0
March 8th, 2023 10:00
Do you have Virtual data movers configured? https://dell.to/3L8rdP7 that should make it easier.
A-G
7 Posts
0
March 8th, 2023 10:00
Hi Josh, no VDM: we're NFS only and don't have many shares so didn't feel that we needed them. We are using cross-site replication using VNX Replicator.
A-G
7 Posts
0
March 9th, 2023 06:00
Hmm, that's a shame! If only the installer had used the vault RG instead. Oh well, we can get by with this pool continuing to exist.
Going back to the original question, is nas_copy the best way to migrate the file systems, or something else? I can remove the cross-site replication and don't mind if there's an associated outage.
Thanks ~ A-G
A-G
7 Posts
0
March 9th, 2023 07:00
Thanks for your help.
Now I just need to write the change document...
~ A-G
DELL-Josh Cr
Moderator
Moderator
•
8.6K Posts
0
March 9th, 2023 07:00
Let us know if you have any further questions.
A-G
7 Posts
0
March 9th, 2023 10:00
Thanks for the offer! Further question: once I've migrated all of the FS away from my old File Storage Pool, and deleted all of the visible LUNs that were supporting the file systems (leaving only the hidden Data Mover LUNs), am I okay to add more NL-SAS disks to the underlying Storage Pool? (It's RAID-1/0 2+2, and I want to add another 1+1 pair. Performance isn't a consideration beyond supporting the Data Movers)
DELL-Josh Cr
Moderator
Moderator
•
8.6K Posts
0
March 9th, 2023 10:00
Yeah, shouldn't be a problem.
dynamox
1 Rookie
1 Rookie
•
20.4K Posts
0
March 10th, 2023 11:00
fs_copy works great for file system migrations.
A-G
7 Posts
0
March 13th, 2023 02:00
Thanks! I'll look into that. I was thinking of using a cascaded Replication at the remote site to populate to the new pool, and a one to many Replication on the live site to move to the new pool (all with a manual, refreshed checkpoint) and then redirect the bits in turn to keep things in sync. That would be quite a hassle to keep everything in order, though!