Hi Adam, hi Schlomo, for me it works with the package that has been put into proposed on the evening of May 11th 2016. Sorry for the late reply. On Wednesday, 11. Mai 2016, 16:42:12 Adam Conrad wrote: > Please help us by testing this new package. See > https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to > enable and use -proposed. Your feedback will aid us getting this update > out to other Ubuntu users. https://wiki.ubuntu.com/Testing/EnableProposed: [...] | | Installation testing using -proposed |---------------------------------------------------------- | | Sometimes you may be asked to test a netboot installer image from trusty- | proposed. The images may be found here (replace "i386" with your | architecture as necessary): | | http://archive.ubuntu.com/ubuntu/dists/trusty-proposed/main/installer-i386/ | | In order to install successfully from these images, you will normally need | to tell the installer to fetch its own components from -proposed as well, | which is not the default. To do this, add the following boot parameter: | | apt-setup/proposed=true | | You must make sure that the mirror you are installing from contains packages | from -proposed. All official mirrors will do so; if you operate your own | mirror or use a site-local mirror, it may need to be modified to pull from | -proposed. Make sure that you do not simply use a loopback mount of a CD or | DVD image as an installation source; this is one of the cases where such | mirrors will not work, since they do not contain the updated kernel packages | required by the new installer. | | [...] Tests I've done: 1. First I compared the versions between archive.ubuntu.com and our mirror, to make sure the Cronjob running at night took over all the changes, and the upstream mirror also did. 2. In the second step I added the "proposed" option to our PXE-Environment, to the preseed-File, and to our automation script creating and updating the directory structure for PXE (including the initrd with the) and the Customized Netboot-ISO-Image. 3. Then I removed the "deb" workaround from our preseed-File, and called the automation script to put our new preeseed-File into initrd. It is now containing: - the apt-setup/proposed=true - and the local0/repository *without* the manually prepended "deb" 4. Now I installed the testing machine with proposed. The installation was successfully done, this failed before (with 1.158ubuntu2). All selected packages were installed, the proposed-Repository was also included after installation and "debconf-get-selections --installer" shows the local0/repository without the prepended deb. This looks fine, now. 5. Then I added the Workarround "deb" again, and called the automation script to put our new preeseed-File into initrd. It is now containing: - the apt-setup/proposed=true - and the local0/repository *with* the manually prepended "deb" Workarround. 6. Now I installed the testing machine with proposed, again. Installation was also successful. All selected packages were installed, the proposed-Repository was also included after installation and "debconf-get-selections --installer" shows the local0/repository without the prepended deb. Also still worked. 7. Now, thuesday (12th) was over, since I of course had some other work this day. ;-) On Friday morning i've read Colins post. To be sure I also added the apt-setup/proposed=true to the kernel command line, and repeated the steps above (within the next past days). The results were the same. We are not distributing self build kernels or something like this via local0/repository at the moment. So the local0 isn't really used at this early point in the installation process, but the installation failed before (with 1.158ubuntu2) due to the syntax error in sources.list, and printed an error that no kernel could be installed. This doesn't happen any more with the packages from xenial-proposed. The not started ntpd in xenial-proposed amd64 (after fully completing the installation) i've ignored as unrelated to this. Looks a bit like something I only experienced on i386/i686 before. Yes, there were changes on other packages in proposed (also mentioned in Bug #1577596) that probably cause this, but I didn't really hunt this down, yet. ;-) Did all the tests above with amd64 only. Tested i386/i686 proposed only with the workarround "deb" in, till now. Since we're only changing something interpreted by Shell, I don't expect any surprises at this point. ;-) The fix works fine (as expected) from my side. Kind regards Lars -- man-da.de GmbH, AS8365 Phone: +49 6151 16-71027 Mornewegstraße 30 Fax: +49 6151 16-71198 D-64293 Darmstadt e-mail: