ONTAP Discussions

Head swap: consolidate 4 node FAS6250 cdot cluster to 1 FAS8200 /move aggregate to other filer

tim_vd
3,421 Views

Hi genies!

I have done a few 1 on 1 headswaps,but I have some questions for this setup. The NetApp documentation I have found focusses mostly on simple head swap procedures.

 

The title basically says it all: a 4 node cdot cluster, serving only nfs data, is being reinvested into a 2 node 8200 ha pair, while reusing the data disks.

4 node cluster is fas 6250, each node has it's own data aggregate and shelfs, which needs to be moved to the new HA pair.
The 8200 comes bundled with a shelf, so i can configure and join it in the cluster. But how do i move the data aggregates to the new filers?

Just reconnecting the old shelfs to the new 8200, and assigning the disks will probably not work. Any suggestions?

This will probably be a disruptive operation, so we are allowed to have downtime.

1 ACCEPTED SOLUTION

robinpeter
3,391 Views

If you are adding the new FAS8200 to existing cluster and later use the vol move.

you might have to figure out the additional space. (its not a headswap procedure..)

 

If you choose that route... My be you will be able to move some volumes and free up one existing aggr, which you can offline

and destroy, then disconnect the shelf from FAS6250 after removing the disk ownership. That way you can move loop by loop from all 4 nodes to FAS8200.

 

Other options will be doing an ARL headswap on node-1 and node-2 using FAS8200, so you dont have to do the vol move on those nodes.

and later based on the free space on node-1 and node-2 aggr, use vol move and free up and disconnect the shelf.. which you can reconnect to FAS8200.

 

The above two methord I suggest don't require any downtime.

 

If you are looking to avoide using the additional space and vol move.. disk-re-assignemnt should work, but you might need to plan for down time. And make sure all the shelf you planning to connect to FAS8200 have the supporting IOM modules. I'll strongly recommend you updating all dIsk and Shelf firmware to the latest before doing any of these procedure.

 

Robin.

View solution in original post

2 REPLIES 2

robinpeter
3,392 Views

If you are adding the new FAS8200 to existing cluster and later use the vol move.

you might have to figure out the additional space. (its not a headswap procedure..)

 

If you choose that route... My be you will be able to move some volumes and free up one existing aggr, which you can offline

and destroy, then disconnect the shelf from FAS6250 after removing the disk ownership. That way you can move loop by loop from all 4 nodes to FAS8200.

 

Other options will be doing an ARL headswap on node-1 and node-2 using FAS8200, so you dont have to do the vol move on those nodes.

and later based on the free space on node-1 and node-2 aggr, use vol move and free up and disconnect the shelf.. which you can reconnect to FAS8200.

 

The above two methord I suggest don't require any downtime.

 

If you are looking to avoide using the additional space and vol move.. disk-re-assignemnt should work, but you might need to plan for down time. And make sure all the shelf you planning to connect to FAS8200 have the supporting IOM modules. I'll strongly recommend you updating all dIsk and Shelf firmware to the latest before doing any of these procedure.

 

Robin.

tim_vd
3,318 Views

Robin,

Thanks for you answer! Indeed, when i use the ARL method to headswap 1 ha pair, I can use the new spindles to vol move the old volumes.

Off course, i will have to upgrade everything to Ontap 9.1 first...

 

Public