"zpool status" interpretation

[TNosredna]

Hi, I setup FreeNAS a while ago and am wanting to upgrade 2 of my drives to two new 2TB drives. I'm trying to determine which two to upgrade that would give me the most bang for my buck. Though I'm starting to think that I will need to destroy and recreate the zpool altogether.

As I don't remember the details of creating the zpool, I'd like to understand what it is I have. I haven't seen any examples online of what I'm looking at, so I'm sure it's a newbie mistake that got me where I am. But, I thought I was just following the prompts of the gui during creation.
Can someone explain what I'm looking at below as well as how I might have gotten that via the FreeNAS gui?

Code:

barn# zpool iostat
               capacity     operations    bandwidth
pool        alloc   free   read  write   read  write
----------  -----  -----  -----  -----  -----  -----
Barn        3.39T   707G     55     18  4.99M   621K
barn# zpool iostat -v Barn
                                           capacity     operations    bandwidth
pool                                    alloc   free   read  write   read  write
--------------------------------------  -----  -----  -----  -----  -----  -----
Barn                                    3.39T   707G     55     18  4.99M   621K
  gptid/d919f2a0-1400-11e3-abde-00219b681af5  1.22T   145G     18     10  1.79M   349K
  gptid/d98da02a-1400-11e3-abde-00219b681af5  1.27T   553G     20      7  1.88M   266K
  mirror                                 919G  9.28G     16      0  1.32M  5.67K
    gptid/da57e68d-1400-11e3-abde-00219b681af5      -      -     12      1  1.28M  8.84K
    gptid/dace52f3-1400-11e3-abde-00219b681af5      -      -     13      1  1.28M  8.84K
--------------------------------------  -----  -----  -----  -----  -----  -----

Code:

 barn# zpool status        
  pool: Barn
 state: ONLINE
  scan: scrub repaired 0 in 9h8m with 0 errors on Wed Feb 12 06:49:41 2014
config:

	NAME                                            STATE     READ WRITE CKSUM
	Barn                                            ONLINE       0     0     0
	  gptid/d919f2a0-1400-11e3-abde-00219b681af5    ONLINE       0     0     0
	  gptid/d98da02a-1400-11e3-abde-00219b681af5    ONLINE       0     0     0
	  mirror-2                                      ONLINE       0     0     0
	    gptid/da57e68d-1400-11e3-abde-00219b681af5  ONLINE       0     0     0
	    gptid/dace52f3-1400-11e3-abde-00219b681af5  ONLINE       0     0     0

errors: No known data errors

Thanks!!

 

[HoneyBadger]

Just trying to sort this - it looks like you've got 2x1TB and 2x1.5TB drives there.

It looks like you made a stripe when you initially set it up with the 2x1.5TB, then made a 1TB mirror vdev out of your other ones and added that to the parent stripe.

If I'm correct about that, you should definitely destroy and recreate because that pool has horrible failure tolerance - only one of your 1TB drives can fail. If either 1.5TB fails, or both 1TBs, your pool is gone.

What's the rest of your specs - CPU, motherboard, RAM, SATA controller?

And if you can't tell by my sig, if you're not running ECC, Honey Badger's gonna care.

 

[TNosredna]

That was kinda what I was afraid I'd done. Bummers! Thanks for the clarification!!
What is the best command to run to get the display for the hardware you are looking for?

And no, I'm not running ECC. I know it's a sin, but this is the "hand I was dealt", so I need to figure out the best I can do with this hardware. Fyi, I'm backing up my data (using FreeNAS CrashPlan plugin) to a USB drive that I've attached to the NAS box (offsite backup is not a high priority for me at the moment).

Though, if we can put that particular aside, I'd love to hear what type of pool you'd recommend I create, assuming I have the following drives available:
2 x 1TB
1 x 1.5TB
3 x 2TB

Thanks again!!

 

[cyberjock]

You do realize that your backups will be trashed if your non-ECC RAM fails... so don't think that your backups are going to save you. You'll back up a trashed pool to Crashplan or whatever else you choose to use long before you realize the error of your ways. That's precisely why ECC RAM is so heavily recommended.

 

[HoneyBadger]

I'm not running ECC.

ಠ_ಠ

I know it's a sin, but this is the "hand I was dealt", so I need to figure out the best I can do with this hardware. Fyi, I'm backing up my data (using FreeNAS CrashPlan plugin) to a USB drive that I've attached to the NAS box (offsite backup is not a high priority for me at the moment).

Mulligan that hand. If you're running ZFS, you need ECC. Period, full-stop, end of discussion. I've been registered here for two weeks as of today and I've already seen three people's pools die from bad RAM.

Though, if we can put that particular aside, I'd love to hear what type of pool you'd recommend I create, assuming I have the following drives available:
2 x 1TB
1 x 1.5TB
3 x 2TB

Thanks again!!

My suggestions:

1. Buy ECC RAM.
2. No, seriously, buy ECC RAM.
3. Sell or repurpose the 1.5TB drive, it's the odd man out.
4. Buy another 2TB and make a 4x2TB RAIDZ2 (4TB usable) for your primary pool.
5. Create a second mirrored pool of 2x1TB to replicate the really critical data from the first pool to.

 

[TNosredna]

awesome! thanks for the info and the feedback! I have another question about the next steps, but it's a different topic, so I'll start another thread!