> > The s8 "boot" slice should contain the > > /usr/platform/i86pc/lib/fs/ufs/bootblk code. > > > > installboot should write the bootblk code to the "s8" / "boot" slice. > > > > I think s8 may not overlap with the root slice; the x86 bootblk code it too > > big for an UFS > > filesystem, so it can't be installed into the root UFS but uses the > > separate "boot" s8 slice. > > > > Maybe your new HDD is missing the "s8" slice, or "s8" overlaps the root > > UFS, so by installing the > bootblk code with installboot to "s8" the UFS root filesystem is thrashed? > > Yes, I did not ufsdump the s8 and s9 slices. I thought they were not > necessary because > /usr/lib/fs/ufs/mboot did take care of this (see step 1 below)
You can't ufsdump them, they do not contain an UFS filesystem. > Therefore are you saying I should do like this: > > 1. fdisk -b /usr/lib/fs/ufs/mboot /dev/rdsk/c0d0p0 > (new active boot disk) > 2. installboot /usr/platform/i86pc/lib/fs/ufs/pboot > /usr/platform/i86pc/lib/fs/ufs/bootblk > /dev/rdsk/c0d0s8 No, use the standard installboot command from the EXAMPLES section in the installboot(1M) manual page, which uses the s2 slice on Solaris x86. But make sure that the s8 slice is defined and starts on cylinder 0. And make sure that the root filesystem slice (or any other filesystem slice) does *not* start on cylinder 0. > I understand /dev/rdsk/c0d0p0 is also mapped to > /dev/rdsk/c0d0s8 isn't it? No, c0d0p0 is the whole disk, ignoring any MBR or Solaris disk label partitions. As soon as we have a Solaris x86 MBR fdisk partition with a valid SunOS disk label the c0d0p0 device should be different from any c0d0sX device. This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list [email protected]
