Solved the problem myself. And found a alternative way to set up raid.
I have to identical disk hda, hdc. 1) Install debian on hda. hda1 /boot hda2 / hda3 Swap 2) config raidtab with hdc af failed. md0 for boot md1 for root 3) config fstab and lilo.conf md0 for boot md1 for root 4) clone the disk dd if=/dev/hda of=/dev/hdc This take some time. 5) start raid raidstart /dev/md0 raidstart /dev/md1 6) write new superblock mke2fs -S /dev/md0 mke2fs -S /dev/md1 7) change /boot to md0 umount /boot mount /dev/md0 /boot 8) run lilo lilo 9) reboot and test reboot (when up again) df this should give /dev/md1 / /dev/md0 /boot 10) hotadd some disk remove failed-disk from /etc/raidtab then raidhotadd /dev/md0 /dev/hdc1 raidhotadd /dev/md1 /dev/hdc2 This works for me. Knud Knud Sørensen wrote: > > I have a raid1 device for which fsck returns a error. > It says that the physical size of the device is 26000k but > that the superblock says that the size is 26066k. > > if I run mkraid --upgrade it tells me that > the physical size is 26066k and that the superblock > starts at 26000k. > > What should i do to make fsck successful? > > > > Knud > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]