Re: SCSI Sector Size Problem

Harald Koenig (
Sat, 2 Nov 1996 12:01:28 +0100 (MET)

> 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.
> The badblocks program (which is what mke2fs calls) also has an option
> for doing a distructive read/write test. The problem is that it takes a
> *lot* longer, and there currently isn't an option to mke2fs which will
> cause it to run badblocks with its -w option.

I needed this only once for a SCSI disk with lots of bad blocks
which was connected to a NCR810 controller using Drew's driver;
"mke2fs -c" or directly running "badblocks" didn't work for me.

after some testing it showed up that at Drew's NCR driver (the BSD driver
wasn't yet available that time) doesn't report any error to the user
program when trying to read bad blocks, it always returns all the data
requested and only sends some error messages to syslog!

connecting that bad drive to an old AHA1542 (using the same Linux kernel)
fixed this problem and "badblocks" and thus "mke2fs -c" worked as designed,
I could mark all those bad blocks and use the driver later connected to the
NCR810 again...

so it might be a problem with Linux's low level disk device driver too...


All SCSI disks will from now on                     ___       _____
be required to send an email notice                0--,|    /OOOOOOO\
24 hours prior to complete hardware failure!      <_/  /  /OOOOOOOOOOO\
                                                    \  \/OOOOOOOOOOOOOOO\
                                                      \ OOOOOOOOOOOOOOOOO|//
Harald Koenig,                                         \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik                              //  /     \\  \                     ^^^^^       ^^^^^