MidasMagoo
Cadet
- Joined
- May 24, 2019
- Messages
- 4
Say I serve up three Windows SMB shares to my team, let's call them child1, child2 and child3. They exist as child datasets to parent, i.e.:
vol1/parent/child1
vol1/parent/child2
vol1/parent/child3
I want the three datasets to share a 20TB pool, not including snapshots. Snapshots are taken at the child level--no parent level snapshots. This is because the different children have different retention rules on the snapshots. I want to use refquota on the children because I need the team to be able to recover space on their own (by identifying files to archive to a different server and removing them from the primary server). I'll manage the total space and snapshots outside of the 20TB presented to them. Here's the problem from my experimentation so far:
1) If I quota parent and refquota each child, then the child snapshots still take up space under the parent quota, so the team doesn't really recover usable space.
2) If I refquota parent and refquota each child, then I can't figure out how to get the three children to share 20TB. For that matter, I suppose I wouldn't need to do anything with parent, just refquota each child. Same problem, though. The children aren't sharing the 20TB. I'd have to decide how to divide the 20TB (which I don't want to do).
3) If I refquota the parent and put no quota/refquota on the children, then the children have access to the whole of the pool. Refquota on the parent seems to not only ignore snapshots, but ignore the child datasets.
I had thought of going with option 1 and putting the quota on the parent as 25TB or 30TB, whatever I thought I might need for snapshots. I still can't figure out how to share the 20TB on the three children, though. If each child reports space available until the parent quota is reached, I end up having the same problem of the team being unable to free up space on their own. If it's not obvious, archiving happens only when sufficient primary space is gone. Nature abhors a vacuum.
Any thoughts on how to achieve what I'm looking for?
vol1/parent/child1
vol1/parent/child2
vol1/parent/child3
I want the three datasets to share a 20TB pool, not including snapshots. Snapshots are taken at the child level--no parent level snapshots. This is because the different children have different retention rules on the snapshots. I want to use refquota on the children because I need the team to be able to recover space on their own (by identifying files to archive to a different server and removing them from the primary server). I'll manage the total space and snapshots outside of the 20TB presented to them. Here's the problem from my experimentation so far:
1) If I quota parent and refquota each child, then the child snapshots still take up space under the parent quota, so the team doesn't really recover usable space.
2) If I refquota parent and refquota each child, then I can't figure out how to get the three children to share 20TB. For that matter, I suppose I wouldn't need to do anything with parent, just refquota each child. Same problem, though. The children aren't sharing the 20TB. I'd have to decide how to divide the 20TB (which I don't want to do).
3) If I refquota the parent and put no quota/refquota on the children, then the children have access to the whole of the pool. Refquota on the parent seems to not only ignore snapshots, but ignore the child datasets.
I had thought of going with option 1 and putting the quota on the parent as 25TB or 30TB, whatever I thought I might need for snapshots. I still can't figure out how to share the 20TB on the three children, though. If each child reports space available until the parent quota is reached, I end up having the same problem of the team being unable to free up space on their own. If it's not obvious, archiving happens only when sufficient primary space is gone. Nature abhors a vacuum.
Any thoughts on how to achieve what I'm looking for?