topic Re: Space calulations when creating new Aggregate on AFF A250 in AFF

Space calulations when creating new Aggregate on AFF A250

MK429 — Thu, 25 Jul 2024 10:06:20 GMT

I have a new AFF A250 running ONTAP 9.13.1P10 it has 24 x 7.68TB NVME SSDs

When I go to create the new aggregates, the system proposes :

Two aggregates each of 23 x 3.48TB = 65.8TB ( 2 disks lost to raid so 21 x 3.48 = 73.08 )

65.80TB is 10% less than 73.08TB - Typically this would be the 10% WAFL overhead, however as I am running out of the box ONTAP 9.13.1 I would expect the WAFL overhead to be 5% and the usable aggregate to be 69.42TB

My question is : when a new aggregate is created does it a by default have a 5% Aggregate Snap Reserve factored in, therefore I am seeing 5% WAFL overhead + 5 SnapReserve or do I have an issue with the AFF where it is not setting the WAFL reserve to 5%

It may seem a small amount of space for that aggregate example but over the whole cluster its around 40-50 TB we are missing. ( other nodes on the cluster are C400 which much larger disks )

Thank you.

Re: Space calulations when creating new Aggregate on AFF A250

TMACMD — Thu, 25 Jul 2024 20:05:27 GMT

Building out an A250 with 24 x 7.68, you whould end up with:

root/vol0 = maxraid-10, 8 root partitions, 2 root-partitions for Parity = 23.4G x x = 168.3G (default min-root)

data/aggr = maxraid-22, 20 data paritions, 2 data partitions for parity = 3.48T x 20 = 66.15T (2 spare data partitions)

-- OR --

data/aggr = maxraid-23, 21 data paritions, 2 data partitions for parity = 3.48T x 21 = 69.46T (1 spare data partition)

If you are deviating from these numbers, maybe the system was not initialized with all 24 disks in the system. I see that all to often. Using the built-in process to auto-create the aggregates, it will usually keep one partition by default as a spare.

Re: Space calulations when creating new Aggregate on AFF A250

MK429 — Sun, 28 Jul 2024 13:16:05 GMT

Traced the problem to the way the system manager UI calculates aggregate capacity, the UI factors in both the 5% WAFL overhead and 5% agg snap reserve when estimating the size of the agg.

However when you actually create the aggregate it doesn't apply the 5% snap reserve setting , thus giving you an extra 5% of space ( which matches what NetApp Fusion quoted ) very frustrating especially when you have large arrays of 15TB disk and suddenly 5% is a lot of space!