[06:11] <lordievader> Good morning
[06:14] <cpaelzer> hi lordievader
[06:15] <lordievader> Hey cpaelzer , how are you doing?
[06:15] <cpaelzer> Friday'ish :-)
[08:10] <V7> Hey all
[08:10] <V7> Could anyone help with one situation
[08:13] <hateball> !ask
[08:16] <V7> So, a probem:
[08:17] <V7> When trying to boot an installer from a CD on old machine it gives: "Loading bootlogo...; graphics initialization failed; Error setting up gfxboot; boot:;
[08:17] <V7> When typing "help" in "boot:" prompt it gives the same
[08:18] <V7> Screen: https://i.imgur.com/tQhsnzL.jpg
[08:18] <V7> This is Ubuntu Server 16.04.4 i386
[08:18] <V7> Checksums the same
[08:20] <V7> So any command I'm trying to run it tells the same as you can see on screen.
[08:24] <V7> So, screen: https://i.imgur.com/LUBCmrT.jpg
[08:29] <V7> Anyone ?
[08:39] <ikonia> V7: it can't load the splash so is failing
[08:39] <ikonia> V7: most likley your video card won't support the framebuffer
[08:40] <ikonia> something like that
[09:32] <olivierbourdon38> Hello everyone. Anybody having experience / has solved issues generating a working initramdisk using initramfs (as opposed to dracut which works) when disk is GPT table and LVM used for all partitions ? the boot process fails as it does not find the root partition. grub.cfg seems perfectly similar (only UUIDs have changed) so must be something missing in the ramdisk content
[09:32] <olivierbourdon38> ubuntu xenial
[09:36] <olivierbourdon38> very minimal set of packages (270 in all)  https://gist.github.com/obourdon/0647f84321aac7cbe8db6a65cf78e100
[09:38] <tomreyn> olivierbourdon38: i think update-initramfs is the default, isnt it? i've got a xenail system which uses it and boots off a GPT disk. there's a mirror RAID, dmcrypt-LUKS, LVM on top, and no manual customizations to how the intramfs is compiled were needed.
[09:40] <tomreyn> /etc/kernel-img.conf has (the default of) do_initrd=yes
[09:40] <olivierbourdon38> tomreyn yes, indeed but I could never had it working as I am using DIB (https://github.com/openstack/diskimage-builder) to build the images. Only dracut seems to do the job for proper LVM booting so I am trying to figure out what to change in DIB to make it work
[09:42] <tomreyn> well you havbe a working (supposedly all needed modules contained), and a non working (supposedly missing modules) initramfs there. just compare what's in there and see what's missing in the initramfs one?
[09:45] <tomreyn> you usually want "MODULES=most" in /etc/initramfs-tools/initramfs.conf unless you have special needs.
[09:45] <tomreyn> (this is also the default)
[09:47] <tomreyn> olivierbourdon38: ^
[09:48] <olivierbourdon38> yes I did try to compare those and running update-initramfs -u -v shows that lvm plugin is called and embbed everything necessary in ramdisk but still no luck when booting
[09:49] <olivierbourdon38> I also changed the underlying volumes FS type from xfs to more "standard" ext4 3 and 2 without more success, LVM and labels seems to be the issue
[09:55] <tomreyn> based on what?
[09:55] <tomreyn> you could also just try to enter commands manually on grub console and thus see what seems to be wrong / missing
[10:04] <olivierbourdon38> based on the fact that rootfs can not be found whereas the same mechanism without LVM works perfectly
[10:12] <tomreyn> !who » olivierbourdon38
[10:12] <tomreyn> !who | olivierbourdon38
[10:13] <tomreyn> olivierbourdon38: so the lvm module is not included when you build with update-initrfamfs?
[10:42] <olivierbourdon38> tomreyn lvm module seems to be there but boot process does not seem to be able to retrieve rootfs by label as far as I can tell
[10:43] <tomreyn> hmm, without a lot more details anylzing this will not be possible. but i trust in you working it out. ;-)
[10:54] <jamespage> coreycb: contining work on py3 migration for neutron/networking pkgs - going OK so far if a little fiddly
[10:56] <olivierbourdon38> tomreyn I understand but I was hoping that this kind of issue to have occured before and that I could grab some ideas on where to look at to try solve it
[11:00] <jamespage> coreycb: I'm introducing quite a few -common pkgs for shared config files
[11:00] <jamespage> hope that's inline with your work so far
[11:08] <tomreyn> olivierbourdon38: thanks for explaining your take on this. understandable. good luck in sorting it out.
[11:12] <olivierbourdon38> tomreyn I think I found a way to get better/easier comparison to solve this. I created a preseed file which partition the disk in the exact same way my DIB images are.
[11:13] <tomreyn> good, keep narrowing it down
[11:14] <olivierbourdon38> I was able to successfully boot over net and LVM were recongnized ;-), will now go on comparing this working pattern with my failing one. Thanks tomreyn because our discussion triggered this idea
[11:15] <tomreyn> glad i could help, if indirectly ;)
[12:16] <Palm_premium>  
[12:20] <tomreyn> Palm_premium: you just sent an empty line there, in case you wanted to send more?
[12:26] <jamespage> coreycb: urgh missed that python3-ostestr still uses 'python' for the actual tests
[12:29] <coreycb> jamespage: oops, hopefully it does support py3.
[12:29] <jamespage> coreycb: yeah but you have to set PYTHON=python3 in the env
[12:30] <Nafallo> heh. I don't think you're meant to deploy landscape server on ubuntu core, right? ending up with a bunch of missing dependencies I think ;-)
[12:30] <Nafallo> stuff like ip and openssl
[12:31] <coreycb> jamespage: it looks like it's doing that in override_dh_auto_test
[12:31] <jamespage> coreycb: yeah that's what I'm just working back through now
[12:33] <Nafallo> s/ubuntu core/ubuntu base/
[12:53] <jamespage> coreycb: well two fail with the switch
[12:54] <jamespage> nice
[12:54] <coreycb> jamespage: not bad i guess
[13:26] <jamespage> coreycb: https://review.openstack.org/575732
[13:26] <jamespage> I have at least three neutron-*'s with the same patching
[13:30] <coreycb> jamespage: +1
[13:56] <jamespage> coreycb: I've also cherry picked a ssl recursion fix for eventlet - will see if that helps some of my failing tests shortly
[13:57] <coreycb> jamespage: ack
[14:53] <coreycb> jamespage: i think we're pretty good for b2. i'm going to switch gears back to py3 and will pick up any fringe b2 packages along with that.
[14:53] <coreycb> jamespage: poking at backports as well