Start a Conversation

Unsolved

This post is more than 5 years old

G

1254

November 8th, 2012 01:00

Unable to migrate to a new VNX because the old array raids have a slightly better usable capacity

Hello,

We are trying to migrate via Mirrorview froma CX-4 480 to a new VNX 5700. Unfortunately we found out that the new raid group haves ~2GB less of usable space. This is causing problems as the old raids are used 100% so we cannot migrate all the original luns mantaining the orignial layout.

I am talking of simple raid groups 4+1 or raid 10 with the same drive types (raw disk drive is equal on both the old and the new boxes)

Is there any workaround? Is this documented somewhere?

Regards

Guido

1.4K Posts

December 2nd, 2012 22:00

Hi Gas,

Apologies for the delay. Could you please, let us know if you are still facing the issue?

If no, Would you be kind enough to share your Resolution with the community?

If yes, What is the size of the source LUN which you wish to migrate?

Ankit

1.4K Posts

December 6th, 2012 09:00

What is the size of the source LUN which you wish to migrate?

19 Posts

December 6th, 2012 09:00

Not yet resolved.

19 Posts

December 7th, 2012 00:00

The same problem happens with different drive types, we have to migrate a lun that occupies a full raid group.
if on the target host (VNX) you create a raid group with the same drive type (same raw size in blocks reported from unisphere) the maximum available space is 2G smaller than on a CX.

Gas

247 Posts

December 10th, 2012 06:00

I fear there will never be a true solution: the limitation that you can only replicate to a same size LUN will never be lifted for obvious data integrity reasons.

As for the reason why the VNX has 2GB less usable space for the same amount of drives, I don't know.

Are you running VMware on those LUNs? If yes, I'd advice migrating using Storage VMotion instead as a workaround. Or use some sort of other

19 Posts

December 19th, 2012 06:00

Unfotunately it is not a VMware lun and i agree that probably this is more of an architectural change than a bug.

4.5K Posts

December 19th, 2012 14:00

Try this to work around the issue. Get the size of the original LUN in "Blocks" as listed in the LUN Properties, then create a new LUN and specify the size using the same Block number.

On the 5700, you are creating raid groups and not Pools correct? Pool LUNs are smalller by 2GB initial and a bit more depending on the size of the LUN.

What size disks are you using on the CX4 and on the VNX?

glen

No Events found!

Top