2011-09-11 11:09 PM
I am new to Netapp storages and want to know how to calculate usable size of a aggregate.
My storage is Netapp FAS3210 (HA) with 300GB 15K SAS 24 disks. I will create 9+2 aggregate for each controller.
Solved! SEE THE SOLUTION
2011-09-11 11:23 PM
2011-09-12 12:11 AM
So as a quick guide, (the KB articles will give you a better idea)
For each aggregate configured in the way you are talking…
Things to consider…
Raw space then is 11x300Gb Drives – lets say 3.3Tb give or a take…
Things to consider
1. Disk right sizing – those 300Gb drives are not 300Gb they are actually 272Gb once they are right sized (leaving 2992gb)
2. 2 parity drives will then remove another 544gb (2448gb)
3. Wafl and OS overhead, lets take around another 10% for that, so about 250Gb (2198Gb)
So around 2.1tb is your usable space…
You then have to consider things like snapshot reserve etc so if you applied to each volume a “standard” (as snap reserve needs to be set to whatever is appropriate for your retention requirements) reserve of say 20% that would remove around 440gb so we are down to 1558gb or data space that can be presented to users and apps….
A really good guide as a starting point, especially in smaller environments is if you half the raw capacity, that will be about the usable space… if you look at the above, that pretty much applies here…
Don’t forget to have each controller with a list one spare disk, per disk type…so if you have SAS and SATA drives for example on a controller there should be at least one of each type as a spare…
Although this may sound a lot of “lost” disk…you have to look at what you are getting…one you start to use snapshots, thin provisioning, clones etc….you’ll soon find yourself over provisioning! And having much more presented capacity than you have real disk….
Check the KB articles as suggested and you’ll find more detail on the above…but those rough guidelines should stand you in good stead.
2011-09-12 12:13 AM
As RAID-DP is effectively a stripe of data drives plus two parity drives, the simpliest way to work out useable space is to multiply the number of data drives with the useable space of the data drives.
The KB articles above show how to work out how much space you will get out of each drive, however once you have been working with NetApp for awhile I recommend creating an internal KB with realworld disk sizes.
For example, In our case we are seeing ~227GB of useable space from a 300GB Drive.
Therefore in your example of a 9+2 Aggregate you will see about 1.99TB of useabale space per Aggregate (calculation was 9 x 227GB / 1024).
Note, this assumes you have no fractional reserve set on the Aggregate (snap reserve -A <aggr name> 0). Ensure that if you are setting fractional reserve to 0% there is no schedule either (snap sched -A <aggr name> 0 0 0).