thanks, was: Re: data corruption with DISABLE_PSE+DISABLE_PG_G: unrelated

[Vallo Kallaste]

On Sat, May 10, 2003 at 11:11:43PM +0200, Heiko Schaefer
<hschaefer at fto.de> wrote:

> > > what i am asking myself:
> > > is there any chance that i still get any data corruption because of the
> > > issues that you write about in some configuration ?!
> >
> > No.  Not with thouse flags set.  If you are getting data
> > corruption with the flags set, then you have some other
> > problem, most likely hardware.
> 
> ok, that's the answer i was looking for - as solid as it can get, with
> such issues, i guess :)

Heiko, I'm not able to reproduce it on Dual Athlon 2400+ system with
1GB of ECC memory and two Hitachi 120GB disks. Created two ~90GB
filesystems on both disks with plain newfs, no softupdates. Then
ran the following to populate source filesystem:
while :; do size=$(jot -nrw %dk 1 4000 7000); split -b ${size} $HOME/film/LOTR1.avi ${size}. ; done

The source filesystem has ~45GB of data in ~7300 files. I used the
following quick script to copy data and control the checksums. It's
seventh run currently and no signs of corruption. I'll let it run
through tomorrow just to be on the safe side. Seems like you have
nasty hardware problem(s).

#! /usr/local/bin/bash

src="/opt/TESTDIR"
dst="/opt1/TESTDIR"

for ((i=0;i<10;i++)); do
	for dir in 1 2; do
		find ${src} -type f -print0 | xargs -0 -J % cp % ${dst}/${dir}
	done

	for dir in 1 2; do
		cd ${dst}/${dir} && \
		find -s . -type f -print0 | xargs -0 md5 -r \
		> ${HOME}/result${dir}.txt
	done

	for dir in 1 2; do
		cmp ${HOME}/result${dir}.txt ${HOME}/checksum.txt
	done

	find ${dst}/{1,2} -type f -print0 | xargs -0 rm
done
-- 
Vallo Kallaste