On Mar 29, 2013, at 10:34 PM, H. Peter Anvin <hpa@xxxxxxxxx> wrote: > On 03/17/2013 09:50 PM, Chris Murphy wrote: >> >> 3. If the installation target is a drive with a 61 sector MBR gap >> (legacy start LBA 63 for 1st partitions), the core.img is too large to >> be embedded if LVM or btrfs are chosen device types. So grub2-install >> fails, then anaconda crashes. >> >> Arguably the legacy partition scheme is flawed... > > ... or grub2 is flawed for blithly thinking that it is okay to use a > specific amount of unpartitioned space? GRUB Legacy did this also, so I'm unsure what you're arguing. On GPT, GRUB2 uses partitioned space (whether EFI or BIOS firmware is used). > (It isn't always 62 sectors, either.) As it turns out I don't consistently get a failure with a 1st partition starting at LBA 63. Usually the GRUB core.img is ~26K which will fit into that MBR gap. But sometimes it's bigger and I don't know why, and then it fails, and subsequently anaconda fails. Chris Murphy _______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list