Raid5 Read stats for parity cal...

Raid5 Read stats for parity cal...

[Linux Raid Study]

Hello:

I understand that parity calculations are not done as part of Raid5
read operation in normal running conditions.

However, it would be done during disk failure or when a sector is bad.
Is there a way (any stats) available to see if parity cal are done
when sector is bad (since this is not as obvious as a disk fail...)

Thanks!

Re: Raid5 Read stats for parity cal...

[Majed B.]

You can force an array check by writing "check" (without double
quotes) to a file named "sync_action" found in /proc/sys

Read this for further details: http://linux.die.net/man/4/md

You can run that check at any time, not just when there's a problem.

On Thu, Sep 17, 2009 at 3:13 AM, Linux Raid Study
<linuxraid.study@gmail.com> wrote:
> Hello:
>
> I understand that parity calculations are not done as part of Raid5
> read operation in normal running conditions.
>
> However, it would be done during disk failure or when a sector is bad.
> Is there a way (any stats) available to see if parity cal are done
> when sector is bad (since this is not as obvious as a disk fail...)
>
> Thanks!
> --
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>



-- 
       Majed B.
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Raid5 Read stats for parity cal...

[Neil Brown]

On Wednesday September 16, linuxraid.study@gmail.com wrote:
> Hello:
> 
> I understand that parity calculations are not done as part of Raid5
> read operation in normal running conditions.

Correct.

> 
> However, it would be done during disk failure or when a sector is bad.

Also correct.

> Is there a way (any stats) available to see if parity cal are done
> when sector is bad (since this is not as obvious as a disk fail...)
> 

Maybe you want
   /sys/block/mdX/md/dev-YYY/errors

which counts the number of sectors for which md found a read error and
had to generate correct data using parity, and then wrote that correct
data back.

NeilBrown