Replacing a failed/failing disk

[Patrick M. Hausen]

Out of curiosity:

What is it about that volume manager button? I prefer to do my maintenance on the command line, so I was never tempted to click it ;)
And second, why the screenshots of the legacy UI? Is anyone still running this in numbers?

 

[danb35]

What is it about that volume manager button?

I don't know, but a whole lot of people in the process of attempting to replace a disk end up adding a new disk to their pool--and the only way that can (could) happen is using the volume manager button.

why the screenshots of the legacy UI?

Because I wrote the guide when it was "the GUI", and haven't gotten around to updating it yet.

 

[Patrick M. Hausen]

I should have looked at the dates. Necroed thread. Thank you.

 

[danb35]

Necroed thread.

Still a resource, so it really should be up-to-date. Just need to make the time to do it.

 

[LittleGreyCat]

It assumes no such thing, only that you're running a pool that can survive having a disk offline. That can be a mirror or RAIDZ1/2/3.

What are you seeing that makes you think anything here depends on the pool being a mirror? I'd like to change it to clarify.

The screen shots showed a mirror, and you used the term "resilvering" which I took to be recreating the mirror pair.

 

[LittleGreyCat]

One important point: In no way is a crummy, overpriced SoC running black box code with bugs proportional to the number of tentacles it tries to grow (A webGUI for the RAID controller? What a horrible idea...) more proper than ZFS.

I don't believe that I described the real RAID controller I worked with in that manner.

FTAOD it was in the mid '90s and it was a Clariion array which managed itself in a very competent fashion, had multiple SCSI cards for resilience, and cost probably more than I am ever likely to earn.
It was one big enterprise beast.
Originally manufactured by Data General.
It was connected to a Data General Unix server.
Serious RAID array - Wikipedia

 

[jgreco]

I don't believe that I described the real RAID controller I worked with in that manner.

FTAOD it was in the mid '90s and it was a Clariion array which managed itself in a very competent fashion, had multiple SCSI cards for resilience, and cost probably more than I am ever likely to earn.

Well, looking at what @Ericloewe wrote,

One important point: In no way is a crummy, overpriced SoC running black box code with bugs proportional to the number of tentacles it tries to grow (A webGUI for the RAID controller? What a horrible idea...) more proper than ZFS.

What you have literally just described is a crummy, overpriced VERY LARGE BOX running black box code with bugs proportional to the number of tentacles it tries to grow.

So shame on @Ericloewe for not thinking big scale.

I think I still have some Mylex DAC960's floating aroun which were popular as SCSI RAID array controllers in the mid to late 90's. Just in case anyone wants to relive the horror.

 

[danb35]

you used the term "resilvering" which I took to be recreating the mirror pair.

You misunderstand "resilvering"--it includes the process of recreating a mirror pair, but it also includes replacing a disk in a parity RAID set.

 

[LittleGreyCat]

You misunderstand "resilvering"--it includes the process of recreating a mirror pair, but it also includes replacing a disk in a parity RAID set.

OK - thanks.
Please note that as a Newbie I was confused over a couple of the aspects - mirrors and resilvering - and by the fact that the screen shot showed a mirror pair - so I assumed the article was for mirrored pairs only.
It could help clarify (when you get round to updating the article) to clearly state up front that this applies to all forms of resilient storage - as you originally replied to me and as you have noted above here.

Not knocking the article in any way - I found it very informative. :)

 

[Ericloewe]

The term resilvering is inherited from the early days of ZFS, when only mirror vdevs (and leaf vdevs, otherwise known as disks) existed (kinda feels like saying "before mitochondria existed").

I'd never thought about it, but I do see how it can be misleading, since the etymology does indeed draw from the mirror metaphor.