Flexgroup aggregate best practice

[STEVEWILLSQ]

As per Tr-4571 best practice for a FAS based flexgroup is 2 aggregates per node.

What would be the conequences of adding a 3rd aggregate per node.

Environment

6 node FAS8200 cluster 3x DS460C shelves per pair loaded with 10TB drives

ONTAP 9.7P5

2 data aggregates per node ~288TB per aggr

Aggregates are 42x 10TB drives 12 drive RAID-TEC raid groups.

4 member volumes per aggr, ~68TB

so far so comforming to the Tr

Single flexgroup spans the cluster and is approaching 95% full

Customer has additional shelves available and would like to expand the flexgroup.

Adding an aggregate breaks the 2 per node best practice rule

Adding a raid group breaks the 400TB aggr size limit

So without adding nodes what would the implications be of adding a 3rd aggregate per node.  I'm loath to break best practice

[parisi · Accepted solution]

No real consequence. The best practice is based on the number of aggregate affinities available in ONTAP; there are only two aggregate affinities per node. Adding a third aggregate doesn't add more affinities so no perf gain. You do get capacity gains however.

[STEVEWILLSQ]

great to know,  thanks for the responding so quickly!