1. Another issue with install and GPT (Scott Robbins)
Date: Sat, 17 Sep 2011 07:52:55 -0400
From: Scott Robbins <scottro(a)nyc.rr.com>
Subject: Another issue with install and GPT
To: For testers of Fedora Core development releases
Content-Type: text/plain; charset=utf-8
I accidentally sent this through a different address, so I think it's
just going to come back to me. If it turns out to be a double post,
please forgive me.
I'm not sure if this is a bug, or a one time fluke. I don't have time
or hardware to do several test installs, but perhaps someone else has
What happened: I used unetbootin (yes, I know it's not supported, but I
believe dd-ing an image to USB is supported, and might have the same
I installed. I was not notified of a problem with not having a small
BIOS partition and thought the machine might have EFI support.
After installation, I was unable to boot. (This was at work, while
doing several things, and didn't write down the error.) Having seen
some bugs having to do with SELinux, also tried to boot with enforce=0
but no luck.
A-) Try using dd directly:
# dd if=Fedora-16-Beta-RC1-i386-netinst.iso of=/dev/sdg bs=2M
132+0 records in
132+0 records out
276824064 bytes (277 MB) copied, 28.6564 s, 9.7 MB/s
The resulting USB Boots
anaconda fine (just tested it on ACER ASPIRE ONE
N450) used it to write to /dev/sdb (2nd target USB)
graphical desktop/customize now /sugar-desktop only
B-) Here are 2 methods that work for me on f15-f16 live isos:
1-)Booted CD: use diskutility on live CD or gparted to erase USB
(/dev/sdb?) then format as ext4
use "liveinst" to USB custom:/ ext4 no swap
(2x time as large USB req - swap is a bad use on a USB)
2_)For fedora liveusb-creator use terminal: "liveusb-creator --reset-mbr"
choose .iso; set persistence size. ( use gparted fat formatted USB here)
satellit_ on #sugar freenode IRC
I then reinstalled, this time making a 1 or 2 MB BIOS partition, It
installed with problem.
What I think might have happened:
When I got to partitioning, the USB stick was shown as /dev/sdb. It had
the EFI partition on it, put there, I imagine, when Unetbootin installed
the netinstall.iso that I used. Therefore, Anaconda saw that there was
a BIOS partition and didn't give a warning.
If this is the case, IMHO, it should just be pointed out in the release
notes. Ideally, Anaconda would figure it out and point it out, but I
imagine that would be a lot of work, and it's late, and certainly not,
again IMHO, a blocker.
I reiterate, I don't know if that is what happened, it's just what I
think happened, and perhaps there was some other thing that I missed
that caused the first install to not boot. Perhaps someone could test
this--install, from a USB onto a drive with no BIOS boot partition (and
no builtin EFI support), see if such a partition is shown on the USB
stick and see if, after installation, it boots.