Stripe to a Z1 or Z2 Pool

[holmes.j]

NAS drives are better (with the possible exception of WD Red). Make absolutely sure you are buying CMR and not SMR drives. SMR and ZFS do not mix. Do not buy WD Green or similar - they aren't really designed for this.

Oh dear, all my drives are WD Red. Are they that bad?

So it sounds like I'll just use my onboard 6 SATA ports with 8gb disks, and make sure I don't connect to the Marvel controller. Swap to a bigger case. This should last for a few years and then I can add to the pool with a new vdev.

 

[NugentS]

WD Red were good, until WD snuck some SMR drives into the mix and called them NAS drives.
In general EFAX (as part of the model) = bad - but its not 100% consistent - so check carefully

 

[pschatz100]

WD Red's are fine as long as you have CMR versions. Check the model numbers to be certain.

By far, your simplest option is to buy a new computer case that can hold more drives. This would probably also help with cooling. A 6x8Tb Z2 pool will double your capacity and provide decent redundancy. Good idea. Backing up to an external 16 Tb drive is also a good idea. Remember: disk redundancy is not a replacement for proper backup.

I would get rid of the usb sticks and use SATA ports on the Marvell chipset for an SSD boot drive. It's true that Marvell chipsets have questionable reliability, but in your case I think it would be fine.

 

[holmes.j]

I'm back! Took me a while to get things done, this isn't really a top priority, but I've run into some problems.

I ended up with a 6x8TB in a Z2 array. I successfully moved my iocage and jails datasets over to the new pool. But I'm running into problems with my media dataset (13TB).

There were bad sectors on the old array in that dataset. When I used zfs send/receive it would fail consistently at something like 6.34TB in. So I said screw it and just used rsync to copy the data over. That seemed to have worked, until I rebooted. Now my dataset is empty, but the pool is claiming to have used the 13TB.

So my question is, where is the data? How do I get it back to the dataset and make it visible? Looking at the screenshot, goofy is the new pool, main is the old. goofy claims to have used 13.4TB but none of the datasets show it.