Re: Find mismatch in data blocks during raid6 repair

Hello Piergiorgio,

On Tuesday, July 03, 2012 10:27:34 PM Piergiorgio Sartor wrote:

Hi Robert,

On Tue, Jul 03, 2012 at 09:10:41PM +0200, Robert Buchholz wrote:
[...]

quoted

quoted

Why always two blocks?

The reason is simply to have less cases to handle in the code.
There's already three ways to regenerate regenerate two blocks
(D&D, D/P&Q and D&P), and there would be two more cases if only
one block was to be repaired. With the original patch, if you can
repair two blocks, that allows you to repair one (and one other
in addition) as well.

sorry, I express myself not clearly.

I mean, a two parities Reed-Solomon system can
only detect one incorrect slot position, so I would
expect to have the possibility to fix only one, not
two slots.

So, I did not understand why two. I mean, I understand
that a RAID-6 can correct exact up two incorrect slots,
but the "unknown" case might have more and correcting
will mean no correction or, maybe, even more damage.

Well, if two slots have failed and you do not know which, or more than 
two have failed, there is no way to recover anything reliably.
I implemented the two slot fix to recover from a case where you *do* 
know which two slots failed (e.g., from syslog messages such as this: 
end_request: I/O error, dev sdk, sector 3174422). Obviously, this 
expects a lot of knowledge from the admin running the command and 
selecting the slots and comes with no guarantees that the "repaired" 
blocks will contain more of the expected data than before.

I would prefer, if you agree, to simply tell "raid6check"
to fix a single slot, or the (single) wrong slots it finds
during the check.

Does it make sense to you, or, maybe, you're considering
something I'm missing?

This makes perfect sense.

quoted

quoted

Of course, this is just a statistical assumption, which
means a second, "aggressive", option will have to be
available, with all the warnings of the case.

As you point out, it is impossible to determine which of two
failed
slots are in error. I would leave such decision to an admin, but
giving one or more "advices" may be a nice idea.

That would be exactly the background.
For example, considering that "raid6check" processes
stripes, but the check is done per byte, already
knowing how many bytes per stripe (or block) need
to be corrected (per device) will hint a lot about
the overall status of the storage.

That piece of information is definitely interesting. What is the 
smartest way to determine the number of incorrect bytes for one failed 
slot?

quoted

Personally, I am recovering from a simultaneous three-disk failure
on a backup storage. My best hope was to ddrescue "most" from all
three disks onto fresh ones, and I lost a total of a few KB on
each disk. Using the ddrescue log, I can even say which sectors
of each disk were damaged. Interestingly, two disks of the same
model failed on the very same sector (even though they were
produced at different times), so I now have "unknown" slot errors
in some stripes. But with context information, I am certain I
know which slots need to be repaired.

That's good!
Did you use "raid6check" for a verification?

Yes, since John Robinson pointed me to it earlier in this thread.

quoted

I am a big supporter of getting it to work, then make it fast.
Since a full raid check takes the magnitude of hours anyway, I do
not mind that repairing blocks from the user space will take five
minutes when it could be done in 3. That said, I think the faster
code in the kernel is warranted (as it needs this calculation
very often when a disk is failed), and if it is possible to reuse
easily, we sure should.

The check is pretty slow, also due to the terminal
print out, which is a bit too verbose, I think.

That is true. The stripe geometry output could be optional, especially 
when there is no error to be reported.

Anyhow, I'm really happy someone has interest in
improving "raid6check", I hope you'll be able to
improve it and, maybe, someone else will join
the bandwagon... :-)

Well, thank you for starting it and sorry for my slow replies.


Cheers

Robert