Hi,

When you run the installer (e.g. flash "install62.fs" onto an USB
memory stick) and install onto a softraid (e.g. by going into (S)hell
mode at boot, creating a GPT partition table on your system drive e.g.
"fdisk -igy -b 960 sd0", creating a softraid on sd0 using "bioctl" that
then appears as sd2, and then install the OS on sd2), the installer
will create an EFI boot partition on the inner softraid, even though
that partition never will be used.

I guess what the installer does is "fdisk -igy -b 960" for any GPT
installation. If the destination block device is a softraid though, the
"-b 960" should be avoided. The user himself already took care of
ensuring that an EFI partition was provided on the parent block device.

Reproduction:

 * Flash install62 to a USB memory stick.
 * Boot the system from it
 * Go into (S)hell mode
 * Do "fdisk -igy -b 960 sd0"
 * Add a RAID partition e.g.
disklabel -E sd0
a a
(default offset)
(default size)
RAID(for RAID)
w
q
 * Do "bioctl -c C -r 8192 -l /dev/sd0a softraid0", type the password
 * Run "/install". When asked about device to install on, choose "sd2".
 * Complete the installation process

Bug:
The system will now have not only an EFI partition on sd0, but also one
on sd2.

This bug cannot be worked around except by modifying the install
script, as there is no suitable point at which the user could ctrl+Z
and then manually go into fdisk to remove the EFI partition and
reallocate that space to use.

This bug applies on OpenBSD 6.2 and I think I saw it in 6.0 or 6.1 also.

Thanks,
Tinker

Reply via email to