powersj | smoser: blackboxsw: rharper: dpb1: ec2 integration test merge: https://code.launchpad.net/~powersj/cloud-init/+git/cloud-init/+merge/335186 | 01:04 |
---|---|---|
powersj | smoser: blackboxsw: incorporated some of the comments from the other two merges I proposed. | 01:04 |
smoser | blackboxsw: yeah, its not a huge deal. its just a convenience to ust run the modules | 02:05 |
smoser | if you think we should ditch it, i wouldnt throw a fuss. | 02:05 |
=== cpaelzer_ is now known as cpaelzer | ||
blackboxsw | smoser: resubmitted with your change suggestion. I get that we should have it. adds granularity of running different parts of cloud-init (even though our typical runtime path uses the bigger 'cloud-init init' stage hammer). Validated behavior on the cmdline is exclusively the init-modules only and not datasource setup etc. also added unit tests | 05:51 |
blackboxsw | https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/335094 | 05:51 |
blackboxsw | powersj: will check tomorrow thx | 05:52 |
doda | hi folks, are the ssh-keys, hostname and /etc/hosts the only thing that cloud-init recreates/modifies on first-boot? | 11:26 |
blackboxsw | ./v | 15:37 |
=== shardy is now known as shardy_afk | ||
=== shardy_afk is now known as shardy | ||
blackboxsw | ok validated bionic ec2 17.1.60 release fresh install & upgrade path works without errors. Azure 'works as expected', still suffering from the hostname bounce https://bugs.launchpad.net/cloud-init/+bug/1722668. | 17:33 |
ubot5` | Launchpad bug 1722668 in cloud-init (Ubuntu) "Azure: bouncing of network device/publishing of hostname fails on artful" [Critical,Confirmed] | 17:33 |
smoser | blackboxsw: https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/335233 | 17:39 |
smoser | you can put that azure comment on the bug mp there and ec2 also | 17:39 |
blackboxsw | interesting from cloud-init analyze of azure bionic setup 44 seconds (15 seconds on apt) trying to see what else takes so long there during boot | 17:45 |
blackboxsw | will do smoser | 17:45 |
smoser | blackboxsw: did you run gce ? | 17:50 |
smoser | blackboxsw: if you ran azure on a smaller type, disk io is horrendous | 17:51 |
smoser | like floppy bad | 17:51 |
smoser | https://bugs.launchpad.net/ubuntu/+source/update-notifier/+bug/1426023 | 17:51 |
ubot5` | Launchpad bug 525674 in update-notifier (Ubuntu) "duplicate for #1426023 apt-check hangs, preventing login via SSH" [Medium,Fix released] | 17:51 |
smoser | i'm going to look at https://jenkins.ubuntu.com/server/view/cloud-init/job/cloud-init-copr-test-7/143/console | 17:53 |
blackboxsw | smoser: yeah ran small instance azure. was currently looking through gce UI for launching daily-ubuntu-1804-bionic-v20171213 | 17:53 |
* blackboxsw goes back to gcloud cmdline | 17:54 | |
smoser | blackboxsw: https://github.com/cloud-init/qa-scripts/blob/master/doc/launching.md | 17:54 |
smoser | yeah | 17:54 |
blackboxsw | thanks again, yeah was digging up that doc again | 17:54 |
smoser | hm.. | 17:57 |
smoser | https://stackoverflow.com/questions/26734777/yum-error-cannot-retrieve-metalink-for-repository-epel-please-verify-its-path | 17:57 |
smoser | powersj: you have any idea on that copr-test error above? | 18:00 |
smoser | i cannot reproduce locally | 18:00 |
powersj | smoser: I believe it is a connectivity issue to the mirror | 18:02 |
powersj | we seem to have very poor connection to any mirror for them | 18:03 |
powersj | smoser: I've determined that I need to add 'content_id=com.ubuntu.cloud:daily:aws' to my filter for simplestreams, anything else to point at aws? | 18:12 |
powersj | https://paste.ubuntu.com/26184412/ is what I have | 18:14 |
=== cj_ is now known as cj | ||
smoser | powersj: well 2 options | 18:32 |
smoser | a.) act like 'image-status' and just go right into the specific content_id url for aws. | 18:32 |
powersj | I was currently looking at sstream-query | 18:34 |
smoser | b.) your suggestiion is "safer" for sure, going in by the index. but officially querying aws you can search by end point. | 18:34 |
smoser | endpoint=https://ec2.us-west-1.amazonaws.com | 18:34 |
powersj | ah | 18:34 |
smoser | yeah, image-status and usquery live on top of sstream-query with just knowledge of filters and endpoints built in | 18:34 |
blackboxsw | ok gce and ec2 re-done w/ 17.2 package. azure to follow. | 18:36 |
blackboxsw | ->errand back in ~25 | 18:36 |
smoser | powersj: http://paste.ubuntu.com/26184503/ | 18:37 |
smoser | i used http and --no-verify, ubt its safe to use https and no-verify | 18:38 |
powersj | smoser: would you prefer i call out to the sstream-query command or continue try down the path of using the simplestreams library directly and replicate sstream-query? | 18:39 |
smoser | no library is fine | 18:42 |
smoser | preferable | 18:42 |
smoser | you have example already in what you did for nocloud | 18:42 |
smoser | just different filters | 18:42 |
blackboxsw | smoser: approved 17.2 release branch https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/335233 | 19:15 |
rharper | huh, seedfrom=<url> is trailing slash sensitive ; seedfrom=http://10.245.168.20:38747/ (works) seedfrom=http://10.245.168.20:38747 (fails) | 19:21 |
smoser | rharper: strict. | 19:36 |
smoser | i'd be willing to acept somethign that added a / onto the end of the host[:port] portion of the url | 19:36 |
smoser | but i'n not intersted in just appending a /user-data or /meta-data. | 19:37 |
smoser | as it is this works as expected: | 19:37 |
smoser | http://myhost/instance-abce- | 19:37 |
smoser | getting | 19:37 |
rharper | smoser: I was just surprised; I'm cleaning up curtin's launch and I noticed that we had duplicate // in a few places, and in fixing that, it dropped the trailing slash on the seed url | 19:37 |
smoser | instance-abce-user-data and instance-abce-meta-data | 19:37 |
blackboxsw | smoser: care if I tryout git git-publish tool on your 17.2 branch to publish to master? | 19:38 |
smoser | duplicate / are generally a bug , and hiding such things doesnt help. | 19:38 |
smoser | blackboxsw: ? | 19:38 |
smoser | what is that ? | 19:38 |
rharper | smoser: right | 19:38 |
smoser | oh. um... well that one *is* special in that it needs the tag to go too | 19:38 |
blackboxsw | smoser: sorry this one https://trello.com/c/Xz1btAu8/591-tool-to-merge-community-authored-branch-into-master-and-ammend-appropriate-author | 19:38 |
blackboxsw | I wanted to land your approved 17.2 branch into master | 19:39 |
blackboxsw | and test the full-run of the git-publish MP tool | 19:39 |
smoser | powersj: https://github.com/canonical-server/test-scripts/pull/2 | 19:45 |
smoser | and then | 19:45 |
smoser | https://github.com/canonical-server/jenkins-jobs/pull/21 | 19:45 |
powersj | smoser: thx will look after lunch | 19:45 |
smoser | i'm not sure that will actually fix our issue, but i'm flummixed as to why the test works on 6 and not 7 | 19:45 |
powersj | smoser: stole your FilterMirror class too for simplestreams | 19:45 |
smoser | blackboxsw: did that make sens e? | 19:47 |
smoser | that mp is "special" | 19:47 |
smoser | int aht you ahve to pull the tag | 19:47 |
smoser | blackboxsw: i'm going to just do that ... take it and push and ake a release. i'll write down the stuff i did | 19:50 |
blackboxsw | ohh I missed your comment before mine smoser. right can't land your branch because 17.2 tag associated with it | 19:50 |
blackboxsw | understood smoser, there are some powers branches I can run against after the cut is made | 19:51 |
smoser | blackboxsw: writing this down at | 19:56 |
smoser | https://hackmd.io/MYEwRg7AHKCsC0wCMYCm8AsWCG8wDYBmDTEZMEVMAJm0LCA= | 19:56 |
smoser | blackboxsw: do you know di di put 'mark-released' anywhere othe rthan my temp dir ? | 20:45 |
smoser | http://paste.ubuntu.com/26185072/ | 20:45 |
blackboxsw | looking, trying to understand what you asked | 20:46 |
blackboxsw | ohh that script? | 20:46 |
smoser | do you know if i put that script somewhere? | 20:46 |
blackboxsw | yeah I thought that as in qa-scripts | 20:46 |
smoser | other than in my temp dir here :) | 20:46 |
blackboxsw | checking | 20:46 |
blackboxsw | https://github.com/cloud-init/qa-scripts/blob/master/scripts/lp-bugs-released | 20:47 |
blackboxsw | I put it there I think | 20:47 |
smoser | yeah. | 20:47 |
blackboxsw | checking to see if it's the sae | 20:47 |
smoser | it needs exe added. | 20:47 |
blackboxsw | same | 20:47 |
smoser | i'll do that. | 20:47 |
blackboxsw | good deal | 20:47 |
powersj | rharper: thanks for the review - as a side note, on __exit__ we run the destroy() method of all the instances. For ec2 this means shutting down the instances and deregistering any custom images. The deregister == delete. | 20:56 |
powersj | This includes any image/instance used as a "base" image for customizing | 20:56 |
rharper | powersj: does __exit__ get called on exception or control-c on nose ? | 20:56 |
rharper | powersj: generally just want to make sure that no matter what (save crashed host) we'd remove/purge those keys; is there a way to autodelete vpcs or other amazon resources ? | 20:57 |
powersj | In my experience it does for exceptions, I have had to clean up 1-2 SSH keys, but I think that's from me doing debugging | 20:57 |
rharper | k | 20:57 |
powersj | we tag all our resources for testing as well, so I'd like to have a clean up script that purges anything with 'cloud-init-testing' | 20:58 |
blackboxsw | +1 | 20:58 |
powersj | but would like to avoid having to do that in the first place :) | 20:58 |
powersj | blackboxsw: do we need CI to be testing the cloud-init CLI? | 20:59 |
powersj | at least run the commands? | 20:59 |
blackboxsw | it's not a bad idea for a full integration test powersj. unit tests only have mocked coverage at the moment | 21:00 |
blackboxsw | very shallow | 21:00 |
rharper | +1 for cli unittests; shouldn't be too bad since the cloudinit cli is pretty sparse | 21:01 |
blackboxsw | could easily run through most of the CLI advertised subcommands /params as a sanity check since most of them don't require any restart ec. | 21:01 |
* rharper has backlog for curtin | 21:01 | |
blackboxsw | etc. | 21:01 |
powersj | smoser: merges approved | 21:12 |
powersj | smoser: Is 'https://ec2.%s.amazonaws.com' % region_name that safe for all regions around the world? | 21:13 |
powersj | blackboxsw: ^ do you know? | 21:13 |
blackboxsw | safe.... umm, you mean are they all a string? | 21:13 |
powersj | do they all follow that same URL pattern no matter the region or location in the world | 21:14 |
smoser | powersj: i think so yes. | 21:14 |
smoser | unless maybe a govcloud | 21:14 |
smoser | but for now that ignoerable | 21:14 |
powersj | k thx | 21:14 |
rharper | smoser: \o/ 17.2 | 21:20 |
smoser | all documented at | 21:20 |
smoser | https://hackmd.io/MYEwRg7AHKCsC0wCMYCm8AsWCG8wDYBmDTEZMEVMAJm0LCA= | 21:20 |
smoser | comments welcome | 21:20 |
smoser | (and we shoudl put that somehwer etoo) | 21:20 |
smoser | and /me has to run | 21:21 |
smoser | and i fixed lp-bugs-released to not let you feed it a bug number as the version | 21:21 |
smoser | (whoops) | 21:21 |
powersj | op me or update topic ;) | 21:22 |
smoser | later | 21:22 |
=== powersj changed the topic of #cloud-init to: Reviews: http://bit.ly/ci-reviews | Meeting minutes: https://goo.gl/mrHdaj | Next status meeting: Monday 1/08 16:00 UTC | cloud-init 17.2 released | ||
powersj | danke | 21:22 |
blackboxsw | nice! woot! | 21:24 |
rharper | powersj: add date to the 17.2 release topic | 21:28 |
rharper | or maybe current release ? | 21:28 |
powersj | Reviews: http://bit.ly/ci-reviews | Meeting minutes: https://goo.gl/mrHdaj | Next status meeting: Monday 1/08 16:00 UTC | cloud-init 17.2 released (Dec 14, 2017) | 21:28 |
powersj | like that? | 21:29 |
rharper | works | 21:29 |
=== powersj changed the topic of #cloud-init to: Reviews: http://bit.ly/ci-reviews | Meeting minutes: https://goo.gl/mrHdaj | Next status meeting: Monday 1/08 16:00 UTC | cloud-init 17.2 released (Dec 14, 2017) | ||
blackboxsw | hmm I need to publish meeting minutes | 21:53 |
ckonstanski | I have a few short weeks in which I can get some contributing done. School is out till January. Will be looking at 1728430 unless something else comes up. | 22:09 |
ckonstanski | Are the recent gcc pie and profile changes wreaking havoc on the gentoo build? | 22:10 |
ckonstanski | Probably should ask that in #gentoo-virtualization | 22:11 |
rharper | bug 1728430 | 22:11 |
ubot5` | bug 1728430 in cloud-init "make NTP tests work in gentoo" [Medium,In progress] https://launchpad.net/bugs/1728430 | 22:11 |
rharper | ckonstanski: sounds good | 22:12 |
blackboxsw | ok minutes from this week published to cloud-init.github.io had to tweak the publishing logic a bit to handle #actions | 22:20 |
smoser | ckonstanski: there are a lot of things in teh ntp area. robjo is working on things there too. those tests really can stand to be refactored. | 23:04 |
smoser | see mailing list for a message from Robert. | 23:05 |
ckonstanski | ok | 23:08 |
ckonstanski | This? https://lists.launchpad.net/cloud-init/msg00111.html | 23:14 |
robjo | ckonstanski: yes, note there is also a bug https://bugs.launchpad.net/cloud-init/+bug/1731619 and a merge request: https://code.launchpad.net/~rjschwei/cloud-init/+git/cloud-init/+merge/334992 | 23:18 |
ubot5` | Launchpad bug 1731619 in cloud-init "Support chrony as a client for ntp" [Medium,Confirmed] | 23:18 |
ckonstanski | Will get caught up this evening. Thanks for the info. Seems I need to coordinate with Robert or work on something else. Maybe the latter is a better option until his refactor is done? | 23:19 |
ckonstanski | Somewhere I have a TODO list... | 23:20 |
ckonstanski | "gentoo packager support" and "support lxc tests on gentoo" | 23:20 |
powersj | something seems up with pylint and how it finds members | 23:55 |
powersj | now that I added simplestreams to another file pylint complains, even though simplestreams has been imported in another file for quite a while :\ | 23:55 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!