Hi, Pascal Hambourg wrote: > Fdisk [...] calls the system call to tell the > kernel to re-read the partition table
Ah yes. That's plausible. With my previous Linux i would have had to unplug and re-plug the stick. But Debian 8 noticed the new partition without that: $ ls -l /dev/sdc* brw-rw---- 1 root thomas 8, 32 Aug 18 20:44 /dev/sdc brw-rw---- 1 root disk 8, 33 Aug 18 20:44 /dev/sdc1 brw-rw---- 1 root disk 8, 34 Aug 18 20:44 /dev/sdc2 brw-rw---- 1 root disk 8, 35 Aug 18 20:44 /dev/sdc3 systemd magic ? i wrote: > > Let's hope the USB stick still boots via BIOS and EFI ... > It boots via any decent BIOS. That would be the more likely candidate for failure, because the BIOS hops on the x86 code in the MBR of which the partition table got manipulated by fdisk. > I haven't tested EFI boot with it, but I don't > see why it would not work as the new partition is "regular". As long as we have only one partition of type 0xef and not overlapping partitions of non-zero type, EFI should be happy. From MBR it takes only the partition table entry to find its FAT filesystem with the boot programs. Now it has to be tested whether or how easily the installer can make use the files in partition 3. > > > https://www.debian.org/releases/jessie/amd64/ch04s03.html.en This looks pre-isohybrid, i.e. two generations behind of debian-cd for i386 and amd64. I understand arm64 is suitable for CD and USB stick, too. Some other arches make no difference between CD and disk at all. Have a nice day :) Thomas
