disk failure and badblocks

disk failure and badblocks

[seth vidal]

Hi,
 I had a disk fail out of sw-raid5 array with a medium error. I'd like
to check the disk for bad blocks, and if it shows up with some, call me
vendor and get them to replace the disk. In the past I've shut the
machine down and run the scsi bios media check. Would it be possible for
me to remove this disk from the array and run badblocks on it to get the
same information? If I could not shut down the system I would prefer to.
My major wonder is will badblocks be effective for looking for a medium
error?

Thanks,

-sv

Re: disk failure and badblocks

[Luca Berra]

On Sat, Nov 22, 2003 at 03:22:21AM -0500, seth vidal wrote:
>My major wonder is will badblocks be effective for looking for a medium
>error?
only if you can disable the disk onboard cache :)
a long time ago i changed badblocks write mode algo to write a pattern
to the full disk THEN check it, instead of write-block, check, next
block which was a victim of the cache. Pity i lost that patch :)

L.

-- 
Luca Berra -- bluca@comedia.it
        Communication Media & Services S.r.l.
 /"\
 \ /     ASCII RIBBON CAMPAIGN
  X        AGAINST HTML MAIL
 / \

Re: disk failure and badblocks

[seth vidal]

On Sat, 2003-11-22 at 09:12, Luca Berra wrote:
> On Sat, Nov 22, 2003 at 03:22:21AM -0500, seth vidal wrote:
> >My major wonder is will badblocks be effective for looking for a medium
> >error?
> only if you can disable the disk onboard cache :)
> a long time ago i changed badblocks write mode algo to write a pattern
> to the full disk THEN check it, instead of write-block, check, next
> block which was a victim of the cache. Pity i lost that patch :)

What if I made the number of blocks written at any time * the block byte
size greater than the onboard disk cache? That seems like it might
circumvent the caching problem.

-sv

Re: disk failure and badblocks

[Luca Berra]

On Sat, Nov 22, 2003 at 09:57:36AM -0500, seth vidal wrote:
>What if I made the number of blocks written at any time * the block byte
>size greater than the onboard disk cache? That seems like it might
>circumvent the caching problem.
it should.
L.

-- 
Luca Berra -- bluca@comedia.it
        Communication Media & Services S.r.l.
 /"\
 \ /     ASCII RIBBON CAMPAIGN
  X        AGAINST HTML MAIL
 / \