Replacing a "healthy" drive to match all capacities in pool/vdev

[Bageland2000]

I have a 5-drive pool (RAIDZ1). 4x3TB + 1x4TB. I have a fifth 3TB that I want to use replace the 4TB so I can free up the 4TB drive. Is there a way to do the replace/resilver so that if another drive were to fail in the resilvering process I could replace the 4TB drive I removed to get the pool back to an intact status?

 

[alpaca]

Nope. RAIDZ1 can tolerate a single drive failure. Manually choosing to offline and then replace a drive is effectively "failing" that drive. If a different drive were to fail during the resilver (a second "failure"), you are going to lose data.

 

[Bageland2000]

Dang, really? That seems counterintuitive. I obviously don't understand anything about the filesystem, but I would think you could "unfail" the drive and it would be like "oh, there your data is, you're ok." I guess it's not setup that way.

 

[NugentS]

What happens if any data on the pool has changed?

 

[Bageland2000]

Fair point, but I guess I thought there'd be a way to "pause" the pool to do a disk swap, which would allow for a failure to be fixed by returning the drive that's being replaced, allowing the revert the pool status back to when you began the replacement (but now in a degraded state with the pool status having a single drive now "actually" failed)

 

[danb35]

I thought there'd be a way to "pause" the pool to do a disk swap

Nope. However, you can replace the 4TB disk without offlining it first, if you can add the 3TB disk as an additional disk to the system while the resilvering is happening. No loss in redundancy that way.

 

[Bageland2000]

This is exactly what I want. How can I do this? I've searched forums, the TrueNAS manual, and YouTube videos and can't seem to figure it out.

 

[danb35]

How can I do this?

Install the new disk, replace the old disk with the new disk, once the resilver is complete, remove the old disk.

 

[Bageland2000]

Done. That was too easy. Every guide I've read says you have to offline the drive first. Oh well. Thanks so much for helping :)

 

[danb35]

If a disk is failing, it's often affecting pool performance. Offlining it means it's not going to slow down the resilver--but it also means you've lost redundancy during the resilver operation. Pros and cons.

 

[Bageland2000]

Makes sense, glad to be learning about the file system.