| *** holser_ has quit IRC | 00:17 | |
| *** holser has joined #zuul | 00:18 | |
| *** tosky has quit IRC | 00:23 | |
| *** holser has quit IRC | 00:43 | |
| ianw | tristanC: your plugin is installed @ https://104.130.172.52/c/openstack/diskimage-builder/+/751610 ... it logs to the console, but i don't see a results box | 01:12 |
|---|---|---|
| fungi | does it want to match on comment user details maybe? | 01:50 |
| *** Goneri has quit IRC | 01:54 | |
| ianw | fungi: i don't think so, but ...maybe | 02:07 |
| *** rfolco has quit IRC | 02:20 | |
| *** armstrongs has joined #zuul | 02:36 | |
| *** armstrongs has quit IRC | 02:46 | |
| *** bhavikdbavishi has joined #zuul | 02:47 | |
| *** bhavikdbavishi1 has joined #zuul | 02:53 | |
| *** bhavikdbavishi has quit IRC | 02:55 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 02:55 | |
| *** zenkuro has quit IRC | 03:20 | |
| *** saneax has joined #zuul | 04:04 | |
| *** bhavikdbavishi has quit IRC | 04:23 | |
| *** bhavikdbavishi has joined #zuul | 04:24 | |
| *** saneax has quit IRC | 04:25 | |
| *** saneax has joined #zuul | 04:29 | |
| *** saneax has quit IRC | 04:51 | |
| *** evrardjp has quit IRC | 05:33 | |
| *** evrardjp has joined #zuul | 05:33 | |
| *** vishalmanchanda has joined #zuul | 05:41 | |
| *** wuchunyang has joined #zuul | 05:47 | |
| *** wuchunyang has quit IRC | 05:47 | |
| *** bhavikdbavishi1 has joined #zuul | 06:23 | |
| *** bhavikdbavishi has quit IRC | 06:25 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 06:25 | |
| *** jfoufas1 has joined #zuul | 06:56 | |
| *** bhavikdbavishi has quit IRC | 07:06 | |
| *** bhavikdbavishi has joined #zuul | 07:43 | |
| *** hashar has joined #zuul | 07:46 | |
| *** jcapitao has joined #zuul | 07:57 | |
| *** saneax has joined #zuul | 08:12 | |
| *** bhavikdbavishi1 has joined #zuul | 08:16 | |
| *** rpittau|afk is now known as rpittau | 08:17 | |
| *** bhavikdbavishi has quit IRC | 08:18 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 08:18 | |
| *** jpena|off is now known as jpena | 08:57 | |
| *** ChrisShort has quit IRC | 09:04 | |
| *** ChrisShort has joined #zuul | 09:06 | |
| *** holser has joined #zuul | 09:20 | |
| *** bhavikdbavishi has quit IRC | 09:49 | |
| *** rfolco has joined #zuul | 10:13 | |
| *** jfoufas1 has quit IRC | 10:43 | |
| *** bhavikdbavishi has joined #zuul | 11:37 | |
| *** bhavikdbavishi has quit IRC | 11:44 | |
| *** jcapitao is now known as jcapitao_lunch | 11:48 | |
| *** tosky has joined #zuul | 11:57 | |
| *** fbo is now known as fbo|lunch | 12:01 | |
| *** bhavikdbavishi has joined #zuul | 12:01 | |
| *** bhavikdbavishi1 has joined #zuul | 12:10 | |
| *** bhavikdbavishi has quit IRC | 12:12 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 12:12 | |
| *** zenkuro has joined #zuul | 12:13 | |
| tristanC | ianw: the author needs to be named zuul | 12:38 |
| *** jpena is now known as jpena|lunch | 12:39 | |
| tristanC | as sean-k-mooney suggested, the match should do '.* CI' (instead of 'Zuul.*' | 12:39 |
| sean-k-mooney | tristanC: well it should match on both | 12:40 |
| sean-k-mooney | the sepecial Zuul author + .* CI | 12:40 |
| *** jcapitao_lunch is now known as jcapitao | 12:41 | |
| zbr | sean-k-mooney: I guess we will not see Zuulander proposing CR very soon. | 12:48 |
| *** bhavikdbavishi has quit IRC | 12:49 | |
| *** bhavikdbavishi has joined #zuul | 12:51 | |
| danpawlik | clarkb: hey, could you check https://review.opendev.org/c/zuul/zuul/+/644927 ? I have made a patch set related yours comment: https://review.opendev.org/c/zuul/zuul/+/761441 | 12:54 |
| danpawlik | also tobiash please check the 761441 | 12:54 |
| danpawlik | thanks! | 12:54 |
| *** fbo|lunch is now known as fbo | 12:55 | |
| *** jfoufas1 has joined #zuul | 12:59 | |
| *** rfolco is now known as rfolco|ruck | 13:00 | |
| avass | zbr: do you know where the ansible modules ar stored nowadays? trying to find the synchronize module and it seems to be part of the posix collection but I don't know where that is | 13:06 |
| avass | oh nevermind I found the link to the repo from the galaxy page | 13:08 |
| avass | also, is synchronized supposed to work with kubectl connection? or should the kubernetes driver use the openshift prepare-workspace? | 13:10 |
| avass | I'm starting to think a unified prepare-workspace that does it best to figure what to do (like the combined windows/linux WIP one) would be nice to have | 13:12 |
| avass | but also for kubectl connections | 13:12 |
| zbr | avass: check https://github.com/ansible-community/protogen/tree/master/plugins | 13:18 |
| tristanC | ianw: fungi: here is how it looks when the author is named zuul: https://104.130.172.52/c/openstack/diskimage-builder/+/554002 | 13:22 |
| sean-k-mooney | zbr: there is no point in merging this until it support the third party cis | 13:24 |
| sean-k-mooney | it would activly discurage people form looking at the third party ci results | 13:24 |
| sean-k-mooney | tristanC: the render i pretty close to the old one so ya looks good. although i would porably put it to the right of the commit message that said that is where related changes and conflitcs will who up so its current location is fine too | 13:31 |
| avass | didn't someone implement a .zuulignore to be able to store playbooks and roles in zuul.d? | 13:32 |
| avass | ah yep: https://review.opendev.org/c/zuul/zuul/+/744811 | 13:34 |
| *** jpena|lunch is now known as jpena | 13:37 | |
| *** bhavikdbavishi has quit IRC | 13:38 | |
| openstackgerrit | Albin Vass proposed zuul/nodepool master: Document ImagePullPolicy for kubernetes driver. https://review.opendev.org/c/zuul/nodepool/+/764463 | 13:38 |
| avass | that was missing in the documentation by the way ^ :) | 13:39 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: ensure-nimble: ensure curl is present https://review.opendev.org/c/zuul/zuul-jobs/+/764465 | 13:51 |
| *** tobberydberg_ has quit IRC | 13:53 | |
| *** tobberydberg has joined #zuul | 13:56 | |
| openstackgerrit | Merged zuul/nodepool master: Changed informations provided in openshiftpods driver https://review.opendev.org/c/zuul/nodepool/+/756671 | 13:59 |
| *** rfolco has joined #zuul | 14:00 | |
| *** rfolco|ruck has quit IRC | 14:01 | |
| openstackgerrit | Sorin Sbârnea proposed zuul/zuul master: Makes console lines wrapping user configurable https://review.opendev.org/c/zuul/zuul/+/723603 | 14:07 |
| openstackgerrit | Sorin Sbârnea proposed zuul/zuul master: Make console lines wrapping user configurable https://review.opendev.org/c/zuul/zuul/+/723603 | 14:08 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: ensure-nimble: ensure depdencies are present https://review.opendev.org/c/zuul/zuul-jobs/+/764465 | 14:21 |
| *** rpittau is now known as rpittau|afk | 14:51 | |
| tristanC | sean-k-mooney: it seems like we have more room under the commit message. I can have a look at adding third-party-ci results too, though i'll wait for feedback on the current design first. | 14:59 |
| *** ashbullock has joined #zuul | 15:01 | |
| openstackgerrit | Sorin Sbârnea proposed zuul/zuul master: Make console lines wrapping user configurable https://review.opendev.org/c/zuul/zuul/+/723603 | 15:07 |
| ashbullock | Hey all, I've been working on a bitbucketcloud driver and have it at a point that's ready for feedback, any reviews would be greatly appreciated https://review.opendev.org/c/zuul/zuul/+/759003 | 15:08 |
| avass | I can't reach https://zuul-ci.org | 15:10 |
| avass | or I guess I can but it's very slow | 15:10 |
| avass | fungi, clarkb: I guess someone wants to take a look at that ^ :) | 15:12 |
| avass | looks like it's back! | 15:15 |
| avass | or maybe not :( | 15:18 |
| zbr | avass: not really, offline for me too. | 15:21 |
| *** jfoufas1 has quit IRC | 15:21 | |
| zbr | or maybe not, well with this speed it counts as offline to me. | 15:21 |
| *** vishalmanchanda has quit IRC | 15:22 | |
| zbr | i do see some timeouts loading fontawesome and a banner.jpg | 15:22 |
| zbr | 29s load time. | 15:23 |
| fungi | avass: zbr: yes, i've been trying to work out what's dragging apache down on the static content server | 15:23 |
| fungi | its connection to the backend seems fine, i can retrieve the files with a shell on that server, but apache is taking ages to serve anything | 15:24 |
| zbr | maybe there are some problems with the host running the vm? or networking. | 15:26 |
| * zbr thinking about introducing S3 joke here.... | 15:27 | |
| *** ikhan has joined #zuul | 15:29 | |
| fungi | yeah, i've checked the usual suspects there, cacti and typical investigative tools suggest everything's fine, including apache's mod_status detail, but even doing a wget from localhost on the server for a static file located on the server itself takes ages | 15:30 |
| avass | have you tried turning it off and on again? ;) | 15:31 |
| *** rfolco is now known as rfolco|ruck | 15:32 | |
| * zbr gives wisdom badge award to avass, no sarcasm included, it may really work. | 15:33 | |
| fungi | indeed, that's sort of what fixed it, unfortunately it doesn't provide a lot of additional insight into why it was being so slow | 15:35 |
| avass | on the other hand my zuul is working \o/ https://zuul.vassast.org/t/vassast/build/1a1124406c164a1e8d8c6677c17755ef | 15:35 |
| zbr | considering the amount of work load we had recently, i would not question such measures. First time it happens is reasonable to flip the switch. | 15:36 |
| zbr | once we see a recurrence we can go deeper | 15:37 |
| *** ikhan has quit IRC | 15:45 | |
| zbr | avass: do you know how to remove the blue contour that pf is adding on the summary page when you click to expand/collapse rows? | 15:46 |
| *** bhavikdbavishi has joined #zuul | 15:47 | |
| avass | the blue contour? | 15:50 |
| zbr | avass: is link/button indication | 15:50 |
| zbr | overflow-y: overlay; | 15:50 |
| zbr | https://sbarnea.com/ss/Screen-Shot-2020-11-27-15-50-50.png | 15:51 |
| avass | I don't have that | 15:51 |
| zbr | it appears when you click on the rows, at least with chrome. i suspect is part of the feature for marking current selection or so, but is distracting. | 15:51 |
| avass | could be something specific to chrome then since I don't get it in firefox :) | 15:52 |
| avass | and no I have no clue | 15:52 |
| zbr | it happens with both chrome and safari, but with firefox is more discrete as Firefox displays a hashed grey rectangle instead, less annoying. | 15:53 |
| avass | oh yeah | 15:53 |
| zbr | but i really doubt we want it, we should find a way to avoid it | 15:54 |
| zbr | btw, body {overflow-y: overlay;} seams to fix the issue where scrollbar forces browser to rewrap the page content. I wonder if it has some undersired side-effects. | 15:55 |
| openstackgerrit | Sorin Sbârnea proposed zuul/zuul master: Avoid page reflow when scrollbar appears or disapears https://review.opendev.org/c/zuul/zuul/+/764471 | 16:00 |
| sean-k-mooney | zbr: tristanC it snot very readable if you enable the dark theme | 16:09 |
| sean-k-mooney | https://104.130.172.52/c/openstack/diskimage-builder/+/554002 | 16:09 |
| sean-k-mooney | just enabled it for the zuul user there | 16:09 |
| sean-k-mooney | the grey box and white text is hard to read | 16:09 |
| sean-k-mooney | the table content is fine its just the header | 16:10 |
| sean-k-mooney | zbr: i assume we dont want any scrollbars in the results | 16:12 |
| sean-k-mooney | we want the rest fo the page to be shifted down | 16:12 |
| zbr | sean-k-mooney: not sure which kind of link you posted because browser seams to instantly close the window when it click it. | 16:12 |
| zbr | i doubt use of SSL with IP addresses is safe | 16:13 |
| sean-k-mooney | zbr: thats the temp vm form the ci run with the gerrit plugin | 16:13 |
| sean-k-mooney | e.g. the one we held to see what gerrit looks like | 16:14 |
| sean-k-mooney | with the plugin tristanC has | 16:14 |
| sean-k-mooney | it using self signed certs | 16:14 |
| sean-k-mooney | its not ment to be secure | 16:14 |
| sean-k-mooney | that is what it looks like with darkmode https://pasteboard.co/JCjZvCW.png | 16:17 |
| zbr | i am trying to explain that Chrome does not open the link at all, no prompts. With Safari I was able to load it after adding my admin pws. we better find an easier way to test. | 16:17 |
| sean-k-mooney | vs the default https://pasteboard.co/JCjZIZM.png | 16:17 |
| sean-k-mooney | zbr: that just because chrome is configured to not allow self signed certs | 16:18 |
| sean-k-mooney | firfox and safariy give you the option | 16:18 |
| zbr | I just had an idea about how we could redesign the results table to make it integrate better with gerrit, mainly no longer a table. | 16:18 |
| fungi | zbr: as a workaround you could temporarily set review.opendev.org to that ipv4 address in your local /etc/hosts to test | 16:18 |
| fungi | assuming chrome doesn't bypass file lookup for host resolution anyway | 16:19 |
| zbr | fungi: good luck editing /etc/hosts on an ipad | 16:19 |
| *** bhavikdbavishi has quit IRC | 16:19 | |
| *** bhavikdbavishi1 has joined #zuul | 16:19 | |
| sean-k-mooney | i thihk the non dark mode version looks pretty good the only isue is the success/failure text is not a link but they are just to the left | 16:20 |
| zbr | here is my idea: make each job a colored label that indicate it result, add tooltip with detail: "SUCCESS in 10 mins". They will wrap very well on any screen width. | 16:20 |
| zbr | also they will look fine with dark mode too. | 16:20 |
| sean-k-mooney | zbr: i think if you care to test configuring yoru broser to allow self signed cert is resonable | 16:20 |
| sean-k-mooney | zbr: not really | 16:20 |
| sean-k-mooney | the dark mode the contrast is not greate | 16:20 |
| sean-k-mooney | the gray and white are too close | 16:21 |
| sean-k-mooney | at least on firefox | 16:21 |
| sean-k-mooney | the text of the tabel looks good | 16:21 |
| sean-k-mooney | but the title of the box is not readable | 16:21 |
| sean-k-mooney | if you compare https://pasteboard.co/JCjZvCW.png to https://pasteboard.co/JCjZIZM.png | 16:22 |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 16:22 | |
| zbr | holly.... gerrit is not usable on dark mode and that has nothing to do with tristan chage. | 16:23 |
| *** bhavikdbavishi has quit IRC | 16:23 | |
| sean-k-mooney | well its alpha | 16:23 |
| sean-k-mooney | i dont mind it really | 16:23 |
| sean-k-mooney | but you dont have the option to chosse your own theme anymore | 16:24 |
| sean-k-mooney | for the review content | 16:24 |
| sean-k-mooney | over all its less clear but its usabel | 16:24 |
| sean-k-mooney | i kind of wish there was a solarized dark mode | 16:25 |
| *** bhavikdbavishi has joined #zuul | 16:25 | |
| zbr | once i enabled dark mode in user preference it became usable. | 16:25 |
| zbr | but i am unable to test tristan change with that as i cannot login to change preferences. | 16:25 |
| zbr | i find it weird that gerrit developers decided to ignore browser theme mode | 16:26 |
| sean-k-mooney | oh you just type zuul then click become account | 16:26 |
| sean-k-mooney | dont try to use your own accont gerrit is configure in developer mode | 16:26 |
| sean-k-mooney | for auth at least | 16:26 |
| zbr | it it reports account not found | 16:26 |
| sean-k-mooney | the link wont work | 16:26 |
| sean-k-mooney | but if you type zuul in account name and click the become account button it shoudl work | 16:27 |
| zbr | thanks | 16:27 |
| sean-k-mooney | that what i did at any rate but im using firefox on a laptop | 16:28 |
| sean-k-mooney | not sure if that will work on an ipad | 16:28 |
| zbr | my hint: best border is no border | 16:28 |
| sean-k-mooney | yep | 16:28 |
| zbr | ask anyone in ireland .. | 16:28 |
| sean-k-mooney | hehe :) just dont ask the uk | 16:29 |
| zbr | sean-k-mooney: what do you think about my idea of having labels one after another instead of a table? | 16:29 |
| zbr | this example is one of the good cases, where the table fits, but I seen lots of changes where you have multiple changes using same topic or lots of conflicts | 16:30 |
| zbr | in these the "UI" becomes bloated and table makes it super-bloated. | 16:30 |
| sean-k-mooney | i actully like having all that info present | 16:31 |
| zbr | look at "merge conflict" badge like label on left top corner and imagine that we would display jobs similar to these. | 16:31 |
| sean-k-mooney | i would still want al jobs to be listed | 16:31 |
| sean-k-mooney | so if we had a lable/badge per job sure | 16:31 |
| sean-k-mooney | but it would still take up the same amount of space i think | 16:32 |
| sean-k-mooney | i feel like im one of the few peopel that actully looks for non voting job failures however | 16:32 |
| zbr | we can expose ( [v] tox-linters ), ( [x] tox-py39 ) kind of. | 16:32 |
| zbr | it would work even for color blind if we add a small emoji for success/failure. | 16:32 |
| sean-k-mooney | there is a unicode tick and X we can use | 16:33 |
| zbr | not sure yet about the time, for me displaying it in tooltip would be fine, but i am sure someone will want it on the screen | 16:33 |
| sean-k-mooney | i would be fine with that but still think the table makes sense | 16:33 |
| sean-k-mooney | the run time i like that there but i guess i can get it form the comment | 16:34 |
| fungi | not only is red/green color blindness the most common variety, but also there are very large parts of the world where red and green have the exact opposite meaning | 16:34 |
| sean-k-mooney | i ocationally looke at it to tell at a glance if devstack ran | 16:34 |
| fungi | (red=good) | 16:34 |
| zbr | the table does not wrap well with narrow screens or builds that have big number of jobs (over 8-10 it becomes a real nuisance) | 16:34 |
| sean-k-mooney | e.g. if a tempest job failed in 38mins its proably becaue devstack failed not the jobs | 16:35 |
| *** zenkuro has quit IRC | 16:35 | |
| *** zenkuro has joined #zuul | 16:35 | |
| sean-k-mooney | zbr: if i reviewing on a pone i use a gerrit app instad of the gui | 16:35 |
| sean-k-mooney | i am not sure we shoudl be optimising for that at the expense of normal dispalys | 16:36 |
| zbr | probably I can play with the tamper monkey script to build a POC around how i see it, and it would be much easier to check on something testable. | 16:36 |
| zbr | i do run into lots of issues with zuul because most zuul developers never try lanscape displays | 16:36 |
| sean-k-mooney | yep you can share the script and we can run it or you can share screan shot in an email | 16:37 |
| sean-k-mooney | zbr: you mean portrait | 16:37 |
| zbr | and while I have 4K+5K displays, I do use both of them in split mode, so my browser windows are most of the time portrait. | 16:37 |
| sean-k-mooney | landscape is the default for monitors | 16:37 |
| fungi | i actually have my workstation monitors rotated 90 degrees into portrait orientation | 16:37 |
| sean-k-mooney | i have one that way for irc | 16:38 |
| fungi | and then i put them together side-by-side | 16:38 |
| sean-k-mooney | and then an ultrawidn 21:9 for my main use | 16:38 |
| sean-k-mooney | then have that split in half | 16:38 |
| zbr | anyway the new gerrit is 100x better with narrow screens that 2.x | 16:38 |
| fungi | my main browser window on the workstation is 1440px wide by 2560px high | 16:39 |
| zbr | with zuul is fixable | 16:39 |
| sean-k-mooney | i need to redeploy or change how my zuul runs actully | 16:39 |
| sean-k-mooney | its really really reaaly slow | 16:39 |
| sean-k-mooney | if you dont serve the static content with a webserver | 16:40 |
| sean-k-mooney | other then the build in one in the container | 16:40 |
| sean-k-mooney | its not setting the cache headers properly | 16:40 |
| zbr | ½ 4K screenshot: https://sbarnea.com/ss/Screen-Shot-2020-11-27-16-40-05.png -- look how much of the build results fits the screen. | 16:40 |
| sean-k-mooney | so even thouhg i put cloudflare cdn in front its still slow to load the static content | 16:40 |
| sean-k-mooney | zbr: yep that looks fine to me | 16:41 |
| sean-k-mooney | you have to scolle to see the files but you have all of the context visabel | 16:42 |
| zbr | i see a lot of wasted space, the entire page is ~2.8x screens in lengh. Ideally it should be ~2x screens. | 16:42 |
| sean-k-mooney | there is space on the left | 16:42 |
| sean-k-mooney | but i dont see much wasted space elsewhere | 16:42 |
| zbr | yes, that is what i was trying to highlight the left-sidebar area has space but not for tables. | 16:42 |
| fungi | sean-k-mooney: we deploy ours with an apache rewrite to zuul-web listening on the loopback, and then use mod_cache to improve performance for ^/api/tenant/.*/status and /static | 16:43 |
| sean-k-mooney | zbr: that space wotn exist normally | 16:43 |
| sean-k-mooney | the merge conflits and restate secto reflowed from the right | 16:43 |
| sean-k-mooney | fungi: ya i saw some docs for that jsut didnt do it yet | 16:44 |
| sean-k-mooney | i was hoping cloudflare woudl do it for me but no since it does not have the right caching headers | 16:44 |
| fungi | sean-k-mooney: that sounds like something we could solve in zuul-web if you want to to add cache-relevant headers | 16:46 |
| sean-k-mooney | fungi: im using traefik currenlty to do ssl termination and reversproxying ill likely see if i can do it with that if not ill use apache or nginx | 16:46 |
| sean-k-mooney | fungi: ya we coudl i just dont really know what they should be | 16:46 |
| sean-k-mooney | there is no cache expiry time so it does not get cached | 16:47 |
| *** hashar has quit IRC | 16:47 | |
| sean-k-mooney | i was going to see if traefik middelware coudl just inject them, doing it in zuul web would be a better long term soltion proably | 16:48 |
| sean-k-mooney | im currently useing traefik to add cors headears to get the logs working | 16:49 |
| sean-k-mooney | that took a while to figure out | 16:49 |
| fungi | we rewrite ^/api/tenant/(.*)/console-stream to ws://127.0.0.1:9000/api/tenant/$1/console-stream | 16:53 |
| fungi | which i think works around that problem? | 16:53 |
| sean-k-mooney | no this is for the serving of the logs that are uploaded after the job | 16:53 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: DNM: trigger upload_logs_s3 https://review.opendev.org/c/zuul/zuul-jobs/+/764476 | 16:53 |
| fungi | oh! that | 16:53 |
| sean-k-mooney | i have zuul at zuul.seanmooney.info but the logs are at logs.seanmooney.info | 16:53 |
| fungi | yeah we use the cors settings in our swift containers | 16:53 |
| sean-k-mooney | and by defualt firefox block that | 16:54 |
| sean-k-mooney | yep i actully set the header in the caddy webserver i have serving the directory tree | 16:54 |
| sean-k-mooney | i should have don it in trafik it was a pain to get working since caddy recently was rewitten | 16:54 |
| sean-k-mooney | so all the guide where for the old version | 16:55 |
| avass | oh no | 16:55 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: DNM: trigger upload_logs_s3 https://review.opendev.org/c/zuul/zuul-jobs/+/764476 | 16:56 |
| sean-k-mooney | oh i never start zuul after i showdown my cloud to do some electical work yesteraday | 16:56 |
| sean-k-mooney | i should fix that | 16:57 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: upload-logs-s3: fix test-jobs files: attribute https://review.opendev.org/c/zuul/zuul-jobs/+/764476 | 16:57 |
| sean-k-mooney | fungi: by the way not all of the s3 log servers are configured correctly | 16:58 |
| sean-k-mooney | fungi: im pretty sure the rackspace one is serving deflate encoded logs without the right header | 16:59 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: DNM: trigger upload_logs_s3 https://review.opendev.org/c/zuul/zuul-jobs/+/764478 | 16:59 |
| *** ashbullock has quit IRC | 17:00 | |
| sean-k-mooney | fungi: half the time i can open the raw logs with baseiclly curl | editor or in my case lnav <url> but the rest fo the time mainly on rackspace i think it get compress files without the header and the editor does not understand that | 17:00 |
| sean-k-mooney | and need to use zcat to fix it | 17:00 |
| sean-k-mooney | since you mentioned settign headers in zuul | 17:00 |
| sean-k-mooney | *swift | 17:01 |
| *** tosky has quit IRC | 17:04 | |
| *** wuchunyang has joined #zuul | 17:09 | |
| avass | upload_logs_s3 role is broken for some reason | 17:09 |
| avass | I'm confused about the error but it looks like it's something with boto | 17:10 |
| *** wuchunyang has quit IRC | 17:13 | |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: GZIPCompressedStream default read length to -1 https://review.opendev.org/c/zuul/zuul-jobs/+/764478 | 17:19 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: GZIPCompressedStream default read length to -1 https://review.opendev.org/c/zuul/zuul-jobs/+/764478 | 17:20 |
| avass | can I get a review on that ^ it fixes it when running it locally | 17:23 |
| avass | oh and the test job should probably be green | 17:24 |
| fungi | sean-k-mooney: yeah, rackspace puts some secret sauce cdn in front of their object store which seems to do some sort of content inspection and then unconditionally compress some things independent of whatever's set in swift | 17:34 |
| fungi | i think you can influence it by setting accept-encoding | 17:36 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Add configuration to make logs public https://review.opendev.org/c/zuul/zuul-jobs/+/764483 | 17:45 |
| zbr | fungi: corvus: https://review.opendev.org/c/zuul/zuul/+/723603 "the wrapping" one | 17:46 |
| avass | I think corvus is on vacation this week | 17:47 |
| avass | sean-k-mooney: are you using s3 for log storage in zuul? | 17:50 |
| tristanC | sean-k-mooney: thanks, i'll have a look to invert the color when dark-mode is enabled | 17:54 |
| *** jpena is now known as jpena|off | 17:55 | |
| zbr | yeah, i do not want to merge the wrapping without getting feeback from him, but this does not mean others are not welcomed to look at it. | 17:58 |
| zbr | tristanC: just remove the borders and it should be fine. | 17:59 |
| zbr | tristanC: here is how native gerrit CI integration looks: https://gerrit-review.googlesource.com/c/gerrit/+/289025 | 18:03 |
| zbr | it may worth going towards that direction. | 18:04 |
| zbr | not ugly at all, and i like that is one line unless you click it. | 18:04 |
| tristanC | zbr: yeah, it seems like we could mount a react component and do a better rendering in a follow-up version | 18:12 |
| *** jcapitao has quit IRC | 18:12 | |
| tristanC | zbr: ideally i would re-use existing an components from the zuul/web directory | 18:13 |
| tristanC | though i'm mostly looking forward getting the same zuul-results table feature from gerrit 2.x | 18:13 |
| avass | 764478 and 764476 fixes upload_logs_s3 if anyone has time to spare | 18:15 |
| sean-k-mooney | fungi: with curl proably i doint think i can wit lnav/vim but the main issue is it does not say its compressed proerly | 18:19 |
| sean-k-mooney | if it did it woudl be fine they support compressed logs | 18:19 |
| sean-k-mooney | it just needs to know its compressed | 18:19 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Add configuration to make logs public https://review.opendev.org/c/zuul/zuul-jobs/+/764483 | 18:20 |
| sean-k-mooney | the browser figures it out but not everything does | 18:20 |
| sean-k-mooney | zbr: we talked about usign that breilfy in the main thread | 18:20 |
| sean-k-mooney | the checks plugin could work but we woudl need to chagne zuul to support it | 18:21 |
| sean-k-mooney | and thrid party CIs would have to be updated | 18:21 |
| sean-k-mooney | zbr: https://zuul-ci.org/docs/zuul/reference/drivers/gerrit.html#checks-plugin-support-experimental | 18:22 |
| sean-k-mooney | there is some experimal suport for it already in zuul so for zuul based third party cis that might be ok | 18:22 |
| sean-k-mooney | but the comment format is basically an api contract we have had for years so it might take some time to migrate everyone | 18:23 |
| *** Goneri has joined #zuul | 18:53 | |
| *** saneax has quit IRC | 18:57 | |
| *** tosky has joined #zuul | 19:03 | |
| *** bhavikdbavishi has quit IRC | 19:04 | |
| *** ikhan has joined #zuul | 19:04 | |
| *** parallax has joined #zuul | 19:08 | |
| *** ikhan has quit IRC | 19:09 | |
| tristanC | zbr: sean-k-mooney: i've deployed https://softwarefactory-project.io/r/20194 on https://104.130.172.52/c/openstack/diskimage-builder/+/554002 , third-party CI and dark mode should now be working | 19:10 |
| fungi | sean-k-mooney: well, zuul itself doesn't need more changes to support the gerrit checks plugin, there's support already implemented, opendev's zuul and other ci systems would need to switch to using it though. and for some (probably very long) period of time you're going to have a situation where some ci systems are reporting there via checks plugin and others via review comment (some marked as "robot" comments | 19:23 |
| fungi | and some not depending on implementation and vintage) | 19:23 |
| *** ikhan has joined #zuul | 19:43 | |
| avass | uuuh, I don't think the example github pipeline should trigger on 'event: check_run' since that caused the check pipeline to loop for me | 19:43 |
| sean-k-mooney | avass: https://github.com/SeanMooney/ci-sean-mooney/blob/main/zuul.d/pipelines.yaml#L156-L164 | 19:51 |
| sean-k-mooney | this seam to work for me | 19:51 |
| sean-k-mooney | at least i have not seen it loop but it might in some cases i dont no | 19:53 |
| avass | hmm I had to remove that otherwise the check restarter over and over again | 19:55 |
| sean-k-mooney | if something fails in the pre playbook zuul will try 3 times to execute the job | 19:56 |
| avass | yeah but this succeeded | 19:56 |
| sean-k-mooney | oh ok | 19:56 |
| *** ikhan has quit IRC | 19:57 | |
| sean-k-mooney | avass: oh i am not subsribed to check run in the app | 19:58 |
| sean-k-mooney | i dont think its one fo the events that is listed in the doc | 19:58 |
| *** mugsie has quit IRC | 19:59 | |
| sean-k-mooney | oh it is https://zuul-ci.org/docs/zuul/reference/drivers/github.html | 19:59 |
| avass | It should be because I only gave it access to what is documented :) | 19:59 |
| sean-k-mooney | but ya i dont have check run ticked | 20:00 |
| avass | everything seem to be working now, though zuul links to the directory containing the logs instead of the index.html | 20:00 |
| avass | https://zuul.vassast.org/t/vassast/build/2071890236394cb4baffcde3a954c13c/logs | 20:00 |
| avass | sean-k-mooney: that's probably why you don't get the looping behaviour then :) | 20:00 |
| *** mugsie has joined #zuul | 20:01 | |
| avass | so the CORS logs work but the link to the logs don't work, heh | 20:02 |
| *** y2kenny has joined #zuul | 20:02 | |
| avass | tobiash: did you use s3 for storing logs? | 20:03 |
| avass | digitalocean is supposed to be s3 compatible so I wonder if this is an s3 quirk or it's just something digitalocean does | 20:03 |
| y2kenny | Hi, is there any reason why ansible_user_dir is not part of the safe_args here: https://opendev.org/zuul/zuul/src/branch/master/zuul/ansible/base/action/add_host.py#L20 ? | 20:03 |
| avass | y2kenny: I would guess someone missed adding it and no one needed it before :) | 20:04 |
| y2kenny | avass: ok I will make a patch | 20:04 |
| y2kenny | ensure_output_dir uses it by default | 20:04 |
| openstackgerrit | Kenny Ho proposed zuul/zuul master: Add ansible_user_dir as safe_arg for add-host https://review.opendev.org/c/zuul/zuul/+/764498 | 20:13 |
| tobiash | avass: we're using swift | 20:15 |
| *** ikhan has joined #zuul | 20:16 | |
| avass | tobiash: alright, I remember you speaking about moving to s3 so I was wondering if you started doing that or not | 20:17 |
| avass | we'll probably do that soon now that I know that it works with digitalocean at least :) | 20:17 |
| openstackgerrit | Kenny Ho proposed zuul/zuul master: Add ansible_user_dir as safe_arg for add-host https://review.opendev.org/c/zuul/zuul/+/764498 | 20:18 |
| y2kenny | avass: do you know who I should add to review for the safe_arg patch? | 20:19 |
| avass | y2kenny: the zuul maintainers usually pick them up whenever they have time | 20:19 |
| y2kenny | ok. I will leave it like that then | 20:20 |
| avass | y2kenny: but shouldn't ansible pick up ansible_user_dir automatically? | 20:20 |
| y2kenny | avass: I guess when using add-host, I have to add everything? | 20:20 |
| avass | oh maybe you need to run setup for that host | 20:21 |
| y2kenny | like... I am manually manipulating the inventory | 20:21 |
| y2kenny | yea | 20:21 |
| avass | you could probably do that for the host you just added | 20:21 |
| *** ikhan has quit IRC | 20:21 | |
| avass | y2kenny: https://docs.ansible.com/ansible/latest/collections/ansible/builtin/setup_module.html | 20:21 |
| avass | I'd expect that to work a bit better | 20:22 |
| y2kenny | oh you mean add the host, gather facts and then have ansible populate the user_idr | 20:22 |
| y2kenny | dir | 20:22 |
| avass | yeah and a lot of other facts | 20:23 |
| *** ikhan has joined #zuul | 20:27 | |
| openstackgerrit | Kenny Ho proposed zuul/zuul master: Clearify terms around pause-job documentation https://review.opendev.org/c/zuul/zuul/+/764499 | 20:31 |
| openstackgerrit | Kenny Ho proposed zuul/zuul master: Clarify terms around pause-job documentation https://review.opendev.org/c/zuul/zuul/+/764499 | 20:32 |
| openstackgerrit | Kenny Ho proposed zuul/zuul master: Clarify terms around pause-job documentation https://review.opendev.org/c/zuul/zuul/+/764499 | 20:47 |
| *** rfolco|ruck has quit IRC | 21:04 | |
| y2kenny | fungi: please let me know if the pause-job documentation modification is ok. | 21:11 |
| *** ikhan has quit IRC | 21:13 | |
| *** ikhan has joined #zuul | 21:16 | |
| *** ikhan has quit IRC | 21:21 | |
| *** lyr has quit IRC | 21:23 | |
| *** pleia2 has quit IRC | 21:24 | |
| *** smyers has quit IRC | 21:24 | |
| *** pleia2 has joined #zuul | 21:25 | |
| *** avass has quit IRC | 21:25 | |
| *** y2kenny has quit IRC | 21:25 | |
| *** avass has joined #zuul | 21:27 | |
| *** smyers has joined #zuul | 21:28 | |
| *** corvus has quit IRC | 21:29 | |
| *** corvus has joined #zuul | 21:30 | |
| fungi | zuul-maint: ^ some food for thought there... digging deeper it looks like we've ended up using parent/child to mean two entirely separate concepts with respect to defining and instantiating jobs, and if we want to avoid confusing our users we should probably deprecate and rename one or the other | 21:46 |
| fungi | probably a good topic to re-raise next week when more folks are around to talk about options | 21:46 |
| fungi | in short we have zuul.child_jobs which is not the inverse of job.parent, but then we document their inverses by talking about "parent jobs" and "child jobs" while referring to both of those concepts in different places | 21:47 |
| *** ikhan has joined #zuul | 21:51 | |
| openstackgerrit | Kenny Ho proposed zuul/zuul master: Clarify terms around pause-job documentation https://review.opendev.org/c/zuul/zuul/+/764499 | 22:11 |
| avass | fungi: yeah that has always been confusing to talk about | 22:37 |
| avass | I think it's especially confusing since zuul_return returns data to child jobs, which are dependent jobs, but if you want to expose data to inheriting jobs (inverse of parent what you'd expect is a child) you should either cache a fact or save a file | 22:40 |
| *** armstrongs has joined #zuul | 22:59 | |
| *** armstrongs has quit IRC | 23:09 | |
| *** wuchunyang has joined #zuul | 23:12 | |
| *** wuchunyang has quit IRC | 23:16 | |
| *** tosky has quit IRC | 23:42 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
