How to wipe all data and rebuild new Raid array?

[Berkyjay]

OS: FreeNAS-11.2-U3
Storage: 1 pool with (3x 4TB raidz1) and (3x 6TB raidz1)

I'm reaching my storage capacity and I want to take advantage of the two extra drive slots I have in my rig to add two new 6TB disks. But I learned from another thread that there's no way for me to just add these two new disks to my existing 6TB raid array. I have all my data backed up offline and now I want to wipe all the disks and make new raid arrays. But I'm going through the UI and I can't clearly see how I would go about doing this.

So I was hoping someone could point me to a guide that walks you through this process.

Thanks in advance.

 

[Redcoat]

See 9.2.1. Creating Pools in the docs. As part of that process you will see that any disks with data will be wiped:

"After the desired layout is configured, click CREATE. A pop-up warning serves as a reminder that all disk contents will be erased. Click Confirm, then CREATE POOL to create the pool. "

9.2.1 is your walkthrough...

 

[Berkyjay]

See 9.2.1. Creating Pools in the docs. As part of that process you will see that any disks with data will be wiped:

"After the desired layout is configured, click CREATE. A pop-up warning serves as a reminder that all disk contents will be erased. Click Confirm, then CREATE POOL to create the pool. "

9.2.1 is your walkthrough...

I actually JUST found this in the docs for 11.2. Thanks for the response though!

 

[Heracles]

Hey Berkyjay,

Because your drives are not the same size, to put them all in a single RaidZ-x device will restrict your bigger drives to the size of the smallest. As such, your 6 TB drives will be used as 4 TB ones if joined in a single vDev.

Once up to 6 drives, Raid-Z2 is probably a better choice compared to RaidZ-1. RaidZ-1 is pretty high risk. Once a drive fails completely, the system may not recover from even a small read error. Because there will be a ton of reads during re-silvering process, the vDev may not rebuild properly.

Should you wish to keep the RaidZ-1 structure and not waste space, you can just create a new RaidZ-1 vDev with your new drives and add it to the pool. It is good practice to have all vDevs in a pool on the same structure like both vDev being 3 drives wide RaidZ-1. To have all drives the same size is not as important. Good news is that to do that, you do not need to empty your actual pool. Just create the second vDev and add it to the pool. To empty the pool and re-create it will re-distribute the load over all the drives in proportion of their capacity, but ZFS would have do it over time slowly as the content of the pool changes.

Good luck in your migration,

 

[Berkyjay]

Hey Berkyjay,

Because your drives are not the same size, to put them all in a single RaidZ-x device will restrict your bigger drives to the size of the smallest. As such, your 6 TB drives will be used as 4 TB ones if joined in a single vDev.

Once up to 6 drives, Raid-Z2 is probably a better choice compared to RaidZ-1. RaidZ-1 is pretty high risk. Once a drive fails completely, the system may not recover from even a small read error. Because there will be a ton of reads during re-silvering process, the vDev may not rebuild properly.

Should you wish to keep the RaidZ-1 structure and not waste space, you can just create a new RaidZ-1 vDev with your new drives and add it to the pool. It is good practice to have all vDevs in a pool on the same structure like both vDev being 3 drives wide RaidZ-1. To have all drives the same size is not as important. Good news is that to do that, you do not need to empty your actual pool. Just create the second vDev and add it to the pool. To empty the pool and re-create it will re-distribute the load over all the drives in proportion of their capacity, but ZFS would have do it over time slowly as the content of the pool changes.

Good luck in your migration,

Hi thanks for the tips! I wasn't planning on mixing my drives in the same raid array. I currently have two vdevs and plan on having two in the new pool. As for having RaidZ1, I'm not too worried about the risk factor. My server is strictly media and so isn't stressed with a lot of read and writes. Plus I always keep a full offline backup. :)