To recreate one and only one aggr on same disks?
a month ago
For some reasons, we need to recreate the one and only one data aggregate on several nodes. Aggregates have either root-data or root-data1-data2 configurations. We could have a shelf of disks for staging purpose, moving data from that data aggr to this staging aggr, and then destroy / recreate that data aggr, and finally move the data back.
My question: Since root aggr is using root partitions on multipel disks, how could I restore all partition configurations back to that data aggregate when I recreate the aggr and before I move data back? or, I have to manually create partitions and also ownership?
Or, What is your recommendation?
5 REPLIES 5
I wasnt 100% sure. Thanks for clarifying.
In that case yeah, you could do the following.
Add a temp shelf
create a aggr_temp just using that shelf.
Copy your data to that shelf.
Delete the aggr1.
Reconfigure partitions how you want.
Create new Aggrs
copy data back to new aggr(s).
unown temp disks.
Hot-remove temp shelf.
>Reconfigure partitions how you want.
>Create new Aggrs
This is the part I have questions with. Remember I have only destroyed those partitions on the data aggr, partitions sharing the same disks are not initialized and still active in root aggr. How do I restore partitions were in data aggr before it got destroyed?
After the data aggr is deleted, the "disks" (partitions) will just be returned back to the spare pool. The root aggr will remain intact.
zero them out and they'll be ready to be added to a new aggr(s) If you need to reassign the disks/partitions betweent the two controllers you use "-data" flag with the disk sub commands.
disk removeowner -disk x.x.x -data true
disk assign -disk x.x.x -data true -owner NODEx
This command to show what is assigned where.
disk show -fields data-owner, root-owner
How do you plan to create the new aggr? and why is it needing to be deleted/recreated. I can only think of a few reasons to do that.