Re: ext2 errors

Ion Badulescu (ionut@moisil.cs.columbia.edu)
Sun, 6 Jul 1997 23:59:40 -0400 (EDT)


On Sun, 6 Jul 1997, Mr. James W. Laferriere Network Engineer wrote:

> > > > Jul 5 15:26:31 soda kernel: hda: dma_intr: status=0x51 { DriveReady
> > > > SeekComplete Error }

[...]

> > > This has nothing to do with ext2, it's your hardware that's failing.
> > > Usually such messages denote a dying IDE hard drive, altough sometimes a
> > > faulty IDE interface can create problems as well.
>
> > I don't think so. I've seen this happening also with 2.0.30,
> > an Intel SMP chipset and a Quantum Fireball. It started right
> > after I made the switch from 2.0.27 to 2.0.30 and happened
> > consistently during one week befor it was noticed when
> > an ugly crash with extensive ext2 corruption happened.
>
> > After going back to 2.0.27 I have not seen it again and it
> > has been under moderate load (8 regular users and a web crawler,
> > which makes heavy use of disk IO to store and retrieve indexes)
> > for about 2 weeks now.
>
> Which controller are you using ?

[...]

> I must have missed it in the previous posts , But
> I can tell you from -experience- that in one version
> of a driver to another can bring out a hiding bug in
> the on drive code or even a latent one in the drive itself.
>
> Everytime I went back to the 1.12c version all was
> well again. Until about 4 weeks later then the drive
> fried , couldn't even lowlevel format the beast .
> It kept coming back with 'drive bad block table full'
> or something very simular.
>
> Be warned it -may well be- a disk on its way to the trash.

.. especially that it's the first report I've seen that links the upgrade
from .27 to .30 to IDE errors. As far as I know there haven't been any
changes to the IDE code recently in the 2.0 series.

My advice is to upgrade gradually from .27 to .30 and see which particular
version breaks your system. Another thing to try is to stick with the
vanilla IDE driver with no fancy custom interface support - my experience
with these interfaces (not Triton, though) has been shaky at best, and
_did_ result in extensive damage to the filesystems in a couple of cases.
As Jim said, it might also be a dying disk.

Again though, this has _absolutely nothing_ to do with ext2.

Ionut

-- 
  It is better to keep your mouth shut and be thought a fool,
            than to open it and remove all doubt.