From: jwaters@deepthought.unm.edu (Joe Waters) Subject: Question on MFM bad block management Date: 29 May 1992 21:18:44 GMT
Okay, here's my problem. I've got a 40 meg Priam MFM drive as my /usr drive -
there are some bad sectors on the disk. When I low-level formatted, they got
put into the bad track table on the drive... But I just started having
problems with bad blocks in Linux - I do a "fsck /dev/hdb1" and it runs for
a while and then either a) reports a HD I/O error (after a couple of HD
resets) or b) it crashes completely (if I'm in X11). Is there any way to
map out bad blocks if I know where they are? Recently I've been trying to
recompile the kernel but when I untarred a new version of the source into
/usr/src/linux, part of the source just happened to fall across those bad
blocks. Needless to say the kernel compile explodes. Before, I was trying
to get the new gcc2.11c onto my system and cpp untarred across bad blocks,
so that would die too... Is there any way to map these blocks out so that
the file system will save around them? Does the linux file system look at
the bad sector map hardcoded onto the drive at format time?
Thanks for any help. Other than those few HD problems, everything is running
exceptionally well; I'm impressed at the directions this OS is going. Thanks
for all the hard work, all the people who have been doing such ground-breaking
development....
Thanks,
Joe