Re: EXT3-fs error

[Louis Hopcraft <lhopcraft telus net>]

First of all, I'd like to thank all of you for responding.  It really
helps in my time of need.  ;-)

And second, Thank God for Backups. ;-)  I had been backing up this box
(my production box, which I never have problems with) to my laptop and
another rack mount box every night. ( I wanted to be extra carefull).  

So when everything went south on it this weekend I had not missed a
step.  All my data and current programs where already "cloned"
(nightly)  And I just went to work.

Since yesterday, I now have alot of "error messages in my log file.  As
Stephen has mentioned 
>"I very much doubt it's anything other than a drive fault."

And I'm thinking that I'll agree with him on this one.  I only wished it
was still under warranty.

Thank you

Louis


On Mon, 2002-01-21 at 05:31, Stephen C. Tweedie wrote:
> Hi,
> 
> On Mon, Jan 21, 2002 at 01:38:55PM +0100, Martin Eriksson wrote:
> > 
> > > On Sun, Jan 20, 2002 at 05:56:48PM -0800, Louis Hopcraft wrote:
> > > > No matter what I do (Y or N) the results are the same.  The problem will
> > > > not go away.  Does anyone haI very much doubt
> it's anything other than a drive fault.ve any ideas as to a fix, or to what the
> > > > problem is?  When I do a fsck and it marks the fs as clean I reboot and
> > > > hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> > > > hdb: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=81423,
> > >
> > >   Your IDe drive is dying. backup immediately and try to get the warranty
> > > to apply.
> > 
> > This can also be because of IDE drives running out of spec, ie. having an
> > overclocked PCI bus, a buggy drive or a buggy controller. One of my drives
> > exhibited typical "bad sector" behaviour, when in fact the drive couldn't
> > run on it's rated UDMA33 speed and had to be manually tuned for MW-DMA2.
> > (Now I run with Andre's IDE patch which automatically tunes "down" the drive
> > because it is in som "quirk" list or something).
> 
> The error was "SectorIdNotFound", which is a specific message coming
> back from the drive that the sector header on the platter was not
> present.  DMA problems can manifest themselves as all sorts of data
> corruption failures, but this is not such an error.  I very much doubt
> it's anything other than a drive fault.
> 
> Cheers,
>  Stephen