| *** patrickeast has quit IRC | 00:02 | |
| *** patrickeast has joined #zuul | 00:05 | |
| *** jamielennox|away is now known as jamielennox | 00:16 | |
| *** saneax is now known as saneax-_-|AFK | 00:43 | |
| jlk | jeblair: hi there. I'm looking at v3 history, and the build descriptions stuff went away. That's not back somewhere else right? if there was github stuff to fiddle with this, I can just throw it out? | 01:14 |
|---|---|---|
| jlk | quick gerrit question, how can I submit a patch (set) that will target the feature/v3 stuff, but uses it's own topic since it'll be it's own set? | 01:19 |
| mordred | jlk: git review -t github | 01:32 |
| mordred | jlk: should work | 01:32 |
| mordred | and put it in the github topic | 01:32 |
| mordred | jlk: the branch has a "defaultbranch" setting in .gitreview | 01:32 |
| mordred | so you don't have to worry too much about the submit-to-feature/v3 part | 01:32 |
| jlk | kk | 01:32 |
| mordred | however, fwiw, if you wanted to do that and didn't have the defaultbranch setting in place | 01:33 |
| mordred | git review -t github feature/v3 would do it | 01:33 |
| mordred | (first unflagged argument is branch to submit to) | 01:33 |
| * mordred wordvomits on screen | 01:33 | |
| jlk | thanks! | 01:35 |
| jlk | oh man, so uh, this refactor of things, well, it's throwing me for a loop in zuul/scheduler.py. It used to iterate over a list of connections and set them up, but now... it doesn't? | 01:36 |
| jlk | I don't know how hte code works to have more than one connection | 01:36 |
| jlk | ah, there is a ConnectionRegistry. | 01:39 |
| dmsimard | jeblair: I think I've addressed most of the width-related problems, let me know if that looks better for you ? http://46.231.133.111/reports/index.html (make sure to force-reload/clear cache) | 02:36 |
| dmsimard | jeblair: and pagination can now be deactivated everywhere :) | 03:57 |
| dmsimard | good night o/ | 03:57 |
| *** saneax-_-|AFK is now known as saneax | 04:22 | |
| *** jasondotstar has quit IRC | 04:35 | |
| *** jasondotstar has joined #zuul | 04:48 | |
| *** abregman has joined #zuul | 06:10 | |
| *** Cibo_ has quit IRC | 06:16 | |
| *** Cibo_ has joined #zuul | 06:19 | |
| *** Cibo_ has quit IRC | 06:37 | |
| *** saneax is now known as saneax-_-|AFK | 06:53 | |
| *** saneax-_-|AFK is now known as saneax | 06:56 | |
| *** isaacb has joined #zuul | 07:43 | |
| *** bhavik1 has joined #zuul | 07:59 | |
| *** bhavik1 has quit IRC | 08:26 | |
| *** hashar has joined #zuul | 08:30 | |
| *** bhavik1 has joined #zuul | 08:50 | |
| *** gk-1wm-su has joined #zuul | 09:55 | |
| *** Cibo_ has joined #zuul | 10:04 | |
| *** isaacb has quit IRC | 10:06 | |
| *** gk-1wm-su has quit IRC | 10:09 | |
| jkt | uh-oh, it seems that projects which are not anonymously readable over https such as this one, https://paste.fedoraproject.org/paste/oo6LerfacAbpAjmHJiH1cF5M1UNdIGYhyRLivL9gydE= effectively result in zuul-merger not merging *at all* | 10:19 |
| *** Cibo_ has quit IRC | 10:32 | |
| *** isaacb has joined #zuul | 10:33 | |
| *** Cibo_ has joined #zuul | 10:36 | |
| *** isaacb has quit IRC | 11:02 | |
| *** hashar is now known as hasharLunch | 11:45 | |
| *** isaacb has joined #zuul | 11:57 | |
| *** Cibo_ has quit IRC | 12:10 | |
| *** fungi has quit IRC | 12:11 | |
| *** fungi has joined #zuul | 12:18 | |
| *** saneax is now known as saneax-_-|AFK | 12:39 | |
| mordred | jkt: there is a patch somewhere I saw about using ssh for cloning for all the things ... | 13:02 |
| mordred | jkt: https://review.openstack.org/#/c/200034/ is one | 13:06 |
| mordred | this does not help your zuul-merger case though | 13:06 |
| *** hasharLunch is now known as hashar | 13:22 | |
| *** Cibo_ has joined #zuul | 13:39 | |
| *** Cibo_ has quit IRC | 13:46 | |
| jkt | mordred: hehe, I just spent a few hours reinventing that particular wheel; the end result is quite similar :) | 13:54 |
| jkt | but anyway, the getInfoRefs() doesn't appear to be called from this context at all | 13:57 |
| jkt | but my josb are still being executed against an already-merged tip, see the log at https://paste.fedoraproject.org/paste/s5TThymVIk~FFoD3DNkUol5M1UNdIGYhyRLivL9gydE= | 13:57 |
| *** openstackgerrit has quit IRC | 14:03 | |
| *** isaacb has quit IRC | 14:03 | |
| * jkt loves the generic name "item" | 14:03 | |
| *** isaacb has joined #zuul | 14:26 | |
| jkt | right, so this is because new Gerrit sends *two* events instead of just one when a new patchset is updated, the refUpdated comes in a separate event apparently | 14:30 |
| jkt | or nope. aargh. coffee. | 14:34 |
| *** mattclay has quit IRC | 14:35 | |
| *** mattclay has joined #zuul | 14:35 | |
| jeblair | jlk: correct, build descriptions are gone (jenkinsism) | 14:40 |
| *** openstackgerrit has joined #zuul | 14:46 | |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Storyboard: Support invalid tasks in update script https://review.openstack.org/438704 | 14:46 |
| *** mattclay has quit IRC | 14:49 | |
| *** mattclay has joined #zuul | 14:51 | |
| pabelanger | morning | 14:58 |
| *** abregman has quit IRC | 15:05 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Remove job trees from pipelines https://review.openstack.org/435944 | 15:12 |
| *** Cibo_ has joined #zuul | 15:14 | |
| *** Cibo has joined #zuul | 15:41 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Plumb job variables through to ansible https://review.openstack.org/438653 | 15:43 |
| *** Cibo_ has quit IRC | 15:43 | |
| jeblair | pabelanger, SpamapS: i added testing and fixed a bug it caught ^; i think that's actually ready now :) | 15:44 |
| *** Cibo_ has joined #zuul | 15:46 | |
| *** Cibo has quit IRC | 15:48 | |
| *** isaacb has quit IRC | 15:50 | |
| *** Cibo_ has quit IRC | 15:51 | |
| *** Cibo_ has joined #zuul | 15:51 | |
| pabelanger | +2 | 15:57 |
| *** isaacb has joined #zuul | 16:19 | |
| notmorgan | pabelanger: i was told to ping you this week re pen testing/security/etc for zuul-y-things :) | 16:27 |
| notmorgan | pabelanger: figured i'd jump on that now(ish) so we can talk soon(ish) ^_^ | 16:27 |
| pabelanger | notmorgan: Yes, it was mentioned. | 16:28 |
| pabelanger | notmorgan: So, we have zuulv3-dev.o.o, which we are running jobs on. Today, people can propose patches to openstack-infra/zuul project and we'll run those playbooks on zuulv3. | 16:29 |
| pabelanger | notmorgan: For me, we should go though the steps to actually try to get data either onto or off the server some how. | 16:29 |
| pabelanger | today, we don't have any secrets, aside from a SSH private key | 16:29 |
| * notmorgan nods. | 16:29 | |
| notmorgan | pabelanger: i think we can have some fun with that. | 16:30 |
| pabelanger | for me, we should attempt to pen the server, document them some place, then patch feature/zuulv3 and update unit tests | 16:30 |
| pabelanger | then, repeat | 16:31 |
| pabelanger | we have talked about wrapping ansible-playbook into some sort of chroot too, but I still think this step is valid | 16:31 |
| pabelanger | We also have zuul specific action plugins we also use today, but we just need to test them more and fix them as we go | 16:32 |
| notmorgan | i agree across the board | 16:32 |
| pabelanger | So ya, I've pretty much accepted out ssh private key on zuulv3-dev.o.o is actually public right now | 16:33 |
| pabelanger | so maybe we should work to protect that more first | 16:34 |
| notmorgan | yeah. i thnk that is def a good plan | 16:34 |
| notmorgan | pabelanger: i think we can agree that exploiting the ssh-key can be excluded from other bits if we're looking at pen, in isolation of chroot-y things. | 16:37 |
| SpamapS | jeblair: hooray for test coverage :) | 16:38 |
| notmorgan | pabelanger: so agreeing the key is public is sufficient to say "that is the first item on a (i'm sure) list of concerns" | 16:38 |
| eggshell | SpamapS: o/ | 16:49 |
| *** Cibo_ has quit IRC | 16:55 | |
| openstackgerrit | Jan Kundrát proposed openstack-infra/zuul master: Prepare correct refspec on new Gerrit https://review.openstack.org/439057 | 16:56 |
| jeblair | notmorgan, pabelanger: that key becoming public is at least annoying enough that i'd rather it not happen, so if we're going to capture the flag, let's use something more harmless. e.g. /etc/passwd. :) | 16:58 |
| notmorgan | jeblair: exactly | 16:58 |
| notmorgan | jeblair: which is why i said we document it as a "known issue" and find something more interesting | 16:58 |
| jeblair | cool | 17:00 |
| notmorgan | jeblair: if it's a known flaw(ish), it's uninteresting as long as we track it and address it. once we've addressed it, we can see if we can snag it. | 17:00 |
| notmorgan | but probably in a more isolated environment with a dummy key | 17:00 |
| SpamapS | eggshell: howdy.. so let's talk about test_tags | 17:00 |
| jeblair | notmorgan: well, the potential flaw would be 'ability to fetch a file outside the job root'. any file on that system would be sufficient to prove it; the only *interesting* file is the key. but it's an actual live key, so i'd rather the specific key not be actually exposed. | 17:01 |
| SpamapS | I believe it's time we write a spec about launcher protections. | 17:02 |
| SpamapS | We need to define the problem really clearly, draw lines, and evaluate what exists. | 17:02 |
| pabelanger | Ya, I would say a file, owned by zuul, out side of the jobdir is something we aim for. /etc/zuul/zuul.conf for example | 17:02 |
| notmorgan | pabelanger: ++ | 17:03 |
| pabelanger | we likley don't want to leak that | 17:03 |
| pabelanger | /etc/passwd might be harder, since if we do get that, then isn't that a security issue in the OS? | 17:03 |
| jeblair | pabelanger: nah, /etc/passwd hasn't been a security issue for years | 17:03 |
| pabelanger | oh right, I am thinking of shadow | 17:04 |
| notmorgan | jeblair: haha i transposed shadow vs passwd too | 17:04 |
| notmorgan | was in the middle of typing "then we have root escalation issues too"... then re-read the target you set | 17:04 |
| notmorgan | pabelanger: ^ ;) | 17:04 |
| jeblair | but zuul.conf is fine too | 17:04 |
| pabelanger | okay, so before we do some of this, we need to restart nodepool.o.o and setup our nodepool-id logic | 17:05 |
| eggshell | SpamapS: I'm wondering if there's a way to feed a test some extra python in zuulv3 since it seems includes is gone. | 17:07 |
| notmorgan | pabelanger: sounds good. | 17:07 |
| jeblair | SpamapS: a spec to describe the container tech would be good. ultimately both the ansible protections and container tech need to be documented in zuul's docs, so that will go a long way toward that as well. | 17:07 |
| eggshell | SpamapS: gandelman mentioned it might be possible to accomplish something similar in the launcher. | 17:08 |
| SpamapS | jeblair: right, that's a good point. It would be good to describe the problem in general terms, and then describe both solutions, so that we can actually look at where we're intending to overlap, and where both are deficient (or what a user is giving up if they find a need to remove one or the other) | 17:09 |
| jeblair | SpamapS: ++ | 17:09 |
| SpamapS | jeblair: and ultimately being able to write those things down for users is adoption Nirvana. :) | 17:09 |
| SpamapS | eggshell: what file are you looking at porting forward? | 17:10 |
| eggshell | SpamapS: tests/fixtures/tags_custom_functions.py | 17:11 |
| SpamapS | eggshell: ah, that's removed in v3 IIRC | 17:12 |
| SpamapS | jeblair: ^^ correct? | 17:12 |
| SpamapS | Seems to me that we should just add the tags to the environment in zuul itself the way we add anything else in. | 17:13 |
| jeblair | SpamapS: yes | 17:13 |
| jeblair | SpamapS, eggshell: http://eavesdrop.openstack.org/irclogs/%23zuul/%23zuul.2017-02-06.log.html#t2017-02-06T20:55:22 | 17:13 |
| SpamapS | wow | 17:13 |
| SpamapS | jeblair: your IRC log search fu is... impressive | 17:13 |
| jeblair | (careful, i'm having 2 interleaved conversations there, so follow the adam_g thread) | 17:13 |
| eggshell | awesome, that answers my questions on how v3 is going to handle build tags. | 17:14 |
| jeblair | eggshell: and for context, it's mostly just metadata that we can pass around to analysis tools | 17:15 |
| jeblair | like "this job is tagged with neutron" so that people can go search for neutron related jobs | 17:16 |
| eggshell | makes sense. | 17:16 |
| SpamapS | eggshell: so you mostly need to port fixtures/layout-tags.yaml to the newer config format under fixtures/config/single-tenant ... | 17:18 |
| eggshell | SpamapS: that part is (mostly) done. | 17:18 |
| SpamapS | eggshell: sweet | 17:19 |
| eggshell | what ends up happening is tags don't get added, so the assertEqual fails since expected results != None | 17:19 |
| SpamapS | eggshell: that's a great place to be failing :) | 17:23 |
| *** isaacb has quit IRC | 17:23 | |
| SpamapS | eggshell: so yeah, now you just need to figure out how to check the ansible variables of a fake build instead of the envvars :) | 17:23 |
| eggshell | ah okay, there's the unknown unknown! | 17:24 |
| SpamapS | and then of course, plumb that through | 17:25 |
| * SpamapS notices we've been using more plumbing analogies lately | 17:25 | |
| notmorgan | SpamapS: as long as you haven't needed the rotorooter .... :P | 17:26 |
| eggshell | SpamapS: re: check the ansible variables of a fake build is that something that currently exists in v3 or is a desirable? | 17:28 |
| SpamapS | eggshell: well we definitely add ansible variables to jobs that run. I just don't know if fake build records them in a convenient place like it does environment variables. | 17:29 |
| SpamapS | eggshell: see zuul/launcher/client.py | 17:30 |
| jeblair | SpamapS, eggshell: FakeBuild.parameters['vars'] should have it; might be nice to assign that to FakeBuild.variables or something for convenience | 17:30 |
| eggshell | SpamapS jeblair: thanks for the help! | 17:33 |
| eggshell | should be able to get going now | 17:33 |
| *** hashar has quit IRC | 18:06 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool master: Fix logic error with nodepool-id https://review.openstack.org/439090 | 18:15 |
| openstackgerrit | Merged openstack-infra/nodepool master: Fix logic error with nodepool-id https://review.openstack.org/439090 | 18:54 |
| *** harlowja has quit IRC | 20:15 | |
| *** harlowja has joined #zuul | 20:16 | |
| dmsimard | rbergeron, jeblair: so a question I asked yesterday remained unanswered -- should we draft a spec for Zuulv3 around the needs for a reporting interface ? | 20:21 |
| dmsimard | It's probably okay if we just end up describing ara if that's what we're looking for but we could end up figuring out what we're missing or how we want to implement it | 20:22 |
| pabelanger | depends what you are reporting. Openstack health would be an interface today | 20:23 |
| rbergeron | dmsimard: i think the answer was sort of yes, except it's probably more like 3.1, and not in the zuul v3 storyboard, since that's literally "get the thing up for openstack" | 20:23 |
| rbergeron | but 3.1 -- I don' tthink there's a board yet - but i think something like that, or saying "hey, a whole ecosystem of magical things can hook up to and show zuul stuff" is probably reasonable | 20:24 |
| rbergeron | or i could be crazypants | 20:24 |
| rbergeron | this happens :) | 20:24 |
| dmsimard | pabelanger: right, so this is a bit what I'm asking | 20:24 |
| dmsimard | there are opportunities to use ara in an openstack-health/centralized kind of way | 20:24 |
| dmsimard | or in one-off report for single jobs | 20:24 |
| dmsimard | It could be toggled per job, say, "reporting: yes" as one of the job parameters or something | 20:25 |
| dmsimard | for example it might not be relevant to have reporting for pep8 things | 20:25 |
| dmsimard | a bit like stackviz vs openstack-health | 20:26 |
| * rbergeron observes the whole internet freak out over aws-east-1 being down | 20:26 | |
| rbergeron | oh, maybe it's s3 | 20:28 |
| rbergeron | something's on fire | 20:28 |
| clarkb | ya its s3 | 20:28 |
| pabelanger | dmsimard: rbergeron: I've had some interest in resultsdb for a while: https://taskotron.fedoraproject.org/resultsdb/results. But haven't done a deep dive yet | 20:29 |
| clarkb | rbergeron: the neat thing is it doesn't affect us so we get to keep working as if nothing was wrong | 20:29 |
| rbergeron | serverless will save them all :) | 20:29 |
| rbergeron | clarkb: exactly. but if you take a break to look at the internet you can see ppl posting many pictures of tire fires and having good giggles | 20:29 |
| dmsimard | pabelanger: unrelated but it's pretty cool to see the actual fedora CI https://openqa.fedoraproject.org/tests/58262 | 20:30 |
| clarkb | rbergeron: its good for us because for all of them they can't even use the internet in their spare time | 20:30 |
| rbergeron | clarkb :) | 20:30 |
| clarkb | rbergeron: we don't hav eto worry about boredom | 20:30 |
| rbergeron | clarkb: no kidding. i got asked by one of the weird ... gate transfer things at heathrow - - where i went from BA to American - as part of my "so what do you do, why are you here, where do you live" thing | 20:31 |
| rbergeron | "what do you do in your spare time at home?" | 20:31 |
| rbergeron | and i was like | 20:31 |
| rbergeron | .... i'm sorry, my what? lol funny | 20:31 |
| clarkb | rbergeron: I hear yo uwatch competitive minecraft | 20:31 |
| clarkb | I think that might count | 20:31 |
| rbergeron | clarkb: only when watching my kids do it in person at minecon, otherwise, not so much | 20:32 |
| *** hashar has joined #zuul | 20:37 | |
| tflink | pabelanger: let us know if you have questions or if we can help with resultsdb | 20:49 |
| *** adamw has joined #zuul | 20:49 | |
| tflink | dmsimard: acutal fedora CI? that hurts a bit :-P | 20:50 |
| pabelanger | okay, zuul is leaving +1 again. I have restarted nl01.o.o | 20:50 |
| dmsimard | tflink: well, I mean, doing CI on screen captures and stuff -- not in a bad way | 20:51 |
| tflink | dmsimard: for context, I work on the non-actual CI in fedora | 20:51 |
| adamw | dmsimard: i work on the openQA stuff - it's a fun tool, but just part of our overall testing story, it's not the best tool for every job | 20:51 |
| adamw | dmsimard: also it's written by SUSE, so we don't get any of the credit :P | 20:52 |
| * tflink was also attempting a joke | 20:52 | |
| tflink | and if you have to explain that it's a joke ... you've failed | 20:52 |
| dmsimard | <3 fedora | 20:52 |
| clarkb | its what tumbleweed uses to do their rolling releases. Seems to work well absed on lack of broken desktop | 20:52 |
| adamw | clarkb: yeah, they're doing neat stuff with it. | 20:52 |
| adamw | for us it started out as a short-term stopgap for automating release validation tests, it kinda turned out better than we expected, so now we kinda expect it to stick around as one test system among many, for running the kinds of tests it's good at | 20:53 |
| adamw | clarkb: btw, https://www.happyassassin.net/nightlies.html is my 'poor man's tumbleweed' - it's integrated with openQA and autocloud, images that have tests and pass all their tests show up green there. | 20:54 |
| pabelanger | clarkb: got time to review: https://review.openstack.org/#/c/438653/ ? plumbs job variables into ansible | 20:57 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Configure regional mirrors for our jobs https://review.openstack.org/439152 | 21:00 |
| *** hashar has quit IRC | 21:00 | |
| rbergeron | adamw!! <3 | 21:04 |
| rbergeron | and tflink too :) | 21:06 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Add generic tox job https://review.openstack.org/438281 | 21:06 |
| jeblair | pabelanger: hrm, nl01.o.o is not outputting errors anymore. i'm afraid we have lost debuggable state by restarting it. if it happens again, let's not restart it so we can look into it. | 21:06 |
| adamw | rbergeron: ahoy | 21:06 |
| * adamw is just being openqa batman | 21:06 | |
| pabelanger | jeblair: ack | 21:06 |
| adamw | :P | 21:06 |
| jeblair | pabelanger: 2017-02-28 21:09:35,275 DEBUG zuul.AnsibleJob: Ansible output: [ubuntu-xenial] 2017-02-28 21:09:35.141004 | # 127.0.0.1:22 SSH-2.0-OpenSSH_7.2p2 Ubuntu-4ubuntu2.1 | 21:09 |
| jeblair | pabelanger: that's pretty much continuously being output in the zuul logs | 21:09 |
| jeblair | pabelanger: can you look into that? | 21:10 |
| pabelanger | jeblair: sure | 21:10 |
| jeblair | i HUPd zuul3 to pick up the new config with recheck | 21:11 |
| jeblair | (though, hopefully that wasn't necessary) | 21:11 |
| mordred | jeblair: just the ssh banner line? | 21:12 |
| jeblair | mordred: yep | 21:12 |
| jeblair | assume it's related to keyscan | 21:12 |
| jeblair | or something weird about known_hosts | 21:12 |
| mordred | jeblair: yah | 21:13 |
| pabelanger | happens after running testr | 21:14 |
| jeblair | pabelanger: yeah, i see that too locally | 21:14 |
| rbergeron | adamw: ossum, im being ansible robyn | 21:15 |
| rbergeron | adamw: but nice to see you and tflink both ;) | 21:15 |
| adamw | haha | 21:16 |
| tflink | rbergeron: we're good at lurking :) | 21:16 |
| * adamw does all the lurks | 21:16 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Remove unused mergeChangeQueue method https://review.openstack.org/435950 | 21:17 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Use hostname in Nodepool requests https://review.openstack.org/435949 | 21:17 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Clarify Job/Build/BuildSet docstrings https://review.openstack.org/435948 | 21:17 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Add note about pipelines and dynamic layouts https://review.openstack.org/435947 | 21:17 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Remove inscrutable TODO https://review.openstack.org/435946 | 21:17 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Handle merge mode in repeated project configs https://review.openstack.org/435945 | 21:17 |
| jeblair | that triggered a node_failure on the last change | 21:18 |
| jeblair | 2017-02-28 21:18:21,255 DEBUG nodepool.NodeRequestHandler: Declining node request 100-0000000797 because it would exceed quota | 21:18 |
| jeblair | oops, i think we have an edge case in our algo | 21:18 |
| jeblair | er, nope, the algo implementation is wrong | 21:19 |
| jeblair | that should hit this point: | 21:19 |
| jeblair | # If request < quota and request > available nodes (due to current usage), begin satisfying the request and do not process further requests until satisfied | 21:19 |
| pabelanger | cool | 21:23 |
| pabelanger | yay for testing | 21:23 |
| pabelanger | jeblair: have we plumbed the timeout setting off the top of your head | 21:23 |
| jeblair | pabelanger: no | 21:23 |
| pabelanger | k | 21:23 |
| jeblair | pabelanger: i think it's in the job, just not passed to the launcher. so only work on that needs to happen in launcher client/server | 21:24 |
| jeblair | pabelanger, SpamapS: low-hanging-fruit ^ | 21:24 |
| pabelanger | sure, I can take a look | 21:24 |
| jeblair | pabelanger: well, i was suggesting that someone else could do it, unless you urgently need it | 21:25 |
| jeblair | Shrews: i have filed https://storyboard.openstack.org/#!/story/2000886 and added you as an assignee since i assume you will want to continue work on that when you return from your well-deserved break. :) | 21:31 |
| pabelanger | jeblair: depends if we want to start testing py27 and coverage jobs, we are right on the edge it looks like | 21:33 |
| pabelanger | jeblair: otherwise, I can make them non-voting for now | 21:33 |
| pabelanger | https://review.openstack.org/#/c/438281/ also create our base job, which is working | 21:33 |
| jeblair | pabelanger: there's a 60 minute default timeout currently; i think we can live with that for a bit. | 21:34 |
| pabelanger | jeblair: Hmm, odd. something is aborting the job then | 21:34 |
| pabelanger | let me figure out why | 21:34 |
| rbergeron | omg | 21:38 |
| rbergeron | jimw came to our ansible meetin :) yay | 21:38 |
| jeblair | mordred: can you review https://review.openstack.org/438653 please? | 21:40 |
| pabelanger | jeblair: oh, hehe. timeout is only 60 (seconds) not 3600 | 21:42 |
| jeblair | pabelanger: ah, the rest of the time must be in pre/post playbooks which do not yet have timeouts | 21:43 |
| jeblair | pabelanger: sounds like this is going to block you; why don't you go ahead and plumb the timeout through | 21:44 |
| pabelanger | jeblair: sure | 21:49 |
| mordred | jeblair: +A | 21:54 |
| jhesketh | Morning | 21:59 |
| mordred | morning jhesketh ! | 22:00 |
| mordred | jhesketh: I feel like you've had a much more tiring time since last I saw you than I have | 22:00 |
| jhesketh | Heh, probably. The jet lag was pretty bad coming home. Still not sleeping at right hours | 22:02 |
| pabelanger | Hmm | 22:09 |
| pabelanger | .tox/py27/bin/python -m testtools.run tests.unit.test_v3.TestAnsible.test_playbook no longer works for me, I fail to properly find ansible-playbook | 22:10 |
| pabelanger | but if I source activate, it will work | 22:10 |
| pabelanger | I'll dive into that tomorrow | 22:10 |
| jeblair | pabelanger: that probably never worked. it needs to be in the path. | 22:12 |
| jeblair | mordred: it would be neat if ttrun could handle that case too. i have no suggestions as to how. :| | 22:12 |
| mordred | jeblair: hrm. yeah - I mean, I betcha it could pretty easily add the bin of the venv to the path | 22:14 |
| jeblair | mordred: that'd probably do it | 22:14 |
| mordred | jeblair: how does this look? http://paste.openstack.org/show/600855/ | 22:23 |
| mordred | jeblair: maybe I should put ttrun somewhere that I could run tests on it ... | 22:29 |
| mordred | of course, then I'd have to _write_ tests | 22:30 |
| mordred | that patch happily passes all the tests that exist | 22:30 |
| rbergeron | happy little passing tests, testing alll the things | 22:45 |
| * rbergeron isnt sure if she's jack handey, bob ross, or happy fun ball | 22:46 | |
| *** hashar has joined #zuul | 23:00 | |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Plumb job variables through to ansible https://review.openstack.org/438653 | 23:02 |
| *** hashar has quit IRC | 23:04 | |
| SpamapS | rbergeron: As long as you're not Eeyore, we'll be fine. | 23:11 |
| jeblair | mordred: that patch looks very plausible! | 23:16 |
| mordred | rbergeron, jeblair, SpamapS: (and anyone else) post live: http://inaugust.com/posts/whats-coming-zuulv3.html email and tweet follow ups coming in just a jiff | 23:16 |
| mordred | jeblair: yay! | 23:16 |
| mordred | jeblair: I _think_ I could just update os.environ and then call subprocess.call and it would work - but explicitly passing it in seemed more explicit | 23:17 |
| *** saneax-_-|AFK is now known as saneax | 23:18 | |
| mordred | https://twitter.com/e_monty/status/836719502092087296 | 23:27 |
| mordred | rbergeron, SpamapS, pabelanger: ^^ for your retweeting pleasure | 23:28 |
| pabelanger | ++ | 23:31 |
| *** adamw has left #zuul | 23:34 | |
| SpamapS | mmmmmm bloggy | 23:44 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool master: Add destructor to SSHClient https://review.openstack.org/439243 | 23:48 |
| pabelanger | boo, there is no way to disable user input for the pause task | 23:49 |
| mordred | pabelanger: I mean - you say that ... | 23:52 |
| pabelanger | mordred: Ya, I am basically trying to do sleep 30 in a test for zuul, but fails because of Executing local code is prohibited. Pause will work, but makes our logging all funky, and CTRL commands will get intercepted by ansible | 23:54 |
| mordred | pabelanger: ah - yes. I agree with you | 23:54 |
| mordred | pabelanger: while pause should not _normally_ be a problem for us because there will be no tty | 23:54 |
| mordred | pabelanger: oh - can we remove the tty from our subprocess invocation? | 23:55 |
| pabelanger | ya, I'll have to see if we can do that | 23:55 |
| mordred | pabelanger: or maybe even just passing in stdin=None or something | 23:55 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!