Re: Weird data corruption?

Haruhiko Okumura (okumura@matsusaka-u.ac.jp)
Fri, 9 Jul 1999 16:04:33 +0900 (JST)


hahn@coffee.psychology.mcmaster.ca wrote:

> > Meanwhile I encountered another data corruption (mutation of four
> > consecutive bytes of a very large file) on my SCSI disk a few days
> > ago, and Linux didn't report a CRC error. Is this normal?
>
> which part? it's not normal to corrupt files. but it can happen
> due to disk faults, or to host bugs or faulty host hardware.
> if it happened due to errors in transfer, the SCSI parity check
> would probably have noticed, and the transfer would be retried
> (assuming your scsi has parity enabled. note that parity is a very
> weak form of corruption-detection, unlike the checksum used by UDMA.)
> I forget whether retries due to detected parity errors are reported
> or not. if your host uses ECC dram, then it looks a lot like a
> kernel, filesystem or app bug.

My DRAMS are not ECC-enabled. On the other hand I never experienced
hangs or kernel oopsen.

Do I have to switch CONFIG_SCSI_CONSTANTS on to get SCSI checksum
errors reported?

My experiences of data corruptions are detailed on

http://www.matsusaka-u.ac.jp/~okumura/linux/corruptdata.html

-- 
Haruhiko Okumura, Ph.D. <okumura@matsusaka-u.ac.jp>
Matsusaka University, 1846 Kubo-cho, Matsusaka, 515-8511 Japan
Phone: +81-598-29-1122  Fax: +81-598-29-1014
http://www.matsusaka-u.ac.jp/~okumura/

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/