Re: File System Corruption with 2.2.18

[Ville Herva]

On Wed, Jan 17, 2001 at 05:14:02PM -0800, you [Andre Hedrick] claimed:
> On Thu, 18 Jan 2001, Tim Fletcher wrote:
> 
> > > Well that is useless test them because you can not test things completely.
> > 
> > I ment that if the partiton has no persient data on it then the test can
> > be run (the test wipes all data on the partition out during the test,
> > right?) with no loss of data on the machine. The partition is still on the
> > same disk so the test data is valid?
> > 
> > I am thinking that the test is somewhat like badblocks -w or have I got
> > the wrong end of the stick?
> 
> Sorry there is no stick to get the end of....
> This is a pure diagnostic tool the determine OS/CHIPSET/DEVICE failures.
> You generate a pattern buffer and write it to the disk and step the buffer
> 1 byte per sector and go head to tail.  Then you read it back head to tail
> and compare what should be there with what is there.  Failures == FS
> corruption is likely under highest loads, period.  Then you attempt 
> to extract any patterns or periodic events to determine if it is driver or
> device or other portions of the OS.
> 
> I am tired of people pointing the finger at me claim my work is the cause
> of FS corruption.
> 
> This is a pattern walk and it will give some performance issue.
> It does not care about the OS, it is doing the direct access that some
> would call bit-bangging in the old days.

But it works on all ATA disks? Does it work for SCSI as well?

I think it would be cool if you'd make it available (on linux-ide.org?),
so that people install servers (and anybody who _cares_) could test their
hardware/driver combination before starting using the box.

Now we have memtest86, cpuburn (what more). It would be nice to have a
good (if not complete) test set to run on each new linux box. It's not
nice to use the box for a month and then go "f..., this box has faulty
memory!" or ..."faulty disk!". Yes, that's what's happened to all of us.

It's much nicer to get a warranty replacement, when you don't have any
data on the disk.

-- v --

[e-mail]