SUMMARY: Fsck sees (Phantom?) bad block

From: Dave Plummer (dlp@medphys.ucl.ac.uk)
Date: Mon Oct 11 1993 - 12:40:33 CDT


Dear all,

First thanks to all those who responded:

heas@chpc.org ()
D.Mitchell@dcs.sheffield.ac.uk (David Mitchell)
hkatz@NUCMED.MED.NYU.EDU (Henry Katz)
poffen@k2.San-Jose.ate.slb.com (Russ Poffenberger)
glenn@uniq.com.au (Glenn Satchell)

And anyone else I missed.

The question (in short) was:

> Sparc 2, SunOS 4.1.3, MAXTOR P17 SCSI disk
>
> Why does fsck report an unreadable block on a disk, when format can be run in
> analyze mode repeatedly over the offending block and the whole disk with no
> errors reported.

My assumption was that some filesystem corruption was at fault as I foolishly
thought that fsck would be more complex and more likely to be the problem.

However David Mitchel indicated, in terms I will not repeat here, that format
was somewhat limited in its ability to detect disk errors, and that I should,
go ahead and repair the apparently good sector.

Indeed, format repair followed by fsck restored the disk to usability.
Dave also suggested running /usr/etc/icheck -b 51936 to confirm that the
block was an inode. It was, and refered to a directory full of other
directories. It was thus no major problem to recover the disk.

Almost all the respondents said go ahead and repair the disk. No one explained
how such a fundimental program as format could fail to pick up what was
obviously a hard error.

Sun are you listening ?

Dave

Dave Plummer, Dept of Medical Physics, University College London, UK



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:08:23 CDT