| openstackgerrit | Merged zuul/zuul master: Add log browsing to build page https://review.opendev.org/671906 | 00:10 |
|---|---|---|
| *** threestrands has joined #zuul | 00:10 | |
| openstackgerrit | Merged zuul/zuul master: Move artifacts to their own section https://review.opendev.org/672379 | 00:21 |
| openstackgerrit | Merged zuul/zuul master: Remember tab location on build page https://review.opendev.org/672836 | 00:34 |
| openstackgerrit | Merged zuul/zuul master: Use base 1 line number anchors in log view https://review.opendev.org/672837 | 00:54 |
| *** panda has quit IRC | 01:06 | |
| *** panda has joined #zuul | 01:08 | |
| openstackgerrit | Merged zuul/zuul master: Add severity filtering to logs https://review.opendev.org/672839 | 01:14 |
| *** igordc has quit IRC | 01:15 | |
| openstackgerrit | Merged zuul/zuul master: Colorize log severity https://review.opendev.org/673103 | 01:29 |
| openstackgerrit | Merged zuul/zuul master: Add raw links to log manifest https://review.opendev.org/673104 | 01:46 |
| *** saneax has quit IRC | 01:48 | |
| openstackgerrit | Merged zuul/zuul master: Rename view to logfile https://review.opendev.org/673105 | 02:03 |
| *** threestrands has quit IRC | 03:30 | |
| *** threestrands has joined #zuul | 03:30 | |
| *** threestrands has quit IRC | 03:31 | |
| *** saneax has joined #zuul | 03:53 | |
| *** jank has joined #zuul | 04:53 | |
| *** jamesmcarthur_ has quit IRC | 05:28 | |
| *** shachar has quit IRC | 06:05 | |
| *** shachar has joined #zuul | 06:05 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Add support for smart reconfigurations https://review.opendev.org/652114 | 06:49 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Add --check-config option to zuul scheduler https://review.opendev.org/542160 | 06:49 |
| *** yoctozepto has quit IRC | 07:00 | |
| *** tosky has joined #zuul | 07:24 | |
| *** jpena|off is now known as jpena | 07:24 | |
| *** themroc has joined #zuul | 07:26 | |
| *** jangutter has joined #zuul | 07:55 | |
| arxcruz | hey guys, is that possible to set a host-vars to several jobs as a variable? so i don't need to repeat the host-vars in all the jobs? It can't be by the parent since there are different parents for those jobs | 07:58 |
| arxcruz | something like host-vars: my-host-vars where my-host-vars have the dictionary | 07:59 |
| AJaeger | arxcruz: you can add vars to all templates like https://opendev.org/openstack/project-config/src/branch/master/zuul.d/projects.yaml#L65 | 08:02 |
| AJaeger | That might even apply to jobs in all pipelines, not sure about exact semantics - let's check docs... | 08:02 |
| AJaeger | so, we have https://zuul-ci.org/docs/zuul/user/config.html#attr-project.vars | 08:03 |
| arxcruz | AJaeger, yeah, here's the problem i have, we have some rhel-8 jobs running that require to set ansible_python_interpreter: /usr/bin/python3 | 08:03 |
| AJaeger | arxcruz: but not project.host-vars, so what you want is not possible today... | 08:04 |
| arxcruz | AJaeger, i was wondering if i could do something like the nodeset | 08:04 |
| arxcruz | that i have a - nodeset: | 08:05 |
| arxcruz | - name: my-nodeset | 08:05 |
| arxcruz | and all the settings | 08:05 |
| AJaeger | arxcruz: So, the quick ideas don't work - sorry, can't help further right now. Best ask later today. But let me check first what you read ;) | 08:05 |
| AJaeger | arxcruz: looking at the docs, that is not possible. Don't you have a common base job? | 08:07 |
| AJaeger | But better discuss with rest of team later today, I'm offline again now... | 08:07 |
| arxcruz | AJaeger, yes, i do, but as i said, it's shared by other jobs that run on python2 | 08:07 |
| arxcruz | AJaeger, sure, thanks | 08:07 |
| AJaeger | ah, fun ;) | 08:08 |
| *** yoctozepto has joined #zuul | 08:09 | |
| openstackgerrit | Matthieu Huin proposed zuul/zuul master: Zuul CLI: allow access via REST https://review.opendev.org/636315 | 08:46 |
| *** SotK has quit IRC | 09:06 | |
| *** SotK has joined #zuul | 09:08 | |
| *** AshBullock has joined #zuul | 09:32 | |
| *** hwangbo has quit IRC | 09:52 | |
| *** ianw has quit IRC | 09:55 | |
| *** ianw has joined #zuul | 09:59 | |
| *** AshBullock has quit IRC | 10:26 | |
| *** bhavikdbavishi has joined #zuul | 10:31 | |
| *** bjackman has joined #zuul | 10:39 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: WIP: Allow ensure-tox to upgrade tox version https://review.opendev.org/672760 | 10:44 |
| *** jank has quit IRC | 10:48 | |
| *** AshBullock has joined #zuul | 10:57 | |
| *** bhavikdbavishi has quit IRC | 11:01 | |
| *** jank has joined #zuul | 11:04 | |
| *** saneax has quit IRC | 11:30 | |
| *** saneax has joined #zuul | 11:34 | |
| *** jpena is now known as jpena|lunch | 11:35 | |
| *** AshBullock has quit IRC | 11:41 | |
| *** jamesmcarthur has joined #zuul | 11:50 | |
| *** jamesmcarthur has quit IRC | 12:03 | |
| *** jpena|lunch is now known as jpena | 12:42 | |
| *** jamesmcarthur has joined #zuul | 12:44 | |
| *** bjackman has quit IRC | 12:47 | |
| *** jamesmcarthur has quit IRC | 12:51 | |
| *** jamesmcarthur has joined #zuul | 13:01 | |
| *** jank has quit IRC | 13:01 | |
| *** jamesmcarthur has quit IRC | 13:02 | |
| *** jank has joined #zuul | 13:03 | |
| *** jamesmcarthur has joined #zuul | 13:03 | |
| fungi | arxcruz: if they're all defined in the same file, you could use a yaml anchor and reference it | 13:38 |
| fungi | not technically a zuul feature, it's a yaml feature | 13:39 |
| mordred | arxcruz: for instance: https://opendev.org/opendev/system-config/src/branch/master/.zuul.yaml#L54 and then https://opendev.org/opendev/system-config/src/branch/master/.zuul.yaml#L66 and https://opendev.org/opendev/system-config/src/branch/master/.zuul.yaml#L73 | 13:43 |
| arxcruz | mordred, fungi is this work also to host-vars right? | 14:01 |
| arxcruz | thanks! that's what i needed :D | 14:01 |
| fungi | yeah, it's just a general yaml feature to deduplicate data structures | 14:05 |
| fungi | since zuul relies on a standard yaml parser, zuul doesn't need to know you're doing that | 14:05 |
| arxcruz | fungi, but it requires to be in the same file correct? or zuul will parse all the files in a big yaml file first ? | 14:06 |
| *** jeliu_ has joined #zuul | 14:09 | |
| *** wxy-xiyuan has quit IRC | 14:10 | |
| fungi | has to be in the same file, yes | 14:12 |
| fungi | zuul doesn't concatenate files | 14:12 |
| *** jeliu_ has quit IRC | 14:13 | |
| *** jeliu_ has joined #zuul | 14:13 | |
| *** jank has quit IRC | 14:19 | |
| *** bjackman has joined #zuul | 14:22 | |
| *** michael-beaver has joined #zuul | 14:27 | |
| *** sanjayu_ has joined #zuul | 14:41 | |
| *** saneax has quit IRC | 14:41 | |
| *** bjackman has quit IRC | 14:48 | |
| *** bhavikdbavishi has joined #zuul | 14:52 | |
| *** jeliu_ has quit IRC | 14:54 | |
| *** bhavikdbavishi has quit IRC | 14:56 | |
| *** jeliu_ has joined #zuul | 15:00 | |
| tobiash | corvus, mordred: is there some important stuff to review atm? | 15:02 |
| tobiash | I was less active here in the last weeks due to operational challenges and vacation but plan to get more active again | 15:03 |
| corvus | tobiash: i think mhu's zuul_admin_web series is close to ready, and since it's half-merged already, it would be nice to get the whole thing in for a release | 15:04 |
| corvus | also we said we'd review Shrews' autohold stuff once that merges | 15:05 |
| mhu | corvus, the GUI part needs a lot of work yet though. I am on it but I am not a react specialist | 15:05 |
| mhu | but if you're fine with just the CLI part at this point in time, then by all means | 15:06 |
| Shrews | \o/ | 15:06 |
| corvus | mhu: yeah, the current stuff up for review is the authz configuration, right? with that in place, someone can at least use the rest api | 15:06 |
| mhu | corvus, yes | 15:06 |
| tobiash | cool, I'll look at it | 15:06 |
| mhu | also, I'll be on vacation in a week, so if I'm not around do not hesitate to take over | 15:07 |
| corvus | all the more reason to get it finished this week :) | 15:07 |
| tobiash | I left a comment on 636315 | 15:12 |
| tobiash | corvus: what do you think? | 15:13 |
| *** bhavikdbavishi has joined #zuul | 15:14 | |
| SpamapS | Is one of the things in the admin interface going to be a "retry" button for buildsets? I have gotten a few requests for that of late, but I don't remember seeing it in the spec. | 15:19 |
| mhu | spamaps, the spec didn't cover the GUI, but if you look at https://review.opendev.org/#/c/643536/15/web/src/pages/Buildsets.jsx I added an "enqueue" button | 15:20 |
| mhu | so yes | 15:20 |
| corvus | or perhaps the web trigger could handle that | 15:22 |
| SpamapS | Cool, just wondering. | 15:23 |
| corvus | tobiash: replied and +3 | 15:27 |
| corvus | mhu: you want to rebase the rest of that stack? | 15:27 |
| mhu | corvus, on it | 15:27 |
| *** chandankumar is now known as raukadah | 15:29 | |
| tobiash | :) | 15:31 |
| openstackgerrit | Matthieu Huin proposed zuul/zuul master: Add Authorization Rules configuration https://review.opendev.org/639855 | 15:32 |
| openstackgerrit | Matthieu Huin proposed zuul/zuul master: Web: plug the authorization engine https://review.opendev.org/640884 | 15:32 |
| openstackgerrit | Matthieu Huin proposed zuul/zuul master: Zuul Web: add /api/user/authorizations endpoint https://review.opendev.org/641099 | 15:32 |
| openstackgerrit | Matthieu Huin proposed zuul/zuul master: authentication config: add optional token_expiry https://review.opendev.org/642408 | 15:32 |
| corvus | clarkb, mordred, tristanC: i *think* opendev's javascript updated, but when i visit a page like http://zuul.opendev.org/t/zuul/build/0c8293045843425b89c14d1f49af52b0 i don't see the tabs at the top | 15:38 |
| corvus | what am i missing? | 15:38 |
| corvus | we should at least see the "Summary" tab even if there's no "Logs" tab; that makes it feel like it's an old version of the js | 15:40 |
| corvus | but isn't the artifacts list new? | 15:40 |
| tobiash | corvus: I saw that some post jobs failed | 15:41 |
| tobiash | corvus: https://review.opendev.org/673105 | 15:41 |
| corvus | tobiash: aha, thanks! | 15:41 |
| corvus | i will follow that trail | 15:41 |
| tobiash | I guess opendev-promote-javascript-content is supposed to do that? | 15:41 |
| corvus | yep | 15:42 |
| corvus | indeed, the artifacts change is the last time that succeeded | 15:43 |
| tobiash | wasn't there some deprecation around download artifact by name? | 15:43 |
| tobiash | that's where it fails: http://logs.openstack.org/05/673105/1/promote/opendev-promote-javascript-content/784f17a/job-output.txt.gz#_2019-07-30_02_04_22_902311 | 15:44 |
| corvus | yeah, but it's supposed to be backwards compat for the next 2 weeks | 15:44 |
| corvus | i guess that didn't work | 15:44 |
| *** hwangbo has joined #zuul | 15:45 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Fix typo in download-artifact https://review.opendev.org/673566 | 15:46 |
| corvus | tobiash: ^ | 15:46 |
| AJaeger | ah, an extra ")" | 15:47 |
| corvus | once that's done, i'll retrigger the latest promotes | 15:47 |
| mordred | stupid )'s | 15:48 |
| openstackgerrit | Merged zuul/zuul master: Spec: Add a Kubernetes Operator for Zuul https://review.opendev.org/659180 | 15:50 |
| tobiash | I was too slow... | 15:50 |
| openstackgerrit | Merged zuul/zuul-jobs master: Update testing section https://review.opendev.org/672820 | 15:51 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: WIP: Allow ensure-tox to upgrade tox version https://review.opendev.org/672760 | 15:54 |
| corvus | tobiash: do you have mirrors or proxies (ie, operating system or language packages, etc) in your setup? | 15:54 |
| tobiash | corvus: both | 15:55 |
| corvus | tobiash: then do you want to look over https://review.opendev.org/669948 ? | 15:55 |
| tobiash | (authenticated) Proxies for internet access and mirrors using artifactory | 15:56 |
| tobiash | corvus: I'll take a look when I'm at home | 15:56 |
| corvus | tobiash: thanks | 15:56 |
| openstackgerrit | Merged zuul/zuul-jobs master: Fix typo in download-artifact https://review.opendev.org/673566 | 16:01 |
| corvus | re-enqueuing 673105,1 for promote | 16:04 |
| AJaeger | corvus: want to re-enqueue 672820 for promote as well? | 16:08 |
| AJaeger | we can also wait for next merge for accurate docs... | 16:08 |
| corvus | AJaeger: i still had the command up so it was easy enough to do, thanks :) | 16:09 |
| corvus | the jobs for 105 ran as expected | 16:09 |
| *** panda has quit IRC | 16:09 | |
| corvus | and afs is updated | 16:10 |
| corvus | so we're just waiting on an opendev ansible/puppet run | 16:10 |
| *** panda has joined #zuul | 16:13 | |
| openstackgerrit | Merged zuul/zuul master: Zuul CLI: allow access via REST https://review.opendev.org/636315 | 16:18 |
| openstackgerrit | Jeff Liu proposed zuul/zuul-jobs master: Add auth config to kubelet user for buildset registries https://review.opendev.org/673351 | 16:19 |
| AJaeger | thanks, corvus | 16:22 |
| *** evgenyl has quit IRC | 16:24 | |
| *** evgenyl has joined #zuul | 16:27 | |
| *** igordc has joined #zuul | 16:33 | |
| *** evgenyl has quit IRC | 16:34 | |
| *** evgenyl has joined #zuul | 16:36 | |
| *** jpena is now known as jpena|off | 16:38 | |
| *** cixx has quit IRC | 16:39 | |
| *** cixx has joined #zuul | 16:40 | |
| *** sgw has quit IRC | 16:42 | |
| *** themroc has quit IRC | 16:52 | |
| *** bhavikdbavishi has quit IRC | 16:54 | |
| pabelanger | Greetings, I think I've asked this before, we have a new requirement for zuul to checkout a project (say https://github.com/ansible-network/sandbox) on disk as src/github.com/ansible_network/sandbox. I know a role has the ability to control the dest name, but I don't believe a project can? This is all related to the upcoming collections work that ansible is going to be pushing in 2.9 and a hyphen in project | 16:54 |
| pabelanger | name is no longer allowed. I am hoping to avoid renaming a complete org in git to support it | 16:55 |
| *** bhavikdbavishi has joined #zuul | 16:56 | |
| mordred | pabelanger: I would suggest that you're not the only person who is going to have an issue with a decision to start disallowing hypens | 16:57 |
| fungi | sounds like a good reason to convince them to allow hyphens | 16:57 |
| fungi | yeah, that | 16:57 |
| mordred | yeah | 16:57 |
| *** jeliu_ has quit IRC | 16:57 | |
| mordred | otherwise its going to make life VERY hard for people who want to use collections but want to install them via git | 16:57 |
| *** sshnaidm is now known as sshnaidm|afk | 16:58 | |
| pabelanger | from what I am being told, it is because python importer breaks (doesn't support it) | 16:58 |
| pabelanger | yes | 16:58 |
| mordred | pabelanger: anyway - I believe you'd want a pre-playbook to do a mv | 16:58 |
| pabelanger | very hard is what I am seeing now, I'm trying to avoid the need to move files around after zuul pushed them | 16:58 |
| pabelanger | mordred: yah, I do that today, but don't like it | 16:59 |
| fungi | probably we need a way for collections to do that remapping. it would be very confusing to need to refer to the on-disk path with something other than its actual repository name. so many places i expect we assume they match | 16:59 |
| pabelanger | I'd also symlink it, but something about testing tool (ansible-test) doesn't like a symlink :( | 16:59 |
| mordred | yup. also - I have roles in repos with -s in them | 16:59 |
| mordred | not just org names :) | 17:00 |
| mordred | like - literally all of my role repos | 17:00 |
| pabelanger | this is also going to be an issue for zuul directly, when it grows the ability to use a collection from executor | 17:00 |
| pabelanger | same | 17:00 |
| fungi | can collections "know" a thing by some symbolic name which is not its directory name, maybe? | 17:00 |
| fungi | maybe via some configurable translation table? | 17:00 |
| *** jamesmcarthur has quit IRC | 17:00 | |
| pabelanger | maybe, but so far there is resistance to using "-", even if a technical solution was found | 17:00 |
| mordred | pabelanger: is it possible to do something (slighly absurd) like "mazer install src/github.com/ansible-network/sandbox" and have that install the thing into a location with the transformations needed? | 17:01 |
| fungi | i guess good luck getting it adopted if they're basically telling users they're not going to support existing names and everyone has to rename stuff before collections will be generally usable | 17:01 |
| fungi | sounds like a cop-out to me | 17:01 |
| pabelanger | mordred: maybe, I can test | 17:02 |
| mordred | fungi: yeah - but ... other things need to be installed, so I can accept the difference between "installed artifact" and "checked out source repo" | 17:02 |
| fungi | right, i mean python packages support this concept | 17:03 |
| mordred | yeah | 17:03 |
| fungi | the distname doesn't have to be the file/directory/archive name | 17:03 |
| fungi | you just have a registry mapping them to one another somewhere, or embed some metadata, or both | 17:03 |
| fungi | it's kinda like saying my new web browser is awesome but it doesn't support domains with hyphens in them so everyone will need to rename their websites | 17:04 |
| fungi | (because the url parsing library i chose gets confused by hyphens, but also because i think hyphens in domain names are ugly and everyone should agree with me) | 17:05 |
| *** bhavikdbavishi has quit IRC | 17:07 | |
| *** bhavikdbavishi has joined #zuul | 17:07 | |
| *** bhavikdbavishi has quit IRC | 17:17 | |
| *** bhavikdbavishi has joined #zuul | 17:18 | |
| *** jamesmcarthur has joined #zuul | 17:26 | |
| corvus | yay! http://zuul.opendev.org/t/zuul/build/2bf9467ddfe74b039a03d81c7d074f01#logs works | 17:32 |
| clarkb | corvus: the raw links have an extra / in them (it will work as is, but would be a nice cleanup) | 17:33 |
| corvus | clarkb: ya | 17:33 |
| * clarkb can write that patch if you want | 17:33 | |
| clarkb | I'm betting that one is simple enough for me to not get wrong :) | 17:33 |
| corvus | clarkb: all yours :) | 17:34 |
| corvus | i need to afk for a bit anyway | 17:34 |
| *** bhavikdbavishi1 has joined #zuul | 17:36 | |
| *** bhavikdbavishi has quit IRC | 17:37 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 17:37 | |
| mordred | corvus: that's just so sexy | 17:39 |
| mordred | SpamapS: ^^ I think you're going to like that | 17:41 |
| *** bhavikdbavishi has quit IRC | 17:41 | |
| pabelanger | looking at http://zuul.opendev.org/t/zuul/build/2bf9467ddfe74b039a03d81c7d074f01#logs in chrome, doesn't show anything execpt summary | 17:42 |
| pabelanger | I do see expections in console | 17:42 |
| pabelanger | hmm, network error | 17:43 |
| pabelanger | so maybe myside | 17:43 |
| *** tosky has quit IRC | 17:43 | |
| *** bhavikdbavishi has joined #zuul | 17:43 | |
| AJaeger | corvus: that is great! | 17:43 |
| pabelanger | http://logs.openstack.org/99/641099/33/check/tox-pep8/2bf9467/job-output.json.gz does load directy | 17:44 |
| pabelanger | directly* | 17:44 |
| pabelanger | so something in js that is issue? | 17:44 |
| mordred | pabelanger: if you refresh the build link does it work? | 17:44 |
| mordred | (and what sort of network error?) | 17:45 |
| pabelanger | http://paste.openstack.org/show/755129/ | 17:45 |
| tristanC | corvus: well done :) | 17:45 |
| pabelanger | that is what I see | 17:45 |
| pabelanger | ctrl-f5 doesn't fix | 17:45 |
| mordred | pabelanger: (the logs tab won't show until the data has been loaded) | 17:46 |
| pabelanger | firefox works | 17:46 |
| pabelanger | some things related to chrome | 17:46 |
| mordred | hrm. it's like it's applying a CORS rule or something - what version of chrome? | 17:47 |
| mordred | pabelanger: ad blocker | 17:47 |
| mordred | or also https://love2dev.com/blog/what-the-heck-does-neterr_blocked_by_client-mean-and-how-can-you-fix-it/ | 17:47 |
| fungi | pabelanger: blocked by am extension maybe? | 17:47 |
| pabelanger | Google Chrome 74.0.3729.131 (Official Build) (64-bit) | 17:48 |
| pabelanger | fungi: checking | 17:48 |
| mordred | oh - which is also mentioning adblock | 17:48 |
| pabelanger | yah, incognito works | 17:48 |
| pabelanger | so like adblock | 17:49 |
| Shrews | pabelanger: you'll accept our zuul ads and like it! :) | 17:49 |
| mordred | the internet seems to think it's related to extension - so maybe the fetch of a thing ending in .gz is bothering it? | 17:49 |
| pabelanger | indeed! | 17:49 |
| pabelanger | privacy badger was issue | 17:49 |
| pabelanger | relaxed settings for logs.o.o | 17:50 |
| mordred | ah. privacy badger | 17:50 |
| pabelanger | which is odd | 17:50 |
| mordred | I wonder if there is any way to detect such a thing and display a warning | 17:50 |
| pabelanger | because premerge it worked | 17:50 |
| pabelanger | I suspect because both were on logs.o.o | 17:50 |
| Shrews | hrm, i use pb and don't have any problems | 17:50 |
| pabelanger | web and logs | 17:50 |
| mordred | pabelanger: pre-merge it was the same site | 17:50 |
| fungi | yeah, it's the cross-site request that does it | 17:50 |
| mordred | both from logs.o.o | 17:50 |
| pabelanger | yah, that worked fine | 17:50 |
| mordred | cross-site to a .txt.gz potentially | 17:50 |
| pabelanger | maybe | 17:50 |
| fungi | pb blocks it for me in ff too, fwiw | 17:51 |
| Shrews | mordred: possibility of an sdk release soon-ish to help track down the builder volume leak issue? | 17:51 |
| mordred | Shrews: https://review.opendev.org/673581 Release 0.32.0 of openstacksdk | 17:53 |
| mordred | Shrews: you mean like that? | 17:53 |
| Shrews | something like that, yeah | 17:53 |
| mordred | Shrews: sweet | 17:54 |
| tristanC | pabelanger: did it worked before for json.gz request, e.g. on failed build page, there should be a panel with the failed task output | 17:54 |
| fungi | i bet if it was linking to logs.opendev.org pb wouldn't care | 17:55 |
| *** fdegir has quit IRC | 17:55 | |
| fungi | but it's a cross-domain request and those seem to raise its ire | 17:55 |
| *** fdegir has joined #zuul | 17:56 | |
| tristanC | fungi: failed build page does a cross-domain request the job-output.json.gz from zuul.o.o to logs.o.o (to display hosts stat and failed task output) | 17:57 |
| pabelanger | http://paste.openstack.org/show/755132/ | 17:59 |
| pabelanger | is what I seen | 17:59 |
| pabelanger | nothing after Summary was rendered | 17:59 |
| mordred | pabelanger: does privacy badger show you anything about why it's blocking that? | 18:00 |
| fungi | yep, and eff privacy badger blocks that request unless i tell it to allow requests to logs.openSTACK.org from zuul.openDEV.org | 18:00 |
| mordred | I'm reading about posting DNT policy files to self-certify that we are not tracking people | 18:00 |
| openstackgerrit | Clark Boylan proposed zuul/zuul master: Cleanup extra /'s in manifest render view https://review.opendev.org/673605 | 18:01 |
| mordred | fungi: any more specific information about what it doesn't like about it? | 18:01 |
| pabelanger | mordred: no, nothing about why, just orange indicator that something was found | 18:01 |
| fungi | mordred: good luck with that. pretty sure the folks assessing the dnt policy submissions only get around to verifying them for *very* large/popular sites | 18:01 |
| fungi | i looked into it at length | 18:01 |
| mordred | boo | 18:01 |
| *** jeliu_ has joined #zuul | 18:02 | |
| mordred | well - this is unfortunate, as I imagine this will be an issue for a non-zero number of our users | 18:02 |
| pabelanger | will also be issue with swift I suspect | 18:03 |
| clarkb | fwiw FF with default do not track stuff enabled and ublock origin and my crazy home firewall did not block it | 18:03 |
| clarkb | pabelanger: with swift we'll drop the .gz suffix and have swift properly serve the content based on headers | 18:03 |
| tristanC | clarkb: same for me | 18:03 |
| pabelanger | clarkb: cool | 18:03 |
| mordred | clarkb: I don't know that we've determined it's the .gz vs. just the cross-domain | 18:03 |
| clarkb | mordred: ah | 18:03 |
| tristanC | clarkb: there needs to be a cors header though | 18:03 |
| clarkb | tristanC: ya that may fix the cross domain problem? | 18:04 |
| mordred | it's _something_ and those are both likely candidates - unfortunately privacy badger isn't telling us what rule is being triggered | 18:04 |
| mordred | a CORS header violation would show differently - and we'd all be seeing it | 18:05 |
| fungi | mordred: fwiw, if i allow cookies for logs.openstack.org when viewing zuul.opendev.org pages, that fixes it | 18:06 |
| fungi | is the logs.o.o site requesting/setting a cookie somehow? | 18:06 |
| tristanC | clarkb: this is the header needed from the logserver: https://review.opendev.org/627903 | 18:06 |
| clarkb | tristanC: isn't that set then? | 18:07 |
| clarkb | I mean the change is merged | 18:07 |
| clarkb | that is probably why we aren't seeing cors errors | 18:07 |
| mordred | fungi: hrm. I don't think we're _purposely_ doing cookies on logs.o.o | 18:07 |
| mordred | clarkb: yeah - we'll need to ensure we tell swift to set that header when we start using swift | 18:08 |
| tristanC | clarkb: this would somehow need to be set on the swift host | 18:08 |
| clarkb | tristanC: I think we may be able to set that via the swift api? | 18:08 |
| clarkb | https://www.swiftstack.com/blog/2013/04/02/using-cors-with-swift/ ya | 18:08 |
| clarkb | (note we aren't using swift yet) | 18:08 |
| fungi | for comparison, http://zuul.openstack.org/build/46add3e269ae4c33a2adfdc95ace58f9#logs works fine for me and pb doesn't claim to have spotted any requests to trackers (allowed or otherwise) | 18:09 |
| fungi | in that case, the requests are not cross-domain though, they're just to other sites in the same domain | 18:09 |
| mordred | fungi: I don't see any cookies in my browers' interaction with logs.o.o | 18:10 |
| fungi | yeah, nor do i | 18:10 |
| fungi | so the "allow cookies" setting in pb may not be as simple as its name would imply | 18:10 |
| mordred | ooh! | 18:12 |
| mordred | https://zuul.opendev.org/t/zuul/build/2bf9467ddfe74b039a03d81c7d074f01#logs does not work | 18:13 |
| mordred | because of ssl / no-ssl | 18:13 |
| fungi | mordred: https://zuul.opendev.org/t/zuul/build/2bf9467ddfe74b039a03d81c7d074f01 doesn't seem to have a logs tab at all | 18:14 |
| mordred | yeah - the logs tab won't load if it can't load the json content that drives it | 18:15 |
| fungi | aha, right | 18:15 |
| fungi | and yeah, doing https://zuul... to http://logs... will be a problem (the browser will generally block that, don't need some extension to get in the way) | 18:15 |
| clarkb | we can create a vhost for logs.opendev and LE it easy enough | 18:16 |
| clarkb | or just switch to swift | 18:16 |
| mordred | yeah. but then zuul.openstack.org would be an issue - also - can we fetch swift objects over https: | 18:17 |
| mordred | ? | 18:17 |
| mordred | (This is a fun rabbithole we've opened here) | 18:17 |
| clarkb | mordred: that will depend on the swift implementation re swift https | 18:18 |
| mordred | clarkb: I guess since the objects are part of the API though that it's highly unlikely any of our public clouds don't have https turned on | 18:18 |
| clarkb | ya | 18:18 |
| mordred | we'll still need to sort out whether the .gz is involved or whether this will be a problem with swift too | 18:19 |
| clarkb | with swift we'll drop the .gz's | 18:19 |
| mordred | what I'm saying is - if the .gz isn't the issue, and the issue is the cross domain ... | 18:19 |
| mordred | then swift will have the same issue | 18:19 |
| fungi | the easiest demonstration is to compare http://zuul.openstack.org/build/46add3e269ae4c33a2adfdc95ace58f9#logs and http://zuul.opendev.org/t/openstack/build/46add3e269ae4c33a2adfdc95ace58f9#logs | 18:20 |
| clarkb | mordred: ah yup | 18:20 |
| fungi | they are the same build but via different vhosts | 18:20 |
| fungi | the former works fine for me and pb reports no trackers, the latter it thinks the logs site is a tracker and blocks it by default (unless i allow it) | 18:20 |
| mordred | *awesome* | 18:21 |
| clarkb | so could be that while the CORS policy says cross domain is fine | 18:21 |
| clarkb | pb is not allowing it because thatcould be a nefarious third party | 18:21 |
| fungi | pb does the same thing when i go to http://graphana.opendev.org/ because it requests files from http://graphite.openstack.org/ (again cross-domain) | 18:21 |
| corvus | i just installed pb and am able to view the job-output.txt of both of those just fine | 18:21 |
| fungi | er, http://grafana.opendev.org/ to http://graphite.openstack.org/ | 18:21 |
| fungi | gah, other way around | 18:22 |
| mordred | fungi: privacy badger source says that if we post https://subdomain.example.com/.well-known/dnt-policy.txt | 18:23 |
| mordred | it won't block the interaction | 18:23 |
| corvus | can someone help me out with a reproducer? | 18:23 |
| mordred | fungi: it reproduces for fungi - I think he's the reproducer case | 18:23 |
| mordred | fungi: which link doesn't work for you? | 18:24 |
| mordred | also, fungi, what version of PB do you have installed? | 18:25 |
| corvus | version 2019.7.1.1 | 18:25 |
| corvus | is what i just installed | 18:25 |
| fungi | corvus: i'm on firefox 67.0.1 with eff privacy badger 2019.7.1.1 | 18:25 |
| corvus | i see that pabelanger was looking at job-output.json earlier -- i suspect there's a problem with direct rendering of that, so we may want to be clear about which files we're accessing | 18:26 |
| corvus | i'm looking at job-output.txt, and so far, i have not had any trouble | 18:26 |
| corvus | i'm on ff 68.0.1 | 18:27 |
| fungi | in general what i see is that if an opendev.org site requests content from an openstack.org site (such as the case with the zuul builds pages) or an openstack.org site requests content from an opendev.org site (such as graphs in grafana) then privacy badger sees the request as going to a tracker | 18:27 |
| mordred | corvus: I thought the original pabelanger issue was that he got no logs tab at all because he got the blocking | 18:27 |
| fungi | so i generally just slide the control for that "tracker" over to allow and forget about it | 18:28 |
| mordred | corvus: I concur though - I just installed PB and I am able to load the page fine | 18:28 |
| corvus | mordred: 17:44 < pabelanger> http://logs.openstack.org/99/641099/33/check/tox-pep8/2bf9467/job-output.json.gz does load directy | 18:28 |
| corvus | mordred, fungi: yeah, i don't know how to reproduce fungi's issue | 18:29 |
| fungi | mordred: corvus: does the pb icon show a yellow-background number like "1" or a green-background "0" on those pages? | 18:29 |
| corvus | green 0 | 18:29 |
| fungi | neat | 18:29 |
| mordred | green 0 for me too | 18:29 |
| mordred | 13:42:51<pabelanger>looking at http://zuul.opendev.org/t/zuul/build/2bf9467ddfe74b039a03d81c7d074f01#logs in chrome, doesn't show anything execpt summary | 18:29 |
| mordred | (sorry for the bad timestamp) | 18:29 |
| corvus | mordred: right, i'm just saying let's be careful about using job-output.json as a test | 18:30 |
| mordred | totally | 18:30 |
| corvus | i'm not saying that pabelanger didn't have a problem | 18:30 |
| fungi | for me http://zuul.openstack.org/build/46add3e269ae4c33a2adfdc95ace58f9#logs has a green "0" (no trackers found) but http://zuul.opendev.org/t/openstack/build/46add3e269ae4c33a2adfdc95ace58f9#logs has a yellow "1" (logs.openstack.org) | 18:30 |
| corvus | i'm saying, "hey folks, there may be 2 problems, let's be careful not to accidentally conflate them" | 18:30 |
| fungi | pabelanger: do you see the same? | 18:30 |
| mordred | corvus: incidentally - job-output.json IS giving me a problem - but it's a different issue, so I'll just put a pin in it and we can come back to it | 18:30 |
| corvus | fungi: pb apparently "learns" ? | 18:30 |
| fungi | yeah, i don't know what sort of learning algorithm it uses | 18:31 |
| mordred | http://paste.openstack.org/show/755133/ is what I get from http://zuul.opendev.org/t/openstack/build/46add3e269ae4c33a2adfdc95ace58f9/log/job-output.json | 18:31 |
| mordred | corvus: so maybe the fact that pabelanger and fungi both enabled it means we don't see the issue? | 18:33 |
| mordred | oh - or the opposite - fungi has been using it for a while so it has learned things about his usage | 18:34 |
| corvus | yeah... maybe something about some other site under openstack.org? | 18:34 |
| fungi | mordred: ahh, the dnt policy thing is different, okay. maybe that's a lot easier. i was thinking of the https://tosdr.org/ database which is how ddg privacy essentials works out the site's "privacy practices" for part of its "privacy grade" | 18:34 |
| mordred | fungi: can you still see the error? | 18:34 |
| mordred | fungi: if so, I'd like to try putting the well-known file in place and seeing if that clears it up for you | 18:35 |
| *** jamesmcarthur has quit IRC | 18:35 | |
| fungi | mordred: that seems to have solved it | 18:35 |
| mordred | I did not put the file in place | 18:35 |
| fungi | oh, you're right | 18:36 |
| fungi | it took a moment for the js to get far enough to request the file and then it switched form green 0 to yellow 1 | 18:36 |
| fungi | so yes, it is reprodicble for me on reload, just takes a moment to show up | 18:36 |
| pabelanger | fungi: yes, that is correct | 18:37 |
| mordred | fungi: ok. I'm going to put the file in place - let's see if it solves it | 18:37 |
| fungi | lmk when it's there and i'll do another forced reload | 18:37 |
| fungi | assuming you can confirm it's directly fetchable | 18:37 |
| fungi | you're installing it on the logs site, right? | 18:38 |
| mordred | http://logs.openstack.org/.well-known/dnt-policy.txt is in place | 18:38 |
| mordred | yes | 18:38 |
| mordred | fungi: any luck? | 18:39 |
| fungi | i confirmed i was able to load that file, but pb behavior is unchanged | 18:39 |
| fungi | we can also check the apache access log for any requests for that file | 18:40 |
| mordred | ok. so putting that file in place does not help. that's at least good to know | 18:40 |
| fungi | of course, i've requested it once already | 18:40 |
| corvus | mordred: when you're done, let's be sure to remove it so it doesn't confuse the issue | 18:40 |
| mordred | corvus: agree | 18:40 |
| mordred | fungi: I see three accesses | 18:40 |
| mordred | I loaded it once | 18:40 |
| mordred | you loaded it once | 18:40 |
| corvus | me | 18:40 |
| mordred | and corvus | 18:41 |
| mordred | so pb did not load it | 18:41 |
| mordred | I will remove it now | 18:41 |
| mordred | gone | 18:41 |
| mordred | corvus: ok - I can't load job-output.json in my browser in either domain - same error each time | 18:45 |
| mordred | line 29 of errors.js - it seems like error.request is undefined | 18:45 |
| corvus | mordred: yeah, i expect that's a different error | 18:45 |
| corvus | are we switching to that? | 18:45 |
| corvus | i can only handle debugging one thing at a time | 18:45 |
| mordred | corvus: oh - sorry - yes. I expect that's a different error. and we can keep a pin in it | 18:46 |
| corvus | mind you, i'm not sure what else to do with the PB thing | 18:47 |
| corvus | i've been browsing a bunch of openstack.org domains to try to get PB to "learn" to distrust it, but no joy so far | 18:47 |
| corvus | looking at the pb config pages, i don't see anything indicating why it may have decided a domain is a tracking domain | 18:49 |
| corvus | (i have accumulated 1073 tracking domains just by browsing openstack.org) | 18:50 |
| mordred | corvus: same here (out of ideas) and same here (can't see anything) | 18:50 |
| mordred | corvus: wow | 18:50 |
| fungi | it could also have to do with the identifier-looking hex strings in those requests | 18:50 |
| fungi | hard to say | 18:50 |
| mordred | corvus: go to openstack.org itself? | 18:50 |
| corvus | mordred: openstack.org, ask.openstack.org, wiki.openstack.org | 18:50 |
| mordred | and maybe log in to your foundation account? | 18:50 |
| corvus | oh, i haven't logged in | 18:50 |
| mordred | (trying to think of something that would set a cookie in openstack.org domain) | 18:51 |
| fungi | oh, great point | 18:51 |
| corvus | no joy | 18:52 |
| fungi | ff says i have an "openstack.org" cookie last used 35 minutes ago | 18:52 |
| fungi | so this might explain 1. why not all of us see this behavior and 2. why allowing cookies for logs.openstack.org gets the content loading for me | 18:53 |
| mordred | yeah - because you had the cookies thing | 18:53 |
| mordred | yeah | 18:53 |
| mordred | like - why is your ff sending a cookie to logs.o.o one wonders | 18:53 |
| fungi | now if only i could figure out how to get it to show me what's in that cookie | 18:54 |
| clarkb | fungi: that should be in the network debug panel of ff | 18:54 |
| fungi | also i think this may explain why pabelanger said switching to an incognito session solved it (you did say that didn't you pabelanger?) | 18:54 |
| mordred | I have a cookie set for logs.o.o | 18:56 |
| mordred | how do I have one of those? | 18:56 |
| corvus | fungi: do you have ff content blocking enabled? mine is set to strict | 18:56 |
| corvus | i wonder if that means that pb isn't seeing as much training material | 18:56 |
| fungi | corvus: i do, but it's possible at some point i allowed cookies for openstack.org to get something working there | 18:57 |
| fungi | the cookie in question is for .openstack.org and has a name of __cfduid | 18:57 |
| fungi | set back in may with a one-year expiration | 18:57 |
| fungi | and ff says it is indeed sending that to logs.o.o | 18:58 |
| clarkb | ya I think cookies are sent to subdomains and parent domains | 18:58 |
| mordred | fungi: that's a cloudflare cookie | 19:00 |
| fungi | yuck | 19:00 |
| fungi | baleeted | 19:01 |
| corvus | is the current hypothesis: openstack.org set a cloudflare cookie at some point in the past, PB saw that cookie go to logs.o.o and therefore learned that it was a "tracking domain"? | 19:01 |
| mordred | yeah | 19:01 |
| mordred | also - I thought we looked in to cookies for subdomains and determined they did not get sent to subdomains - when we were looking in to the zuul-preview stuff | 19:02 |
| clarkb | mordred: I think there are settings you set in the cookie for that | 19:02 |
| mordred | yes - this is accurate - I just re-learned that | 19:03 |
| clarkb | so if the cloudflare cookie doesn't do that then you lose | 19:03 |
| clarkb | (we might be able to work with jimmy to fix that) | 19:03 |
| mordred | yeah - and fungi said it was set for .openstack.org | 19:03 |
| fungi | or was until i baleeted it | 19:03 |
| corvus | new browsers ignore the leading dot, so i think that means that cookie should be sent to *.openstack.org and openstack.org | 19:05 |
| fungi | to takeaway is for opendev/osf folks to work out the openstack.org cookie. this is not a zuul problem so i think we've discussed it as much as needs be in here | 19:13 |
| fungi | can move on to the json file issue i guess | 19:13 |
| mordred | fungi, corvus: maybe we should put in a $something where the logs tab will go after $time or something $somehow - like with a little flag and a popup "don't see logs, you may have an issue with PB or browser blocking, please contact your admin" or something like that? (I don't think we get an actual exception in the javascript we can respond to) | 19:18 |
| corvus | mordred: that's a good question -- are we sure we don't get an error? (like, axios returns a i couldn't get this file message?) | 19:19 |
| corvus | cause if so, we can probably do a thing where we say "this build has a manifest, but i can't retrieve it, popup an error" | 19:19 |
| fungi | the js debugger shows a "blocked by client" error | 19:19 |
| fungi | so maybe that's doable? | 19:19 |
| corvus | and that error could say "unable to fetch $URL" which might be enough for the user to debug the issue | 19:19 |
| corvus | (that would let them triage "oh hey i think the logserver is down" as well as "hrm, maybe that orange privacy badger thing has something to do with this" | 19:20 |
| fungi | could even say "unable to fetch $URL, check your browser plugins and ad blockers" | 19:20 |
| fungi | or... whatever | 19:20 |
| fungi | (though also we're going to be unable to generally support https->http requests so need to remember that needs solving by adding https to more things) | 19:21 |
| fungi | ((or switching to things that already https, like swift)) | 19:21 |
| mordred | well - the js debugger shows that in the console - but that doesnt' mean an exception got tossed to the javascript code | 19:21 |
| mordred | fungi: yeah | 19:21 |
| corvus | mordred: but... something should, right? | 19:22 |
| corvus | i mean, somewhere, there's an http request that needs a response | 19:22 |
| mordred | corvus: yeah. I'd like to know what that response looks like I guess | 19:22 |
| corvus | too bad fungi removed the cookie | 19:23 |
| mordred | corvus: maybe we could manaully add logs.openstack.org to the PB list? | 19:23 |
| corvus | mordred: oh maybe | 19:23 |
| corvus | might could do that by exporting, modifying, importing data | 19:24 |
| fungi | i have a feeling i can reacquire said cookie easily | 19:24 |
| mordred | yeah | 19:24 |
| fungi | but not until after the infra meeting | 19:25 |
| mordred | corvus: I tried the modifying route. it did not work as well as I otherwise might have wanted | 19:30 |
| corvus | :( | 19:31 |
| *** bhavikdbavishi has quit IRC | 19:39 | |
| fungi | so... on another computer i actually have two .openstack.org cookies (the aforementioned __cfduid which is probably from cloudflare and also a __ga which looks to be google analytics) | 19:59 |
| corvus | fungi: is it broke there too? | 20:01 |
| fungi | testing now | 20:01 |
| fungi | yes, privacy badger claims to have detected logs.openstack.org as a tracking site on this computer's firefox as well | 20:02 |
| fungi | (when trying to view the zuul.opendev.org build details) | 20:03 |
| corvus | cool | 20:03 |
| corvus | so i can add some error handling for this, but i think we would have to actually merge and deploy it to test it with PB | 20:04 |
| corvus | since the logs.o.o hosted version of this won't hit the error | 20:04 |
| corvus | oh, i wonder if we could use zuul-preview | 20:04 |
| corvus | yes! | 20:10 |
| fungi | wow, a new use case for it! | 20:10 |
| corvus | fungi: can you visit http://site.1e8199ad98454be297311a2542315ed8.zuul.zuul-preview.opendev.org/build/3c841507c8c345f0a4c83dc8fb8b6d00 and tell me if the problem shows up? | 20:10 |
| clarkb | wow hax | 20:11 |
| corvus | that's just the most recent build from the old patch series | 20:11 |
| corvus | but if it works, then it means i can mock up a solution and we can test it pre-merge | 20:11 |
| fungi | "File Not Found" | 20:11 |
| corvus | what's the context for that? | 20:11 |
| fungi | maybe i got the url wrong | 20:11 |
| corvus | fwiw, that url works for me, including the log tree and rendering job-output.txt | 20:12 |
| corvus | oh ha | 20:12 |
| corvus | a reload gets 404 | 20:12 |
| corvus | so, we might be hitting a zuul-preview bug | 20:12 |
| corvus | oh no i know what it is | 20:12 |
| corvus | i deep linked, but we can't do that with preview sites | 20:12 |
| corvus | fungi: go to http://site.1e8199ad98454be297311a2542315ed8.zuul.zuul-preview.opendev.org/status | 20:13 |
| corvus | grr | 20:13 |
| corvus | fungi: go to http://site.1e8199ad98454be297311a2542315ed8.zuul.zuul-preview.opendev.org/ | 20:13 |
| corvus | fungi: then navigate to a buld using the builds tab | 20:13 |
| corvus | any build should do | 20:13 |
| corvus | (so click "builds" then click the first "success" link on the list) | 20:13 |
| fungi | privacy badger is currently blocking content from zuul.openstack.org there for me. will set it to allow | 20:14 |
| fungi | allowing cookies for zuul.openstack.org got the initial url working | 20:14 |
| corvus | (that sounds like the same error) | 20:14 |
| fungi | i concur | 20:15 |
| corvus | we have 3 sites involved now, so 2 places where we can hit that error | 20:15 |
| corvus | happily we don't really need to solve this one | 20:15 |
| fungi | and yeah, on a build it is flagging logs.openstack.org as a blocked tracker | 20:15 |
| corvus | fungi: w00t | 20:16 |
| corvus | i will go add some error handling based on just a 404 and we can see if that's the same | 20:16 |
| jeliu_ | ansible | 20:16 |
| jeliu_ | ^sorry accident | 20:16 |
| fungi | i will go to the grocery store, but shouldn't be gone long. pabelanger may also be able to test | 20:17 |
| pabelanger | same, GET https://zuul.openstack.org/api/info net::ERR_BLOCKED_BY_CLIENT for url above | 20:18 |
| corvus | other than the config-errors, i can't find a place where we report an error to a user; can anyone think of one? | 20:21 |
| *** themroc has joined #zuul | 20:24 | |
| corvus | found it | 20:24 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: js: correct action names https://review.opendev.org/673680 | 20:28 |
| corvus | i have learned something about the way our app is structured. :) | 20:29 |
| corvus | once the preview build is ready for that, i'll paste in the new url | 20:29 |
| corvus | i'm going to debug the json issue while we wait | 20:33 |
| *** pabelanger has quit IRC | 20:37 | |
| corvus | okay, the json issue is (this may not surprise anyone) that we're given in actual javascript object when we fetch that, rather than a string. | 20:38 |
| *** mhu has quit IRC | 20:38 | |
| *** sgw has joined #zuul | 20:39 | |
| mordred | corvus: hahahaha | 20:41 |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: use opendev image building system for zuul-operator test https://review.opendev.org/673020 | 20:46 |
| openstackgerrit | Jeff Liu proposed zuul/zuul-jobs master: Add auth config to kubelet user for buildset registries https://review.opendev.org/673351 | 20:48 |
| *** igordc has quit IRC | 20:50 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: js: don't transform json logfiles https://review.opendev.org/673688 | 20:54 |
| corvus | that should take care of that one | 20:54 |
| mordred | \o/ | 20:55 |
| corvus | fungi: can you try http://site.7ee9de4db7ca498ead5c80429d0b6830.zuul.zuul-preview.opendev.org/ ? | 21:37 |
| *** jeliu_ has quit IRC | 21:46 | |
| fungi | i am back and can definitely try that | 21:50 |
| fungi | my browser is really, really slow at the moment. please wait :/ | 21:51 |
| fungi | okay, so that initial page worked because i was allowing zuul.openstack.org cookies. retrying with them disabled | 21:52 |
| corvus | fungi: i expect you're going to need to allow from zuul.openstack.org | 21:54 |
| fungi | yeah, so with cookies to zuul.openstack.org blocked, http://site.7ee9de4db7ca498ead5c80429d0b6830.zuul.zuul-preview.opendev.org/ just spins forever at "Fetching info..." | 21:54 |
| fungi | i'll reallow those and then see about the builds page | 21:54 |
| corvus | i think we're only interested in what happens when you go to the builds page | 21:54 |
| fungi | fair | 21:55 |
| fungi | re-set it to block cookies for logs.openstack.org now | 21:56 |
| fungi | no error message... i just don't get a logs tab when i select a build's detail | 21:57 |
| corvus | fungi: what happens if you enable the network pane? | 21:58 |
| corvus | fungi: oh, also, try going to the developer console and clicking the "XHR" button | 21:59 |
| corvus | maybe mordred's idea of not receiving an error response is at play, in which case, i'd love some suggestion as to what that actually looks like | 22:00 |
| *** themroc has quit IRC | 22:02 | |
| *** jamesmcarthur has joined #zuul | 22:09 | |
| fungi | will check | 22:09 |
| fungi | aha, under web developer -> web console | 22:11 |
| fungi | ooh, this is fun | 22:12 |
| fungi | Cross-Origin Request Blocked: The Same Origin Policy disallows reading the remote resource at http://logs.openstack.org/81/668181/1/gate/build-openstack-releasenotes/f863b3a/zuul-manifest.json. (Reason: CORS request did not succeed). | 22:12 |
| fungi | Cross-Origin Request Blocked: The Same Origin Policy disallows reading the remote resource at http://logs.openstack.org/81/668181/1/gate/build-openstack-releasenotes/f863b3a/html/job-output.json. (Reason: CORS request did not succeed). | 22:12 |
| corvus | okay, so it looks like it's the same mech for cors | 22:14 |
| corvus | that's a less esoteric error, so that may be a thread we can pull on | 22:14 |
| *** panda has quit IRC | 22:15 | |
| fungi | right, that doesn't 100% jive with my experience that pb is blocking cookies for logs.o.o and if i allow cookies for logs.o.o the content loads | 22:17 |
| *** panda has joined #zuul | 22:18 | |
| corvus | fungi: yeah, but i'm thinking that's how pb implements its blocking -- it somehow triggers the same codepath that cors does | 22:18 |
| fungi | oh, entirely possible | 22:19 |
| corvus | (maybe behind the scenes if it wants to block something, it updates firefox's cors value for a site) | 22:19 |
| *** sanjayu_ has quit IRC | 22:22 | |
| *** sanjayu_ has joined #zuul | 22:23 | |
| *** jamesmcarthur has quit IRC | 22:39 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Attempt to report CORS and related errors https://review.opendev.org/673707 | 22:47 |
| corvus | fungi: ^ that's based on a cors failure i manufactured; once we have that available through the preview site, we should see if that handles the PB case too | 22:48 |
| openstackgerrit | Clark Boylan proposed zuul/zuul master: Cleanup extra /'s in manifest render view https://review.opendev.org/673605 | 22:51 |
| *** rfolco|ruck has quit IRC | 23:02 | |
| *** michael-beaver has quit IRC | 23:15 | |
| clarkb | corvus: I'm not sure ^ is working beacuse I get weird urls back out except not weird in the way I expected (missing /s or extra /s) weird beacuse there are extra works in the urls. Maybe that is beacuse of how I am viewing them in js off of the logserver? | 23:38 |
| clarkb | http://logs.openstack.org/05/673605/2/check/zuul-build-dashboard/5d07d5f/npm/html/build/bc7c11147a464bd48c06f44677608611#logs has an extra /html/ in the raw url | 23:39 |
| corvus | clarkb: it's because the log_url has an html in it | 23:42 |
| corvus | i think that was overwritten with success_url | 23:43 |
| clarkb | oh so not related to my change but likely a bug we need to fix anyway? | 23:43 |
| corvus | yep | 23:43 |
| corvus | i think we need to update zuul to store the original log_url in the database, not what was reported back | 23:44 |
| corvus | also, i think we're going to need to deprecate success-url and failure-url altogether | 23:44 |
| clarkb | or turn them into something the builds page can link to? | 23:44 |
| corvus | yes -- that part is already there -- many of our jobs now report those as "artifacts" | 23:45 |
| clarkb | oh right | 23:45 |
| corvus | not that one though, so we should update it | 23:45 |
| corvus | and there's still ongoing cleanup for that https://review.opendev.org/673327 and https://review.opendev.org/672382 will make that much better | 23:46 |
| corvus | so anyway, i think we should change the log_url in the mysql reporter now, then over the next 2 weeks we can update the success-url jobs to return artifacts. then when we make the switch, it should be pretty straightforward to get to preview builds | 23:49 |
| corvus | fungi: can you try http://site.142c8d68a65c4ca48341390fae29998a.zuul.zuul-preview.opendev.org/ | 23:51 |
| *** sanjayu_ has quit IRC | 23:52 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!