ONTAP Discussions

'disk replace' to create bigger raid-group size

coreywanless
3,802 Views

Hello All,

 

I'm wondering if there is someone has tried to do it, and has some solution.

 

We had an flash pool aggregate that contained 8 400gb SSD drives. We recently got in 800gb SSD drives. In an attempt to increase the flash pool size we did a 'disk replace' on all the SSD drives of the flashpool and replaced them with the 800gb SSD drives. We were expecting the raidgroup to be upgraded after the last disk swap, but it doesn't appear to do that.

 

Has anyone ever tried this with success or have some other workaround for it? 

 

Thanks,

Corey

1 ACCEPTED SOLUTION

bobshouseofcards
3,783 Views

Hi Corey -

 

You're running into a standard operation on aggregates.  When a disk is "replaced" you are really simulating a failure and a standard swap in of a spare drive.  This follows all the rules of the raidgroup rebuild including rapid rebuild via copy when available.  

 

For your situation the key thing is that a disk, once in a raidgroup, can only be replaced with another disk of the same logical size.  A physical disk of the same type but with larger capacity can be used when no other spares are available, but the larger disk will only ever be formatted for and used as if it were the same size as the smaller disk.  This applies to all disk types.

 

To get a larger flash pool, you'd have to replace the flash pool part of the aggregate.  Unfortunately, that means a complete aggregate rebuild from scratch, as you cannot remove the existing SSDs from a hybrid aggregate once created.  You'd need to migrate all volumes off the aggregate, then rebuild the aggregate with the 800GB SSDs to use that capacity, or just use them as 400's.  

 

Depending on platform and DoT version you might be able to use ADP on the 800's to split them up and not waste half of the disk, though I leave confirmation of whether you can do that with an existing non-ADP set of disks to others as I'm not as up to date on ADP at that level of detail.  Although with the 800's now added you are likely in a complete rebuild situation to fully use those 800's in any case.

 

 

Hope this helps

 

Bob Greenwald

Lead Storage Engineer | Consilio LLC

NCIE SAN Clustered, Data Protection

 

 

Kudos and accepted solutions are always appreciated.

View solution in original post

4 REPLIES 4

aborzenkov
3,788 Views
No, that's not possible. Depending on your environment, you may be able to non-disruptively relocate volumes to new larger aggregate.

aborzenkov
3,785 Views
Oh, sorry, did notice tag. So in cDOT the simplest way is to create new aggregate using new disks and move volumes to it.

bobshouseofcards
3,784 Views

Hi Corey -

 

You're running into a standard operation on aggregates.  When a disk is "replaced" you are really simulating a failure and a standard swap in of a spare drive.  This follows all the rules of the raidgroup rebuild including rapid rebuild via copy when available.  

 

For your situation the key thing is that a disk, once in a raidgroup, can only be replaced with another disk of the same logical size.  A physical disk of the same type but with larger capacity can be used when no other spares are available, but the larger disk will only ever be formatted for and used as if it were the same size as the smaller disk.  This applies to all disk types.

 

To get a larger flash pool, you'd have to replace the flash pool part of the aggregate.  Unfortunately, that means a complete aggregate rebuild from scratch, as you cannot remove the existing SSDs from a hybrid aggregate once created.  You'd need to migrate all volumes off the aggregate, then rebuild the aggregate with the 800GB SSDs to use that capacity, or just use them as 400's.  

 

Depending on platform and DoT version you might be able to use ADP on the 800's to split them up and not waste half of the disk, though I leave confirmation of whether you can do that with an existing non-ADP set of disks to others as I'm not as up to date on ADP at that level of detail.  Although with the 800's now added you are likely in a complete rebuild situation to fully use those 800's in any case.

 

 

Hope this helps

 

Bob Greenwald

Lead Storage Engineer | Consilio LLC

NCIE SAN Clustered, Data Protection

 

 

Kudos and accepted solutions are always appreciated.

coreywanless
3,774 Views

Thanks Guys. I figured that was the case, but needed someone to validate.   

We have an AFF install coming in that cluster. I will probably use that as the temporary location as we evacuate and rebuild the aggregate.

Public