Drive Storage Space

[Baconmanic]

This maybe common knowledge. Now I know that drives, once formatted lose some space, but I feel I am losing alot of space.. Maybe this is normal.

I create a stand alone drive, following the how too, but the full size of the disk never shows up correctly. It only ever shows 300. This is a 320 dive btw.

So by the time I go through the walk through. Wipe the drive (zero out) mount and the CIFS share the grand total of storage space is around the 280 mark. It looks as if I am losing almost 40 gig of space! Now either this is completely normal and I just never noticed it, or I am missing something completely.

I have been banging my head an asking other techies I know, but no one seems to have an answer and all feel this is alot of "lost" space.

Ideas?

 

[titan_rw]

A 320 "gig" drive is typically 320,000,000,000 bytes. Translate that into "gigabytes" that 99% of the world uses, and you get: 320000000000 / ( 1024 ^ 3 ) = 298 gig.

The rest of the difference is reserved space and overhead of zfs. Everything is normal.

 

[gpsguy]

It's normal. Search the forum for the answer. Even in Windows 7, I only see 465Gb on a 500Gb drive.

 

[Baconmanic]

Thanks for the answers. I figured it was normal.
Titan_rw I am running UFS currently as they are older drive and my system only has 2 gig ram.

Would the UFS overhead be different? Then if I RAID it would it eat more?

 

[titan_rw]

UFS reserves a percentage of the filesystem for root only use. This can be adjusted to 0 however. I'm not too familiar with ufs, so I don't know the pitfalls of doing this. UFS does have less overhead from a memory perspective too. With only 2 gigs of ram, ufs is really your only 'safe' option.

What do you mean by 'raid' it? With ufs, you can use a hardware raid controller to do whatever you want. Mirror / stripe / raid 5/6, etc. Overhead from ufs's perspective doesn't change. There's the amount of space used by whatever parity level you choose, but that's overhead used by the raid card, and not freenas.

The only software raid supported by ufs is mirroring and striping. Neither should have any affect on the overhead of ufs. Of course, available space will change with the raid level chosen, as with zfs. That's raid redundancy (or lack thereof), and not filesystem overhead.

 

[Baconmanic]

I think I get what you are saying now. You must forgive me this is my first time jumping in an building a NAS unit.

 

[Baconmanic]

titan_rw - After thinking about it a little more and doing some more research I am thinking I want to try to do this in a ZFS. I understand that there is a recommended amount of RAM at 8 gig, but the minimum is 4 gig (in the official documentation, but contradictory to the 'real world' applications I have read about) . You stated that UFS is my only 'safe' option. What do you mean by that? Is this just based on the recommendations?

With such a small disk size 278 gig (mirrored between a 320 and 300 drive), would I really see that much of an issue with only 2 gig?

I really am only setting this up as a storage for home pictures and movies and want to make sure I have as much redundancy as possible. (NAS with mirrored disks > then upload to a cloud based online storage. [Single disk off to side as a "gold master" kind of deal]) {neurotic much? yes..}

 

[titan_rw]

'safe' as in 2 gigs of ram and zfs can cause kernel panics, etc. Which can result in data corruption worst case scenario.

You'd probably be ok with 4 gigs an a 300 gig mirror zpool.

What do you mean by 'single disk off to the side'? Mirrors are not designed to be split apart, and stored separately. They're designed to run concurrently providing 'live' redundancy.

 

[Baconmanic]

By single disk I mean just a stand alone drive that I can back up via USB

 

[titan_rw]

Yes you can do that. You'll have to manually mount the usb disk each time you plug it in. And you'll have to manually update whats on the usb disk after it's mounted. rsync will do that.