SOLVED Unreadable (pending) sectors

[Joseph Lennemann]

And this not only means that you didn't follow the youtube instructions "to a T", you weren't even in the same county--the youtube video has you using the GUI to replace, not the CLI.

I guess you didn't read the previous posts before making such a comment. Thanks for the reply.

 

[Joseph Lennemann]

So, if you offline the drive and wipe it, you should be able to re-ad it via replace via the gui.

If you add it via cli, you need to add it via gptid. I believe it's possible for the device to get lost otherwise down the track.

Be *sure* you offline the right one ;)

I did this once too.

^ This resolved the issue. Thanks Stux!

 

[emk2203]

I want to piggyback on the post with the same issue, but a different question:

During replacement, the resilvering takes a lot of time (50+ hrs). I know zfs recommendations are primarily for a "always-on" requirement. If I can take the server offline for some hours, is there a way to get this done quicker?

I was thinking of taking out a drive, dd'ing the contents to the new drive and then replacing the old disk before switching the server on again. Is this feasible? Can you grow a pool if you use larger disks if you use this method?

 

[Ericloewe]

If I can take the server offline for some hours, is there a way to get this done quicker?

No. All ZFS operations happen on online pools.

I was thinking of taking out a drive, dd'ing the contents to the new drive and then replacing the old disk before switching the server on again.

No, bad idea. That's just begging for errors to happen.

Can you grow a pool if you use larger disks if you use this method?

That one is not only a bad idea, but it doesn't work at all. The larger drive needs to be resilvered in.