Hard disk PP's per disk is smaller than actual disk size 1

[jpn1]

I have a Pwr6 520 express server, with 6 drives: 2 146GB and 4 450GB drives.

I have installed a single VIOS on the 2 146GB drives and created 4 storage pools(VG's) in VIOS and assigned 1ea 450GB drive to ea storage pool.

I have done this in order to assign 4 virtual disks to my VIOC's (client partitions). 2 for rootvg and 2 for datavg. Then I mirror my VIOC rootvg and datavg at the VIOC level.

Since you are not supposed to LVM mirror on the VIOS if they are to be assigned as virtual disks to client lpars.

My question is how the VIOS allocated the space on the 450GB drives. I actually built the above config with the system planning tool and deployed.

Each of my 4 volume groups has 1 ea 450GB drive assigned with a PP size of 256MB and total PP's available is 1676 which is only 429GB or 31GB short the total drive space of 450GB. I would have thought there would be 1800 PP's at 256MB for a total of 450GB.

My 146GB drives also are short. They have 546 PP's available with PP size of 256MB. This is total space of 136.5GB or short 10GB.

Is this just drive space overhead that the LVM needs in order to make things work. I was a little sad I lost 30GB of space on my 450GB drives.

 

[DukeSSD]

Isn't about the same on any OS, you lose a whole lot on the raw size as soon as you format the disk?

 

[sjm2]

Storage devices use "decimal" GB and AIX is using "binary" GiB. You'll find that the drive is officially listed in the IBM Sales Manual as 146.8GB.

(146.8GBx1000x1000x1000)/(1024*1024*1024)=136.7GiB.

http://en.wikipedia.org/wiki/Gigabyte