I have another question: The maximum aggregate size on most FAS platforms is now 800TiB. Does that include the Parity, and the Flash Pool? In other words, do I need to make the size less than 800TiB to accommode parity discs, and the flash pool of a hybrid aggregate?
#Maximum Aggregate Size
cyan bluff
jade kelp
Max aggr size / right sized capacity per disk = Max number of netto disks in the Aggregate... As in Party, flash or fabric Pool ist extra
wise bronze
To my knowledge it is usable capacity, so 800 without parity and cache
jade kelp
To sum it up: 800TB is quite a lot
forest marsh
Parity drives haven't been factored into maximum aggr size since ONTAP 7.3 was released
Prior to that, everything counted
jade kelp
yeah, I still remember how much more space I got after upgrading to 7.3... also, kahuna improvements got >+25% performance
cyan bluff
Yes, we have a lot of data, and it is growing by leaps and bounds
jade kelp
you can have several aggregates per node and 12 or 24 nodes per cluster.... shouldn't that be enough? How many petabytes are you planning?
wise bronze
If it is a file workload, consider StorageGrid if S3 is an option
Or if your active set is not that large, a C-series with FabricPool
dusk nexus
ONTAP fas platforms have their own maximums for cache (flash cache + flash pool)
As far as I know they do not count against the size the aggregate, but I admit I may be wrong on that.
dusk nexus
When using fabric pool, the volume size can be many PB.
The max sizes are documented and the criteria for the size is that the performance tier must be a percentage of the capacity in order to minimally hold the meta data.
sharp hearth
For information on the new capasity AFF C80 model you can have aggregate with a size of 1600TiB.
An AFF C80 with 48x60TiB internal disks gives you about 2PiB on a 4U unit. Pretty nice 🙂