On Saturday 29 May 2004 13:31, Erwan MAS wrote: > On Sat, May 29, 2004 at 09:50:38AM -0700, Philippe Troin wrote: > [../..] > > | > the first 512 byte of /dev/hda and the 512 byte of /dev/hda1 are > | > identical . > | > | That's the case if hda1 starts at cylinder 0. And it is unlike the > | PC where the first 512 bytes (the MBR) is not part of any > | partition. > > [../..] > > Ok in this case ,the partition table , is in the first 512 of > /dev/hda1 . > > When you format /dev/hda1 , the first 512 bytes is not > use by any filesystem ?
Nothing except for swap partitions. So, you should be fine as long as you don't start a swap partition on cylinder 0. (Swap isn't a real filesystem anyways, which is why is uses the first sector of disk). > In this context ,i have lost my partition table . After this > command : > mke2fs -c -c -j /dev/hda1 . > > " -c -c " checks for bad blocks with writing tests . Well, don't do that. Just a single -c, the read test, should be enough to find bad blocks. What you did is basically the same as "dd if=/dev/zero of=/dev/sda1" followed by a mke2fs. If you do a normal mke2fs (or one -c), it won't overwrite the first disk sector (512 bytes), which is where the disklabel lives. Pat -- Purdue University ITAP/RCS --- http://www.itap.purdue.edu/rcs/ The Computer Refuge --- http://computer-refuge.org