=== balloons is now known as Guest90681
=== Guest90681 is now known as balloons
[17:49] <dobey> how do i figure out why adt-run is hanging when i try to run autopilot tests with it?
[18:36] <dobey> elopio: come back to the dark side. we miss you
[18:37] <elopio> dobey: heh, I miss you too :)
[18:37] <elopio> dobey: adt-run -d -s gives more information and lets you log into the testbed after it times out.
[18:38] <dobey> elopio: it doesn't get far enough for that i don't think
[18:38] <dobey> Services restarted successfully.
[18:39] <dobey> that's the last message printed, where it hangs
[18:39] <dobey> it never times out
[18:39] <dobey> that's with a vivid image
[18:39] <elopio> dobey: paste the output you get with -d.
[18:39] <dobey> i can't run with a utopic image, because utopic is eol :-/
[18:40] <elopio> if it doesn't time out, sounds like something bad in autopkgtest.
[18:40] <dobey> and i can't create a wily image. it seems to just hang
[18:41] <dobey> and i don't know how to create an ubuntu-rtm image
[18:42] <dobey> yeah, -d doesn't show anything additional there
[18:42] <dobey> top shows qemu is doing stuff, but no messages on console
[18:46] <dobey> :(
[18:49] <elopio> dobey: with the snappy ssh script, if you pass -b it prints the vm boot console.
[18:49] <elopio> not sure what test bed you are using.
[18:49] <dobey> i'm trying to run the pay-ui click autopilot tests with adt-run
[18:50] <dobey> qemu has a serial console socket, but i'm not sure how to connect to it
[18:50] <dobey> but i can't seem to open it with tail or minicom
[18:50] <dobey> minicom: cannot open /tmp/adt-virt-qemu.9bhk7o7g/ttyS1: No such device or address
[18:50] <dobey> meh
[18:57] <elopio> dobey: for qemu, the option is --show-boot
[18:58] <dobey> option to what?
[19:03] <elopio> dobey: to redirect the vm boot messages to your terminal. adt-run ... --- qemu --show-boot ...
[19:04] <elopio> that should show more info. Not sure useful info, but more.
[19:07] <dobey> doesn't work here
[19:07] <dobey> adt-virt-qemu: error: unrecognized arguments: --show-boot
[19:08] <dobey> and -d there doesn't show naything
[19:10] <dobey> and this was working fine 6 months ago when i last ran this stuff :(
[19:12] <elopio> um, I am on vivid but have adt-run from wily. Maybe this is new.
[19:13] <elopio> but well, you shouldn't be surprised it stops working after 6 months without running it :)
[19:19] <dobey> why? i haven't changed distro versions. i haven't modified the image i'm testing on. the click package hasn't changed (though i'm trying to add new tests for making a bug fix now)
[19:19] <dobey> it shouldn't just stop working
[19:20] <dobey> the only thing this tells me is that i should just scrap all autopilot tests everywhere, because nobody actually uses these tools, so they're totally unreliable
[19:21] <dobey> and unreliable tests (and the tools needed to run them), are a waste of time
[19:26] <elopio> dobey: +1 to that. If nobody is running the tests daily, they are worthless.
[19:26] <elopio> however, I would suggest to fix them and start running them daily, of course.
[19:27] <dobey> elopio: there's no need to run the tests daily if changes aren't being made daily
[19:27] <dobey> i shouldn't have to run my own tests daily to ensure that the tools used to run those tests aren't breaking
[19:27] <elopio> dobey: you should run the tests against the development image, which changes daily.
[19:28] <dobey> nope
[19:29] <elopio> well, that's probably another problem. If the error is in qemu + image, nobody is running that combination either.
[19:30] <dobey> well, it's also not like i'm running the latest packages in development ubuntu all the time either
[19:30] <dobey> i'm doing this on an LTS host
[19:30] <dobey> i think there have been some security updates for qemu, but i wouldn't expect that to break things
[19:31] <elopio> there's also the chance that it has never worked.
[19:31] <elopio> but well, there are so many projects involved in this that things are delicate and break in unexpected ways.
[19:32] <dobey> well i know it worked because i ran the tests
[19:32] <dobey> and i refused to land the autopkgtest stuff in the click until i had it working reliably
[19:35] <elopio> I get an error creating the chroot as explained in the README ^_^
[19:35] <elopio> many broken things.
[19:36] <elopio> none of them your fault, I bet.
[19:36] <elopio> but if you don't run your tests daily, you won't find when somebody breaks your stuff, and you won't be able to make sure that they don't break it again.
[19:42] <dobey> i shouldn't have to manually run my tests daily
[19:45] <dobey> if the tests should be run manually, then it should be done via CI, where the image builds automatically run the autopkgtests configured for all the packages installed by default on that image
[19:47] <elopio> yes. By you I don't mean *just you*. Your team, somewhere, ideally automatically. But every day, that's the important part.
[19:57] <dobey> i have to disagree