Active IQ Unified Manager Discussions

Automatic volume provisioning with protection manger

detstorage
2,602 Views

Hi All,

I need to know how protection manager distributes auto provisioned volumes across Aggr’s

I have a provision pool with 20 Aggr’s each 20TB in size across 4 filers.

I’ve tested creating 6 relationships with OSSV 1TB,  its picking the 1st two Agrr’s from filer1 and filer2  in the pool.

We were told by the NetApp that the volumes would be evenly created across all Aggr’s in the pool.

Regards,

Mark

1 ACCEPTED SOLUTION

smoot
2,602 Views

Generally, ProtMgr will pick the aggregate with the most free space but it can get a little involved.

If I recall correctly, we first list all the aggregates in all the resource pools. We filter out any which are unacceptable (because of licensing, ONTAP versions, that sort of thing). Of the remainder, we sort by available space.

Here's where it gets a bit complicated. If there is more than one aggregate with roughly the same amount of free space, we use some tie breaking heuristics (where "roughly" means "within 25% of each other"). We try to reuse aggregates that a given dataset is already using and failing that, we try to pick an aggregate on the same controller that the dataset is already using.

I suspect in your case each aggregate is only filled to 5% (1 TB out of 20) so they are all considered roughly the same fullness. In this case, the "reuse an aggregate" and "reuse a controller" rules are probably kicking in. I can't say definitively and we'd need to turn on some extra logging to be certain what is going on.

View solution in original post

2 REPLIES 2

smoot
2,603 Views

Generally, ProtMgr will pick the aggregate with the most free space but it can get a little involved.

If I recall correctly, we first list all the aggregates in all the resource pools. We filter out any which are unacceptable (because of licensing, ONTAP versions, that sort of thing). Of the remainder, we sort by available space.

Here's where it gets a bit complicated. If there is more than one aggregate with roughly the same amount of free space, we use some tie breaking heuristics (where "roughly" means "within 25% of each other"). We try to reuse aggregates that a given dataset is already using and failing that, we try to pick an aggregate on the same controller that the dataset is already using.

I suspect in your case each aggregate is only filled to 5% (1 TB out of 20) so they are all considered roughly the same fullness. In this case, the "reuse an aggregate" and "reuse a controller" rules are probably kicking in. I can't say definitively and we'd need to turn on some extra logging to be certain what is going on.

detstorage
2,602 Views

Hi Smoot,

Thanks for the fast reply. I guess I need to create a large number of datasets and relationships to ascertain what's going on.

thanks again for you help.

Regards,

Mark

Public