[GLLUG] how to remove

[Mike Rambo]

----- Original Message -----
From: tigner <tigner at msu.edu>
To: Mike Rambo <mrambo at lsd.k12.mi.us>
Cc: linux-user at egr.msu.edu
Sent: Fri, 16 Jan 2009 16:28:57 -0500 (EST)
Subject: Re: [GLLUG] how to remove

It is very important IMHO to perform a low level test of
the disk anytime possible corruption occurs. A sector
or group of sectors can become marginal. Manufacturers
of disk drives provide free low level testing software.

I would run the low level test before I tried to fsck the
drive, it could mask the problem -or- make it worse.

I recommend turning on S.M.A.R.T on your drives and
running S.M.A.R.T reporting software on your system.
Many times S.M.A.R.T testing or low level testing will
let you know if the drive is failing or about to fail and
if this is the case, then you are wasting your time
running high level OS file recovery routines suck as
fsck, chkdsk, drvchk etc.

Know your drive is good before you mess with the data.


Barry



This is an array of 6 old 9GB 10000 rpm scsi drives on a compaq
hardware raid controller. I don't think they support smart.
Since the data is striped across all disks in a raid 5 array
any single failed disk shouldn't cause this problem anyway. The
remaining files and directories (all two dozen of them) seem
to be ok.

I'll probably try to copy the data to other media on Tuesday
when I get back to work and then rebuild the array - perhaps
with another raid controller. I use this particular array mostly
to hold linux distro iso's until I'm done with them. Fortunately
there is almost nothing there that result in any lost sleep were
it to be lost anyway. I was mostly curious about what I could do
to recover this data just in case I ever have the problem where
the data is actually important.

Thanks to everyone for the suggestions.


On Fri, 2009-01-16 at 15:20 -0500, Mike Rambo wrote:

> On Fri, 16 Jan 2009 14:31:50 -0500
> "Stanley C. Mortel" <mortel at cyber-nos.com> wrote:
> 
> > Mike,
> > 
> > Have you tried badblocks and fsck?
> > 
> 
> fsck was the first thing I did. It doesn't see anything wrong with the
> file system. Guess whatever is the cause is not something that offends
> fsck. I did not try bad blocks though. I thought that was for marking
> bad blocks on the disk so they were not used at all. Not so?
> 
> All I really need to do is delete the now extraneous entry although if it
> were it recoverable I think it would be an interesting learning exercise.
> 
> 
> > Stan
> > 
> > Mike Rambo wrote:
> > > I had a problem with a raid 5 array. When I got it back up I found
> > > everything looks ok except one directory that now appears as
> > > follows.
> > >
> > > ?r---w-rwx 33444 3714798861 2557849272 2593351921 2017-08-26 12:38
> > > mdk
> > >
> > > Notice the ? instead of the usual d at the far left plus the weird
> 
> <snip>
> 
> > >
> > > I can't even delete this item.
> > >
> > >
> > >   
> 
>