[14:21] <smoser> icee: hm... dns should work at that point.
[14:21] <smoser> i'm not aware of that bug.
[14:22] <smoser> but artful transitioned over to systemd-resolve and somet hings have been less than perfect.
[16:31] <icee> smoser: systemd-resolved is definitely not running, and /etc/resolv.conf is a symlink to the systemd-resolved thing that is not there
[16:31] <smoser> hm.
[16:31] <icee> the rescue shell opens and i'm able to open it and poke around.. and if i remove that symlink and add a 'nameserver 4.2.2.1' file to the cloud image it all works
[16:32] <smoser> rescue shell ?
[16:32] <smoser> hm.. i can try to reproduce.
[16:32] <icee> yah, on console it tells me to hit ctrl-d to continue to boot, or enter to get a shell
[16:32] <smoser> could you open a bug with what you're doing ?
[16:32] <icee> so i hit enter and poke around in the quasi-single-user-mode
[16:32] <icee> i opened a bug
[16:33] <smoser> link ?
[16:33] <icee> uhhh heh let me try and figure that out sec
[16:33] <icee> https://bugs.launchpad.net/cloud-init/+bug/1734167
[16:35] <icee> just added a command line with the entire command lines
[16:35] <icee> er
[16:35] <icee> comment with the entire command lines :P
[16:35] <icee> sorry i'm half awake :P
[16:36] <icee> so i guess it's of note that i'm bypassing initrd, too
[16:43] <icee> smoser: interesting.  it's also broken without my own kernel.  But i don't get the rescue shell to troubleshoot :P
[17:13] <smoser> icee: yeah. verified.
[17:13] <icee> smoser: cool, thanks :D
[17:13] <icee> i don't understand the rescue shell thing
[17:14] <icee> where exactly it comes from, etc... it's really, really nice :P
[17:37] <smoser> icee: i suspect you're missing something in kernel that is required and seeing the fallout that way.
[17:38] <smoser> but, dont know.
[17:38] <smoser> anyway, i put some more information there and pinged our systemd people.
[17:41] <icee> well, with correct dns in the image i get a login prompt and all works :P
[17:41] <icee> so there's something baffling/magical-- the default failure behavior is different :P
[17:43] <smoser> well i verified fail on artful and bionic
[17:43] <smoser> not tried xenial
[17:43] <smoser> i'm kind of holding out hope on that
[17:43] <smoser> but yeah, broken by default :-(
[17:44] <icee> xenial works
[17:45] <icee> i've been using it
[17:45] <icee> systemd makes me cranky.  i tried to jiggle the dependencies around myself to make this work and just couldn't in 20 minutes of fooling around
[17:46] <smoser> "systemd makes me cranky."
[17:46] <smoser> +21
[17:46] <smoser> (the 2 there was a typo, but it isnt really wrong)
[17:48] <icee> haha
[17:53] <icee> smoser: thanks for all your help with getting this properly reported and marshalled
[17:53] <smoser> thanks for reporting
[17:53] <smoser> curious why you are using artful
[17:54] <smoser> (perfectly fine to do that, just most people pick the lts and thus things like this dont get seen)
[18:45] <icee> smoser: so i'm doing automated testing of kernel functionality, and i'd like to do it with non-ancient userlands too
[18:46] <icee> not that 17.10 is really so bleeding edge, but..