Hi list,
I'm trying to create an automatic install configuration for a multiboot
host. It will be used on lab machines for testing LinuxCNC software
built for a dozen+ distros/release/architecture combinations running on
various motherboards and machine controller PCI interface cards. OSs
include non-EOL releases of Debian, Ubuntu, EL6 and Fedora;
architectures include i386, amd64 and arm (where available).
In the current scheme, the disk is pre-partitioned with a large, shared
/boot partition, and the remainder in an LVM VG where the installer
creates the root file system. This kickstart configuration is fairly
straightforward, but partman is a challenge.
Web searches revealed no examples of reusing a VG. After a day of
reading and experimenting with the partman-auto and partman-auto-lvm
shell scripts, I haven't found how to stop partman from erasing and
recreating the root file system LV's existing VG + PV, but instead reuse
the existing one.
Is it possible to reuse an existing VG from partman-auto-lvm, and I'm
missing something obvious? If not, could partman scriptlets be used?
Another scheme is to do away with LVM and use logical partitions. I'm
assuming partman-auto can reuse an existing partition (with method{ keep
}). The disadvantage is lack of a mechanism to keep track of which
distro/release/architecture combo should be installed on which
partition, since partitions are referred to by number instead of by
name. I'm confident this could be done in Anaconda, but can't say how
difficult compared to writing partman scriptlets for the LVM scenario.
Thanks-
John
--
To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/5213e601.6030...@zultron.com
