On Tue, Nov 21, 2017 at 09:36:02AM -0500, David Kaylor wrote:
[ 2.9] Planning how to build this image
0: itags: +filename=/home/dkaylor/.cache/virt-builder/rhel-7.4.x86_64.2 +size=6442450944
+format=raw +template +xz
0: task : pxzcat
0: otags: +filename=test.img +size=6442450944 +format=raw
1: itags: +filename=test.img +size=6442450944 +format=raw
1: task : virt-resize
1: otags: +filename=test.img +size=8589934592 +format=raw
This plan is completely bogus. Why is it choosing to write to the
output file?
I think what's happened here is that when I rewrote the planner
weights
(
https://github.com/libguestfs/libguestfs/commit/a6a982f004ac0e861d1889219...)
it has exposed a latent bug.
It should be dropping any plan which involves running virt-resize on a
same-named file, since that is impossible. Probably adjusting the
weights caused that plan to become the best one, exposing the
(existing) bug.
I'll have to think about this one ...
Rich.
--
Richard Jones, Virtualization Group, Red Hat
http://people.redhat.com/~rjones
Read my programming and virtualization blog:
http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers. Over 100 libraries supported.
http://fedoraproject.org/wiki/MinGW