[13:42] <Laney> xnox: want to review https://code.launchpad.net/~laney/ubiquity/+git/ubiquity/+merge/387747 ?
[13:42] <Laney> tried to keep it small, change to apt update for ubuntu-drivers
[13:53] <realtime-neil> Just how obsolete is this? https://help.ubuntu.com/community/InstallCDCustomization
[13:54] <realtime-neil> On a scale from "I did this yesterday" to "The last person that did that retired a decade ago".
[13:54] <xnox> Laney:  looks ok. And it is backportable to focal branch too, right?
[13:54] <xnox> Laney:  i do hate the 'blah blah\'n blah blah'
[13:54] <Laney> the what?
[13:54] <Laney> multiline strings?
[13:55] <xnox> i normally go for "asdfsafsaf'asdfasdfs" or just "No, we did not find any OEM packages again.'
[13:55] <Laney> lol
[13:55] <Laney> well the second one makes a test fail and the thing ftbfs
[13:55] <Laney> and I don't understand the first :>
[13:56] <Laney> WAIT
[13:56] <xnox> hahahhahahhahahhaa
[13:56] <Laney> you're talking about the apostrophe
[13:56] <xnox> yes
[13:56] <xnox> i did inline comment, with context on the merge proposal
[13:57] <xnox> it's fine, if it builds, we should ship it.
[13:57] <Laney> I'll change it, it doesn't matter
[13:57] <Laney> thx
[13:57] <Laney> I'm not 100% confident in it because I don't have a device to test
[13:58] <Laney> but we can ask those guys to do it
[13:58] <Laney> just makig sure I didn't break kubuntu
[14:05] <xnox> Laney:  i can build you a focal iso without pool.
[14:05] <Laney> we should totally do a one off black run over ubiquity
[14:05] <xnox> Laney:  then your existing oem device should discover "over the air", no?
[14:06] <xnox> Laney:  i guess we somehow need oem-ubuntu-test-in-updates which is not on the iso.
[14:06] <xnox> maybe like forever in proposed, but then we need to boot with proposed enabled. sigh
[14:07] <Laney> it should be ok to test something which is just not in the pool I guess
[14:07] <Laney> can you try it pls
[14:07] <Laney> https://people.canonical.com/~laney/temp/
[14:34] <Laney> what
[14:34] <Laney> is console-setup/Keyboard/KeyboardNames.pl and why is it being created
[14:38] <xnox> d-i thing does that
[14:38] <xnox> always on clean, and recreates it at build
[14:38] <xnox> i don't know how to make it stop
[14:38] <xnox> we didn't used to care as we didn't track console-setup in git before
[14:45] <Laney> ok for some reason it's not there this time
[14:46] <Laney> /shrug
[19:28] <realtime-neil> xnox: that `iso=` trick isn't working for me --- I've got `ip=dhcp iso=http://_gateway:3003/ubuntu-20.04-live-server-amd64.iso` (where I'm serving it) and qemu is erroring out with "No init found. Try passing init= bootarg"
[19:49] <CarlFK> realtime-neil: what is your boot media and files?
[19:50] <CarlFK> this is my qemu line to test a usb stick: https://salsa.debian.org/debconf-video-team/ansible/-/blob/master/usbinst/test_thumb.sh
[19:51] <CarlFK> this script constructs URLs and downloads the files to put on the stick: https://salsa.debian.org/debconf-video-team/ansible/-/blob/master/usbinst/mk_usb_installer.sh
[20:06] <realtime-neil> CarlFK: this is what I'm trying to do: https://salsa.debian.org/snippets/461
[20:07] <realtime-neil> I don't need to test removable media so much as tell the early userspace to grab an iso from the _gateway and boot _that_.
[20:56] <realtime-neil> xnox: I'm trying to use the functionality you added here: https://launchpad.net/ubuntu/+source/casper/1.422  ; every time, it fails saying "runi-"
[20:57] <realtime-neil> "run-init: current directory on the same filesystem as the root: error 0", and five times at that
[20:58] <CarlFK> realtime-neil: you are booting the kernel and initrd from the iso image, and then trying to get it to read the iso over http.   do I have that right ?
[20:58] <realtime-neil> yes, that's correct
[20:58] <realtime-neil> I don't know another way of doing that.
[20:59] <realtime-neil> CarlFK: well, I'm sure there are others, but this is the way I want to do it.
[21:00] <CarlFK> the cd kernel/init may be expecting local media, not over the network
[21:00] <CarlFK> the "net install" images may be needed for over the net
[21:01] <realtime-neil> CarlFK: then I've clearly misunderstood the changes mentioned here: https://launchpad.net/ubuntu/+source/casper/1.422
[21:01] <CarlFK> but I don't know if/where they are for the new ubuntu installer
[21:02] <CarlFK> ok, never mind me.  Yeah, looks like what you are doing is spozed to work
[21:04] <realtime-neil> I'm looking for something inside that *.iso that reports the version of casper it's using, but I don't know where to look.
[21:04] <CarlFK> realtime-neil:   I think you  need to change iso=http... to url=http...
[21:05] <realtime-neil> CarlFK: tried both; they both fail the same way.
[21:06] <CarlFK> at least make it match the docs
[21:06] <realtime-neil> CarlFK: I've updated my kvm-test snippet accordingly
[21:06] <CarlFK> but the error does make me think it isn't getting 'there' yet
[21:07] <CarlFK> also run this in it's own term without the >null python3 -m http.server 3003
[21:07] <CarlFK> see if it gets a hit
[21:07] <realtime-neil> checking...
[21:08] <CarlFK> does this put the files in casper/ ?   7z x "${PWD}/ubuntu-20.04-live-server-amd64.iso" "casper/vmlinuz"
[21:10] <realtime-neil> ...yeah, nothing is hitting the server
[21:11] <realtime-neil> CarlFK: yes, all extracted paths are created relative to $PWD
[21:12] <realtime-neil> I'm looking at the [casper man page](http://manpages.ubuntu.com/manpages/focal/man7/casper.7.html) and wondering if the `netboot=` parameter (implicitly set by `url=`) has anything to do with this?
[22:24] <realtime-neil> Here's something... the `_gateway` notation is (apparently) okay for a `seedfrom=` (`s=`), but `url=` seems to hate it.
[22:25] <CarlFK> realtime-neil: oh... try using the IP
[22:25] <CarlFK> what is _gateway ?
[22:26] <realtime-neil> I saw it here:
[22:26] <realtime-neil> https://ubuntu.com/server/docs/install/autoinstall-quickstart
[22:26] <realtime-neil> but that's the first time I saw it.
[22:29] <realtime-neil> Ha! As long as you feed it the IP address of the machine running that `python -m http.server 3003 `, then it actually downloads the iso.
[22:30] <realtime-neil> Wow. This ubuntu-20.04-live-server-amd64.iso needs a kvm with more than 2048 MiB RAM
[22:43] <CarlFK> well.. at least you are making progress
[22:46] <realtime-neil> CarlFK: I know, right!
[22:47] <realtime-neil> Okay, cool, cool. CarlFK, I'm going to borrow heavily from those scripts you linked. That's exactly what I need to be doing for my remastered installation media image.
[22:49] <CarlFK> woo hoo!!!
[22:57] <mwhudson> realtime-neil: glad you got things working
[22:58] <mwhudson> slightly surprised 2G is required but yeah, netboot does require quite a lot of ram
[23:03] <realtime-neil> mwhudson: you and me both ... between the ~900MiB compressed iso and the ~225MiB uncompressed initrd, I guess it blew through 2GiB in hurry.
[23:04] <mwhudson> accessing the iso over iscsi would help
[23:04] <mwhudson> but well one thing at a time