Is there any way to determine storage used on each vdev?

[maglin]

Maybe the reason you didn't see a decrease on the original vdev with the move is because a snapshot is keeping that data in tact. The scary thing of wanting to balance out the data is you have to delete the snapshots so those blocks can be freeded up. If the data somehow gets lost due to user error it's gone and you have to restore from backup.

And if you do move the data around I believe zfs send | zfs recieve would be faster than rsync, mv, or cp but requires a snapshot to use. I may be wrong as I've never moved a lot of data with anything other than zfs send or just over the network.

 

[philhu]

I agree on all your points. Moving data around just causes snapshots to hold it. Jgreco explained about multi-vdevs in a volumes. So I will just leave it

I just hate to have one vdev approaching 88% while the other is at 21%