On Sun, Dec 28, 2003 at 11:47:31PM -0800, Larry Price wrote: > This is kind of a borderline question; > > A disk was intentionally zero'd out using > > dd if=/dev/zero of=/dev/hda > > however the DOS fdisk utility couldn't rebuild the partition table > afterwards.
Generally speaking some fdisk programs really can't cope with that. I thought the DOS one could. I know that if the disk isn't screwy, cfdisk will be able to fix it. A zeroing should not affect the low-level format because from the perspective of zeroing the drive, you're operating at a logical level rather than a physical level. A poorly kept secret is that HDs have things like bad sectors all the time and that the drives themselves have known for ages how to recognize sectors that are going bad before they do. When this happens, they silently reroute your bits to unused sectors they do not advertise as actually being available. This screws with things like interleaving, but nobody interleaves for a speed boost anymore. (Why? RAID finally really does mean inexpensive disks...) Anyway, you only see bad blocks when the drive runs out of spare sectors. In order to pull this off transparently, the PC needs a very different picture of the drive than the drive electronics get. The PC sees a virtualized disk, much the way Linux applications believe that they have some 3 gigs of memory, regardless of how much RAM and swap you actually have. Writing to the Linux memory space, or to the drive's virtual space, has little or no real effect on the low-level representation of what's actually stored on the platters or in the DRAM latch matrices. Low-level formats of drives are scary undertakings. Usually when I've needed to do it, it was because the drive was damaged. Usually, the reformat doesn't help matters even a little bit. Just a word of warning. _______________________________________________ EuG-LUG mailing list [EMAIL PROTECTED] http://mailman.efn.org/cgi-bin/listinfo/eug-lug