General rebuild question with loss of 2 disks

[foo_fighter]

Let's say you lose 2 disks, a 100GB and a 4TB disk. Technically, you should be able to rebuild all but 100GB of the 4TB drive. But in practice I suspect it won't let you do that. Does that mean that keeping around smaller disks is inherently more risky as the loss of that disk(during multiple failures) impacts more than just the data on it?

[eroz]

If you lose two disks, you lose the data on the two disks.  it does not matter on the size of the disks.  But if one of those disks is your parity drive, then you only lose one data disk.  And you would need to rebuild parity again.

[itimpi]

If you lose two disks, you lose the data on the two disks.  it does not matter on the size of the disks.  But if one of those disks is your parity drive, then you only lose one data disk.  And you would need to rebuild parity again.

It is worth pointing out that because unRAID has a full file-system on each disk, if the disk is not physically failed, then it is often still possible to mount a 'failed' disk outside unRAID to recover data from it.

[JonathanM]

That's one way to look at it. You should always be 100% confident in all your disks, all the time. If you suspect a disk is bad before it fails, the only proper response is to deal with it before it fails. Surprise failures can bite you at any time, so the less risk the better.

 

In your example, you could replace the 100GB with a precleared disk of any size up to the size of your parity, replace the 4TB with a 4TB or larger depending on parity, and use the procedure documented somewhere on here to force the 4TB replacement to be rebuilt with the 100GB slot set as valid. That would allow you to have valid data above the 100GB mark, which may allow some recovery with a full disk rebuild tree scan.

[foo_fighter]

Ok, that makes sense.

 

In your example, you could replace the 100GB with a precleared disk of any size up to the size of your parity, replace the 4TB with a 4TB or larger depending on parity, and use the procedure documented somewhere on here to force the 4TB replacement to be rebuilt with the 100GB slot set as valid. That would allow you to have valid data above the 100GB mark, which may allow some recovery with a full disk rebuild tree scan.