Re: SCSI Sector Size Problem

Nicholas J. Leon (
Fri, 1 Nov 1996 20:24:51 -0500 (EST)

Philip Blundell was overheard whispering ...
# On Fri, 1 Nov 1996, Nicholas J. Leon wrote:
# > # mke2fs will do bad-block checking if you use the -c option. This, like
# > # MS-DOS reads the entire disk drive during its high-level format. The
# > # problem is that this doesn't catch all errors.
# > #
# >
# > It doesn't even catch the errors it reports. I have to disagree,
# > mke2fs (with the -c option) doesn't do anything about the bad blocks.
# >
# > Example: I was reformatting a drive the other day and had a HOST of
# > errors. In one window I had a "tail -f /var/log/kernel" running,
# > watching the errors appear:
# >
# > hdb: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
# > hdb: read_intr: error=0x40 { UncorrectableError }, LBAsect=306946, sector=306945
# > end_request: I/O error, dev 03:41, sector 306945
# >
# > Then after all is said and done, I do a "dd if=/dev/zero of=/dev/hdb"
# > and sure enough, I get errors in the exact same locations.
# >
# > If mke2fs was mapping (or marking) those blocks away, I wouldn't get
# > the error again, would I? Or am I missing something vital here?
# You're missing the point. mke2fs doesn't remap the sectors, because
# there's no generic way to do it (and many drives can't). Nor can it make
# them magically go away. All it does is mark them as bad in the
# _filesystem_ so that no data will be stored there. If you do "dd
# if=/dev/hdb" you bypass the filesystem layer altogether, and so any bad
# blocks marked there will be ignored.

Oops! The second part of my argument (which I apparently forgot) was
that I then remade the filesystem (to compare the bad sectors,
*MOSTLY* the same) and then

cat /dev/zero > /mnt/foo

and it also generated the same errors??

Sorry 'bout that!

Nicholas J. Leon
"Elegance through Simplicity"

Binary 9 offers free mail and web storage to certain
individuals and groups.