Description of problem: Composing DVD of rawhide fails during construction of install root filesystem. Version-Release number of selected component (if applicable): today's git HEAD 631f59be448e5891227291a3af9693e9eb41a503 strongly related to lorax-0.8-1 How reproducible: every time Steps to Reproduce: 1./usr/bin/pungi -c rawhide-fedora.ks \ --destdir=$DESTDIR --name Fedora --ver $VERSION --nosource 2. 3. Actual results: [snip] (606/606) [100%] installing specspo-16-4.noarch Non-fatal POSTTRANS scriptlet failure in rpm package kernel-3.0-0.rc7.git3.1.fc16.x86_64 writing .buildstamp file removing locales creating keymaps creating screenfont moving stubs F: Failed to install eject Traceback (most recent call last): File "/usr/bin/pungi", line 222, in <module> main() File "/usr/bin/pungi", line 124, in main mypungi.doBuildinstall() File "/usr/lib/python2.7/site-packages/pypungi/__init__.py", line 852, in doBuildinstall workdir=workdir, outputdir=outputdir) File "/usr/lib/python2.7/site-packages/pylorax/__init__.py", line 292, in run self.installtree.make_dracut_initramfs() File "/usr/lib/python2.7/site-packages/pylorax/installtree.py", line 557, in make_dracut_initramfs outfile, kernel.version]) File "/usr/lib64/python2.7/subprocess.py", line 511, in check_call raise CalledProcessError(retcode, cmd) subprocess.CalledProcessError: Command '['chroot', '/ext4/Fedora16/work/x86_64/yumroot', '/sbin/dracut', '--nomdadmconf', '--nolvmconf', '--xz', '--modules', 'base dmsquash-live', '/tmp/initramfs.img', '3.0-0.rc7.git3.1.fc16.x86_64']' returned non-zero exit status 1 Expected results: success Additional info:
dracut-011-1.noarch
Is the eject binary in the installtree?
Re-run fails today, bug #723538 [pungi]. Using the last DVD which did compose [July 14 ??], it looks like there is no 'eject', but the situation is complicated: ----- $ cd $DVD/isolinux $ file initrd.img initrd.img: XZ compressed data $ xz --decompress -c initrd.ing >/tmp/initrd xz: initrd.img: Compressed data is corrupt $ file /tmp/initrd /tmp/initrd: ASCII cpio archive (SVR4 with no CRC) $ cpio --list </tmp/initrd | grep eject 20486 blocks ## no 'eject' $ ----- So the first part of initrd.img is an XZ compressed output of a CPIO archive, and the CPIO archive contains no 'eject'. [Evidently initrd.img is a "complicated" file. Where are the utilities that know how to diagnose it?]
The new initrd is complicated. It's a dracut initramfs glued together with our own xz initrd. But what I meant is if the eject binary could be found in the yumroot directory in pungi's work directory. I guess you removed it in the meantime, so we just have to wait until the pungi bug gets fixed.
Do you still get this error? Looks like other people don't hit it.
pungi has been aborting earlier, I haven't gotten to 'eject' in a week. Today pyorbit cannot be found. (That's after fixing the is_beta vs. isfinal rewrite.) Before today, there was a conflict between netxen-firmware and linux-firmware over file /usr/lib/libphan.fw. (The string "netxen-firmware" appears nowhere except in /var/cache/pungi/ourtree/{*.sqlite, <<others>>}.)