zfs/raidz: seems like I'm failing with math

[Antony Uspensky]

On Mon, 17 Oct 2016, Eugene M. Zheganin wrote:

> Thanks ! It does explain it. But then again, on a pool that has been just 
> created, I check the properties of the root dataset (I'm posting all the 
> properties, just to display there's no child datasets or data on the pool):
>
> ===Cut===
> # zfs get all gamestop
> NAME      PROPERTY              VALUE                      SOURCE
> gamestop  checksum              on                         default
> ===Cut===
>
> Only 4.03T is available. Looks like it's the actual size, since it's zfs and 
> not zpool. But 960197124096 bytes * 5 / 1024^4 gives me 4.366 Tb, and not the 
> 4.03 T. Where did about 300 gigs go ? I'm really trying to understand, not to 
> catch some questionable logic or find errors.

Checksums and zfs data structures - less then 8% of total space. Not too 
expensive. And do not confuse checksums (zfs level) and parity (zpool).
But this is estimated available space for empty pool. Real allocated space 
for filled up pool will be a bit less then 4.03T due to padding - read
https://www.delphix.com/blog/delphix-engineering/zfs-raidz-stripe-width-or-how-i-learned-stop-worrying-and-love-raidz
A.