| *** Goneri has quit IRC | 00:39 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: install python2-pip on SuSE when required https://review.opendev.org/724777 | 01:55 |
|---|---|---|
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: Install backported pip for Xenial https://review.opendev.org/724788 | 01:55 |
| *** swest has quit IRC | 01:55 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Add plain nodes to testing https://review.opendev.org/724776 | 01:55 |
| *** swest has joined #zuul | 02:10 | |
| *** bhavikdbavishi has joined #zuul | 02:40 | |
| *** bhavikdbavishi1 has joined #zuul | 02:43 | |
| *** bhavikdbavishi has quit IRC | 02:44 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 02:44 | |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Handle ipv6 literal zookeeper addresses https://review.opendev.org/725157 | 03:38 |
| *** evrardjp has quit IRC | 04:35 | |
| *** evrardjp has joined #zuul | 04:36 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: install python2-pip when running under Python 2 https://review.opendev.org/724777 | 04:46 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: Install backported pip for Xenial https://review.opendev.org/724788 | 04:46 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Add plain nodes to testing https://review.opendev.org/724776 | 04:46 |
| *** bhavikdbavishi has quit IRC | 04:53 | |
| *** ysandeep|afk is now known as ysandeep | 05:20 | |
| *** dpawlik has joined #zuul | 06:04 | |
| *** dpawlik has quit IRC | 06:04 | |
| *** dpawlik has joined #zuul | 06:07 | |
| *** dpawlik has joined #zuul | 06:08 | |
| *** rpittau|afk is now known as rpittau | 06:33 | |
| *** saneax has joined #zuul | 06:33 | |
| AJaeger | I added a check that makes docs job fail in post if there's no html output. This now leads to a POST_FAILURE when the normal build fails as well. So, a FAILURE turned into POST_FAILURE, see https://zuul.opendev.org/t/zuul/build/afbc8c39929a46039aa4d987971cdb4f . Is there a way to check in the post playbook whether the job passed so far to avoid creating the POST_FAILURE? | 06:55 |
| AJaeger | checking whether tox_output exist and then for tox_output.failed is one option but would tie those roles together. Is there a better way? | 07:03 |
| *** jcapitao has joined #zuul | 07:12 | |
| *** yolanda has joined #zuul | 07:20 | |
| *** bhavikdbavishi has joined #zuul | 07:22 | |
| *** bhavikdbavishi1 has joined #zuul | 07:24 | |
| *** bhavikdbavishi has quit IRC | 07:26 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 07:26 | |
| avass | AJaeger: you mean something like: https://zuul-ci.org/docs/zuul/reference/jobs.html#var-zuul_success ? :) | 07:29 |
| *** tosky has joined #zuul | 07:32 | |
| *** bhavikdbavishi has quit IRC | 07:33 | |
| *** sshnaidm|off is now known as sshnaidm | 07:33 | |
| *** jpena|off is now known as jpena | 07:49 | |
| AJaeger | avass: exactly! | 08:02 |
| AJaeger | thanks, avass ! | 08:03 |
| *** guillaumec has joined #zuul | 08:06 | |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Fix fetch-sphinx-tarball fails https://review.opendev.org/725210 | 08:09 |
| *** ysandeep is now known as ysandeep|lunch | 08:35 | |
| *** bhavikdbavishi has joined #zuul | 09:10 | |
| avass | AJaeger: does it show POST_FAILURE if it fails in both run and post-run? otherwise? | 09:30 |
| *** panda|ruck is now known as panda|pto | 09:40 | |
| AJaeger | avass: see the link I posted above, it shows POST_FAILURE | 09:43 |
| avass | AJaerger: oh yeah, got it | 10:01 |
| *** rpittau is now known as rpittau|bbl | 10:14 | |
| *** bhavikdbavishi has quit IRC | 10:30 | |
| *** bhavikdbavishi has joined #zuul | 10:34 | |
| *** ysandeep|lunch is now known as ysandeep | 10:47 | |
| *** bhavikdbavishi has quit IRC | 11:20 | |
| *** bhavikdbavishi has joined #zuul | 11:21 | |
| *** guillaumec has quit IRC | 11:22 | |
| *** jcapitao is now known as jcapitao|lunch | 11:23 | |
| *** jpena is now known as jpena|lunch | 11:30 | |
| *** ysandeep is now known as ysandeep|brb | 11:36 | |
| avass | We don't have a role to upload artifacts to a simple rest api right? | 11:36 |
| AJaeger | avass: I'm not aware of one. I'm only aware of the upload roles in zuul-jobs. | 11:39 |
| avass | AJaeger: yeah I couldn't see anything that looked liked it in the zuul-jobs at least, will probably push one later then :) | 11:40 |
| avass | almost done with it | 11:40 |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: Connect executor to Zookeeper https://review.opendev.org/716262 | 11:41 |
| *** ysandeep|brb is now known as ysandeep | 11:52 | |
| *** rfolco has joined #zuul | 12:04 | |
| *** dmellado has quit IRC | 12:10 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: spec: add a zuul-runner cli https://review.opendev.org/681277 | 12:15 |
| *** dmellado has joined #zuul | 12:16 | |
| *** rpittau|bbl is now known as rpittau | 12:22 | |
| *** gundalow has joined #zuul | 12:23 | |
| *** jcapitao|lunch is now known as jcapitao | 12:29 | |
| *** rlandy has joined #zuul | 12:31 | |
| *** jpena|lunch is now known as jpena | 12:31 | |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: Move fingergw config to fingergw https://review.opendev.org/664949 | 12:34 |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: WIP: Route streams to different zones via finger gateway https://review.opendev.org/664965 | 12:34 |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: Add --validate-tenants option to zuul scheduler https://review.opendev.org/542160 | 12:41 |
| *** hashar has joined #zuul | 12:53 | |
| *** sgw has joined #zuul | 13:01 | |
| *** Goneri has joined #zuul | 13:08 | |
| *** bhavikdbavishi has quit IRC | 13:10 | |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: Connect executor to Zookeeper https://review.opendev.org/716262 | 13:29 |
| *** hashar has quit IRC | 13:33 | |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: Connect executor to Zookeeper https://review.opendev.org/716262 | 13:51 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Handle ipv6 literal zookeeper addresses https://review.opendev.org/725157 | 13:52 |
| *** cdearborn has joined #zuul | 14:14 | |
| *** ysandeep is now known as ysandeep|brb | 14:18 | |
| mordred | corvus: when you're ready for more debugging, I held the nodes for the multiarch job failure: https://zuul.opendev.org/t/zuul/build/bdcee9695b864a1e86c6aedc2b8f9a13 | 14:26 |
| corvus | masterpe: excellent; i think maybe after breakfast in about 30m | 14:32 |
| corvus | mordred: ^ sorry masterpe (tab completion failure) | 14:36 |
| mordred | corvus: woot | 14:41 |
| *** hashar has joined #zuul | 14:45 | |
| *** panda|pto has quit IRC | 14:46 | |
| avass | I think build container image is broken: https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/build-container-image/tasks/main.yaml#L10 :( | 14:49 |
| avass | but it also looks like it's never used anywhere? http://codesearch.openstack.org/?q=build-container-image&i=nope&files=&repos= | 14:50 |
| *** hashar has quit IRC | 14:50 | |
| *** panda has joined #zuul | 14:50 | |
| corvus | avass: yeah, it's part of a new set of jobs to use non-docker tooling; work in progress | 14:50 |
| avass | corvus: ah, that makes sense I guess | 14:51 |
| corvus | avass: it should have a test job i thought? | 14:51 |
| avass | corvus: yep found a test playbook | 14:51 |
| corvus | zuul-jobs-test-registry-podman | 14:51 |
| corvus | avass: it's actually the same playbooks that test the docker versions of the roles | 14:52 |
| corvus | we just tell it to use podman instead of docker | 14:52 |
| *** ysandeep|brb is now known as ysandeep | 14:52 | |
| corvus | (the roles are meant to be 1:1 compatible) | 14:52 |
| avass | corvus: I'll look around and see why that didn't fail, maybe it didn't trigger when it should have, since using a bare 'item' in tasks/main.yaml doesn't seem like it should work | 14:52 |
| avass | unless the role is included in a loop without setting loop_ar | 14:53 |
| avass | loop_var* | 14:53 |
| avass | this looks like it should fail: https://opendev.org/zuul/zuul-jobs/src/branch/master/test-playbooks/registry/test-registry.yaml#L28 | 14:54 |
| *** sassyn has joined #zuul | 14:56 | |
| sassyn | Hi All Again | 14:56 |
| sassyn | Hope everyone is safe! | 14:56 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: DNM: trigger registry tests https://review.opendev.org/725294 | 14:57 |
| corvus | sassyn: hi! how's it going? | 14:59 |
| sassyn | HI corvus, all good here in IL | 14:59 |
| sassyn | great weather and we can go out no more then 1 KM | 14:59 |
| sassyn | how at your place? | 15:00 |
| corvus | sassyn: i can roam a bit more than that, but it's frowed upon. a nearby county just increased the permitted range from 5 miles to 10 miles. all told, in california it's not as bad as other places, and people are expecting to open more businesses in a few weeks. | 15:02 |
| avass | sassyn: hi, wow no more than 1KM | 15:02 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: spec: add a zuul-runner cli https://review.opendev.org/681277 | 15:05 |
| avass | sassyn: here in sweden (almost) everything is like normal. except that people have a greater distance between eachother than normal, and stores put out markers so people keep their distance from eachother | 15:06 |
| avass | and most people who can work from home | 15:06 |
| sassyn | Sweden is cool! | 15:08 |
| sassyn | Consider the following: | 15:09 |
| avass | whelp, just noticed I missed that it defaults to something else if 'item' is not defined | 15:11 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Do not use bare 'item' in build-container-image https://review.opendev.org/725298 | 15:11 |
| sassyn | Now B failed since the repo now is not in potion 0, but in 0+A. However it is know (with manual build - not via zuul) that B+C is working. | 15:12 |
| avass | corvus: how about we do that ^ instead | 15:12 |
| *** ysandeep is now known as ysandeep|away | 15:13 | |
| avass | since it's probably better to loop over that role and set the variable instead | 15:13 |
| *** sanjayu_ has joined #zuul | 15:14 | |
| *** saneax has quit IRC | 15:17 | |
| sassyn | Rephrasing: | 15:18 |
| sassyn | Consider the following: | 15:18 |
| sassyn | so we must run the test again). In our case the repo is still at postion 0 so test pass and at the end A was merged. | 15:18 |
| sassyn | failed since the repo now is not in potion 0, but in 0+A. However it is know (with manual build - not via zuul) that B+C is working. | 15:18 |
| sassyn | submit B+C ? | 15:18 |
| sassyn | since I know B+C is working.. | 15:19 |
| avass | sassyn: do you mean that HEAD+B fails but HEAD+B+C succeeds? | 15:20 |
| sassyn | Yes | 15:20 |
| avass | sassyn: I think multiple people have wanted a pipeline manager that works like that, but don't think that's possible to do in zuul at the moment | 15:20 |
| avass | my idea would be to either squash those commits, or use feature toggles. or something like that, depends on the situation really | 15:21 |
| clarkb | I think tobiash is working on the atomic merges stuff? but ya if HEAD+C works land C before B. If B and C are required bug fixes maybe they should be squashed. If B needs C for interface updates, etc then perhaps land new interfaces alongside old ones in C' first then land B then land C'' to remove old interfaces | 15:22 |
| sassyn | so B will fail, and C will be run, if pass it will be commited. B will say with +2 for code review, but with -1 for verified. How to I tell Zuul to retigger it? remove the verified ? | 15:23 |
| avass | oh so HEAD+C succeds? | 15:23 |
| avass | so then HEAD+C+B succeed as well? then you should be able to recheck/regate, if you | 15:24 |
| avass | you've set up something like that | 15:24 |
| sassyn | yes C worked. | 15:24 |
| sassyn | but B have to be recheck, and I can't do it automatically right? | 15:25 |
| avass | no I don't think so | 15:25 |
| avass | we've set it up so that if the gate fails it sets -2 verified, and allows you to comment 'regate' to rerun the gate if it has a -2 verified | 15:26 |
| avass | so you can skipp the check pipeline and just run the gate directly | 15:26 |
| avass | but this depends a lot on how things are set up | 15:26 |
| avass | I believe opendev just requires verified +2 workflow +1 for the gate to start | 15:27 |
| avass | code-review, not verified | 15:27 |
| sassyn | yep | 15:27 |
| sassyn | OK | 15:27 |
| avass | and that you can retrigger the gate with a comment | 15:28 |
| sassyn | avass is use feature toggles? | 15:29 |
| sassyn | what is* | 15:30 |
| avass | sassyn: like preprocessing directives and things like that to turn parts of the software on and off easily :) | 15:30 |
| sassyn | can u explain more, not sure I got it | 15:31 |
| avass | sassyn: it's so you can turn off features that maybe are not yet ready. I guess one example is if you're preparing to use new hardware that has more features, instead of waiting to merge everything until you've got the new hardware version you can toggle off the features you know require the newer hardware. That way you can keep building and testing | 15:37 |
| avass | if that makes sense | 15:37 |
| sassyn | make sense but how to toggle ? | 15:37 |
| avass | it can either be regular 'if' statements so that code is just not executed, or preprocessing directives so the compilers removes that part of the code. | 15:38 |
| sassyn | OK | 15:39 |
| avass | note that you can also end if with a lot of dead code in your project if you do that :) | 15:39 |
| sassyn | understand now | 15:39 |
| avass | s/if/up | 15:40 |
| clarkb | feature flags are also useful to do partial deployments or A/B testing in production etc | 15:41 |
| avass | yeah | 15:41 |
| fungi | that's where coverage tracking can come in handy, to find dead branches of source code you've forgotten about | 15:41 |
| *** zxiiro has joined #zuul | 15:42 | |
| sassyn | thank u | 15:42 |
| sassyn | hi fungi | 15:42 |
| sassyn | how u are doing well:-) | 15:42 |
| fungi | oh yes, just in lots of (remote) meetings | 15:44 |
| *** bhavikdbavishi has joined #zuul | 15:51 | |
| *** hashar has joined #zuul | 15:51 | |
| *** bhavikdbavishi1 has joined #zuul | 15:54 | |
| *** bhavikdbavishi has quit IRC | 15:55 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 15:55 | |
| *** dpawlik has quit IRC | 15:59 | |
| AJaeger | zuul-jobs maintainer, please review https://review.opendev.org/724132 https://review.opendev.org/723643 https://review.opendev.org/725210 | 16:01 |
| avass | done :) | 16:02 |
| AJaeger | thanks tristanC and avass | 16:07 |
| *** rpittau is now known as rpittau|afk | 16:08 | |
| AJaeger | tristanC: the promotion change mentioned in https://review.opendev.org/#/c/724132 was merged last week shortly after mnaser gave his +1. | 16:08 |
| AJaeger | we wanted to decouple those two changes | 16:08 |
| openstackgerrit | Merged zuul/zuul-jobs master: go: Use 'block: ... always: ...' and failed_when instead of ignore_errors https://review.opendev.org/723643 | 16:15 |
| AJaeger | zuul-jobs maintainer, zbr has a change up for using yamllint which looks fine to me as defined - https://review.opendev.org/725091 Feedback welcome, please | 16:16 |
| openstackgerrit | Merged zuul/zuul-jobs master: ara-report: use failed_when: false instead of ignore_errors: true https://review.opendev.org/723644 | 16:17 |
| openstackgerrit | Merged zuul/zuul-jobs master: fetch-subunit-output: use failed_when: instead of ignore_errors: https://review.opendev.org/723653 | 16:20 |
| *** sassyn has quit IRC | 16:22 | |
| mordred | zbr, AJaeger : one small thing on 725091 - and then I left a note about followup | 16:23 |
| AJaeger | thanks, mordred and avass | 16:25 |
| corvus | if we want to do that, let's proceed with caution and make sure we're solving real problems and not making busy-work. | 16:28 |
| avass | yeah, I can agree to that | 16:29 |
| corvus | mordred: i turned on debug logging in buildkitd, and it looks like after HEADing the manifest from the buildset registry, it then decided that it should GET the underlying blob from registry-1.docker.io | 16:30 |
| corvus | mordred: i don't know why it decided it should get that blob from a different host | 16:30 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Do not use bare 'item' in build-container-image https://review.opendev.org/725298 | 16:30 |
| *** panda is now known as panda|pto | 16:35 | |
| *** evrardjp has quit IRC | 16:35 | |
| *** evrardjp has joined #zuul | 16:36 | |
| *** bhavikdbavishi has quit IRC | 16:36 | |
| mordred | corvus: that's really weird | 16:37 |
| corvus | my guess is that it's deciding to fallback to the upstream registry because it doesn't like something about the response from this "mirror" | 16:41 |
| corvus | even though i can see no errors reported and i think we're providing everything asked for | 16:41 |
| mordred | corvus: maybe we still arent' configuring mirrors properly? | 16:43 |
| corvus | mordred: it's talking ok up to that point | 16:43 |
| mordred | yeah | 16:43 |
| zbr | avass: AJaeger mordred: fixing now, also removing the "temporary" comment, as soon it is merged I can try to fix them | 16:45 |
| openstackgerrit | Sorin Sbarnea (zbr) proposed zuul/zuul-jobs master: Enable yamllint https://review.opendev.org/725091 | 16:48 |
| avass | tristanC: re 723654 | 16:50 |
| corvus | webknjaz: i think we may have disproven the theory that our web server memory leak is due to newest cherrypy; we still have a mem leak, but it doesn't seem related to the cherrypy upgrade. | 16:52 |
| corvus | so there must be some other change that happened that we don't understand yet | 16:53 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: Support multi-arch image builds with docker buildx https://review.opendev.org/722339 | 16:58 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: DNM Check to see if images from intermediate work https://review.opendev.org/724751 | 16:58 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: Write a buildkitd config file pointing to buildset registry https://review.opendev.org/724757 | 16:58 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: Write buildkitd.toml in use-buildset-registry https://review.opendev.org/724837 | 16:58 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: Add zuul labels to images and ability to user-define more https://review.opendev.org/725339 | 16:58 |
| *** jpena is now known as jpena|off | 17:04 | |
| *** zenkuro has joined #zuul | 17:06 | |
| zenkuro | hi is there a way to increase timeout for job for longer then 7 hours? | 17:07 |
| *** y2kenny has joined #zuul | 17:07 | |
| openstackgerrit | Clark Boylan proposed zuul/zuul master: Don't exit zuul-web when signals are received https://review.opendev.org/724946 | 17:08 |
| clarkb | tobiash: ^ is that what you had in mind for the signal handling? I think that is a lot cleaner. THanks | 17:08 |
| clarkb | zenkuro: Zuul allows jobs to set arbitrary timeout values but then that can be limited by the zuul admins via a global (or maybe tenant) max timeout | 17:09 |
| mordred | corvus: I just had a dumb idea | 17:11 |
| clarkb | zenkuro: that means the answer is "yes" but it may depend on what your zuul admins have set the limits to | 17:13 |
| zenkuro | clarkb: in other words if my manual timeout fails I have to check general zuul conf | 17:13 |
| clarkb | zenkuro: yes, you could be externally limited | 17:13 |
| zenkuro | by manual timeout a mean my job config fails on basic validation | 17:14 |
| mordred | corvus: we have a mirror configured for docker.io right? do we need to configure one for registry-1.docker.io too? | 17:15 |
| corvus | mordred: i don't think so | 17:16 |
| clarkb | zenkuro: are you able to share that failure? | 17:16 |
| zenkuro | nope... it is config failure on software factory, and for some reasons Im not able to see logs on it... | 17:17 |
| zenkuro | and I havent configured sf from scratch to post logs =( | 17:17 |
| clarkb | there are other reasons a simple job update could fail | 17:18 |
| clarkb | (and I don't think setting a value beyond the limit is a failure you'll just get capped instead. But I'm not sure of that) | 17:18 |
| corvus | mordred: wow | 17:24 |
| corvus | mordred: zuul-registry needs to *not* return the digest header if it gets a HEAD on the manifest | 17:24 |
| y2kenny | just a quick note/observation (I am not sure if there's a better way)... when the build job exceed the disk_limit_per_job, zuul seems to just keep retrying without failing (up to 41 attempts before I noticed it.) (I left the limit to default 250 when the kernel source size is like 1+ GB.) I ended up finding the executor responsible and look at | 17:25 |
| y2kenny | the executor log to see the warning. | 17:25 |
| corvus | y2kenny: zuul should have a default of 3 attempts for each job: https://zuul-ci.org/docs/zuul/reference/job_def.html#attr-job.attempts | 17:26 |
| y2kenny | corvus: I remember things failing after 3 attempts for other type of failure but this one didn't for some reason | 17:27 |
| zenkuro | clarkb: ive got failure by increasing timeout from 3600 to 36000 | 17:27 |
| corvus | y2kenny: hrm, maybe it's failing early enough that doesn't catch it; thanks for the report | 17:28 |
| clarkb | zenkuro: if you simply update the job description (for that same job) do you also get a failure? | 17:28 |
| corvus | mordred: i believe that's a spec violation, but i'm guessing they coded buildkit to the docker.io implementation, and the implementation doesn't return that header because of their weird cloudflare offloading. | 17:29 |
| y2kenny | corvus: the failure is at the executor level I think... before job getting run. first checkout of the repository and all that. | 17:29 |
| avass | corvus: I've seen zuul retrying past 3 tries if pywinrm isn't installed, not sure if that has been fixed yet though | 17:30 |
| zenkuro | clarkb: nope. I will try again now and upfate you | 17:30 |
| corvus | avass: ah, that's also tied into executor behaviors, that could also be a very early failure like y2kenny describes | 17:31 |
| mordred | corvus: wow | 17:32 |
| avass | corvus: I could probably take a look at that later in the week | 17:33 |
| openstackgerrit | James E. Blair proposed zuul/zuul-registry master: Don't return Docker-Content-Digest header on HEAD https://review.opendev.org/725350 | 17:33 |
| corvus | mordred: ^ i think that should do it | 17:34 |
| corvus | mordred: i edited the registry in-place on the test node to discover/verify that | 17:34 |
| corvus | mordred: i have found this sequence useful for experimenting with both the registry and the buildkit config: 'docker exec <do something to modify files>'; docker commit ...; docker restart ...; | 17:35 |
| mordred | corvus: I feel like we should file a bug somewher e- but I'm not 100% sure where | 17:35 |
| corvus | that's easier than reconstructing the corresponding containers from scratch | 17:35 |
| mordred | corvus: oh! good point | 17:35 |
| corvus | mordred: definitely in docker. or moby. or buildkit. or containerd. | 17:35 |
| tobiash | clarkb: yes | 17:35 |
| mordred | clarkb, tobiash: have a sec for a +A on 725350 from corvus there? | 17:36 |
| clarkb | mordred: looking | 17:37 |
| *** jcapitao has quit IRC | 17:52 | |
| openstackgerrit | Merged zuul/zuul-registry master: Don't return Docker-Content-Digest header on HEAD https://review.opendev.org/725350 | 17:53 |
| zbr | there is one change I want to make to tox role, to make it split TOXENV and loop over each targe, so we have different task for each environment. | 17:57 |
| zenkuro | clarkb: yes, Ive rechecked. adding 0 to 3600 cause config error | 17:57 |
| zbr | this should fix the problem of confusing errors for those few calling it with multiple environments like TOXENV=foo,bar,zoo | 17:58 |
| zbr | if foo fails, tox will run the others and finding the error in the console would be a tremendous task, example: https://dashboard.zuul.ansible.com/t/ansible/build/c975b2639a144d4b82f8158bb29dc65b | 17:58 |
| zbr | anyone against this idea? | 17:59 |
| clarkb | zenkuro: in that case I guess we don't silently cap and do error. You'll need to see what your global limit is I guess | 17:59 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: Write buildkitd.toml in use-buildset-registry https://review.opendev.org/724837 | 18:04 |
| *** sshnaidm is now known as sshnaidm|afk | 18:07 | |
| zenkuro | hm Ive found timeout for basejob, but increasig it resulted in a failure | 18:09 |
| clarkb | zenkuro: the base job tiemout is a default not a cap. Let me find the config option to cap it | 18:09 |
| clarkb | zenkuro: https://zuul-ci.org/docs/zuul/reference/tenants.html#attr-tenant.max-job-timeout | 18:10 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: merger: warn about invalid object type https://review.opendev.org/725363 | 18:22 |
| AJaeger | zuul-jobs maintainers, please review https://review.opendev.org/725210 , https://review.opendev.org/724132 , https://review.opendev.org/#/c/724840/ | 18:35 |
| zenkuro | clarkb: thanks clarkb, looks like it worked | 18:38 |
| openstackgerrit | Merged zuul/zuul master: Don't exit zuul-web when signals are received https://review.opendev.org/724946 | 18:38 |
| zenkuro | sorry for dumb questions, havent familiarized myself with zuul yet | 18:39 |
| corvus | zenkuro: no apology necessary :) | 18:55 |
| corvus | mordred: iiuc, 724757 being green now means you can squash 724757, 724751, 722339 together and we should be good to merge? | 18:55 |
| mordred | corvus: the multiarch job is green!!! | 18:55 |
| mordred | yup | 18:55 |
| corvus | mordred: neat-oh :) | 18:56 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: Support multi-arch image builds with docker buildx https://review.opendev.org/722339 | 18:57 |
| mordred | corvus: thereyago | 18:57 |
| mordred | clarkb: if you have a sec, ^^ that implements multi-arch building so should let us update nb03 with docker containers | 18:59 |
| clarkb | mordred: I will in about 10-15 minutes | 18:59 |
| mordred | clarkb: and I think we're about as confident as we could be about it working | 18:59 |
| mordred | tobiash: ^^ you might want to look too? (or you might not - I don't know if multi-arch docker images are important to you at all) | 18:59 |
| mordred | corvus: actually - let me clean something up real quick | 19:01 |
| mordred | corvus: there is an entry in the job vars that doesn't work (and never has) | 19:01 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: Support multi-arch image builds with docker buildx https://review.opendev.org/722339 | 19:02 |
| mordred | corvus, clarkb, tobiash: ^^ removed cruft from job definition | 19:03 |
| mordred | mnaser: ^^ multi-arch docker builds | 19:03 |
| *** yolanda has quit IRC | 19:16 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Add zuul labels to images and ability to user-define more https://review.opendev.org/725339 | 19:19 |
| *** y2kenny has quit IRC | 19:20 | |
| *** cmurphy has left #zuul | 19:27 | |
| AJaeger | mordred: could you check https://review.opendev.org/725298 , please? | 19:27 |
| AJaeger | mnaser: can we merge https://review.opendev.org/724132 now? | 19:28 |
| clarkb | mordred: in the pull and tag at https://review.opendev.org/#/c/722339/25/roles/build-docker-image/tasks/buildx.yaml does that pull and tag for all built arches? | 19:28 |
| clarkb | mordred: and does the build tag not apply at all? because we add tags on the build command line | 19:30 |
| mordred | clarkb: the second tag is probably not needed - the pull pulls for the arch that the node is running on | 19:33 |
| mordred | clarkb: the pull basically ensures that after a build on that host, normal docker commands can work | 19:33 |
| mordred | clarkb: the buildx sequence is inverted - basically instead of "build, tag, push" which results in the images being both in the local cache and the buildset registry, we do "build+tag+push, pull" which results in the images being in both the local cache and the buildset registry | 19:34 |
| clarkb | mordred: oh gotcha so if there is overlap between target and build host then the imgae will be present for use on the build host | 19:35 |
| clarkb | mordred: so we can remove the last task there to do the tag? | 19:35 |
| mordred | oh - no - the second tag is necessary | 19:35 |
| mordred | the second tag is to make sure that in the local cache the image shows up as the target production tag | 19:35 |
| mordred | we tag it only with the buildset registry name in teh buildx command so that it pushes it there | 19:35 |
| mordred | in normal docker we tag as the target tag during build, then re-tag with the buildset registry tag so that we can push the refs there | 19:36 |
| clarkb | mordred: lines 18 and 40 look the same though? | 19:36 |
| mordred | clarkb: 40 is tagging buidset_registry/repo/tag as repo/tag | 19:36 |
| mordred | clarkb: 18 is tagging the image as buildset_regigstry/repo/tag | 19:37 |
| clarkb | I see the suffix on 40 is important | 19:37 |
| mordred | prefix - but yes | 19:37 |
| mordred | you know - we could PROBABLY just pull it as the correct name and skip the re-tagging - but I'd want to do that as a followup | 19:38 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: Don't pull and retag in buildx workflow https://review.opendev.org/725380 | 19:39 |
| mordred | clarkb, corvus : ^^ I *think* that should also work | 19:40 |
| clarkb | mordred: why dod we use a specific sha for docker/binfmt? | 19:46 |
| mordred | clarkb: there is no general tag | 19:47 |
| mordred | clarkb: they strongly recommend using explicit shas | 19:47 |
| mordred | clarkb: also - is your brain bleeding yet? | 19:47 |
| fungi | i feel like these people are all rediscovering problems which got solves decades ago | 19:48 |
| mordred | fungi: but nobody solved them in golang decades ago | 19:48 |
| clarkb | mordred: I'm enjoying that ipv6 continues to hate us | 19:49 |
| *** lseki has joined #zuul | 19:51 | |
| clarkb | mordred: do we want to keep the zuul-registry -d flag? I think that is for debug output? | 19:51 |
| clarkb | I left that as a comment on the change | 19:56 |
| mordred | clarkb: yeah - corvus commented on that on a previous ps - I think it's a good idea - helps with debugging | 19:57 |
| clarkb | zuulians https://review.opendev.org/#/c/723764/ is a somewhat minor gerrit driver update, but it helps cut down on tracebacks that are expected and the user can't do much about | 20:06 |
| clarkb | and that in turn aids debugging when you are trying to skim logs to find unhappy things | 20:06 |
| mordred | clarkb: are you sure that cuts down on tracebacks? | 20:14 |
| mordred | clarkb: you're using log.exception | 20:14 |
| clarkb | mordred: ya but we don't retry | 20:14 |
| mordred | ah - nod | 20:14 |
| clarkb | mordred: before we were getting 3 of them | 20:14 |
| clarkb | we should only get one now | 20:14 |
| mordred | do we want the traceback at all? | 20:14 |
| clarkb | mordred: maybe not. I could see that being log.info | 20:15 |
| mordred | seems like we could go to log.error (or log.info) - it's not an exception in the code - we'v ehandled it, it's not a thing the admin needs to worry about | 20:15 |
| mordred | I mean- I think we can land this | 20:15 |
| mordred | just saying - maybe we want to further clean up | 20:15 |
| clarkb | ++ | 20:15 |
| avass | mordred: I'm going to nit at you not generating tempfiles in 722339 :) | 20:19 |
| mordred | avass: oh no | 20:24 |
| mordred | avass: I was hoping nobody would notice that | 20:24 |
| mordred | avass: mind if we fix it in a followup? I'll work on one right now | 20:25 |
| avass | mordred: I was planning on doing that after the merge :) | 20:25 |
| avass | it doesn't really matter that much | 20:26 |
| mordred | avass: cool :) | 20:28 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-jobs master: Use tempfile in buildx build https://review.opendev.org/725387 | 20:29 |
| mordred | avass: ++ | 20:29 |
| avass | I hardly believe someone without a /tmp is going to start using the multiarch feature soon anyway ;) | 20:30 |
| mordred | avass: never say never | 20:33 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: ansible-lint: use matchplay instead of matchtask https://review.opendev.org/724910 | 20:33 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: use zj_image instead of image as loopvar https://review.opendev.org/725012 | 20:33 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: use zj_log_file instead of item as loop_var https://review.opendev.org/725013 | 20:33 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Check blocks recursively for loops https://review.opendev.org/724967 | 20:34 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Update ansible-lint-rules testsuite to only test with the relevant rule https://review.opendev.org/725014 | 20:34 |
| *** threestrands has joined #zuul | 20:34 | |
| avass | ^ I also pushed a patch to update matchplay in ansible-lint to return linenumber since at the moment it give you the line the block starts at | 20:36 |
| avass | but that stack should be pretty ready | 20:36 |
| *** y2kenny has joined #zuul | 21:00 | |
| y2kenny | Is this https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/fetch-output-openshift/tasks/rsync.yaml#L4 suppose to be updated to zj_pod.1.context instead of item.1.context? | 21:00 |
| avass | y2kenny: oh no, YES | 21:01 |
| y2kenny | ok... I was going crazy wondering why things broke | 21:02 |
| y2kenny | and the no_log from is not working for some reason (even though I set the verbose to True) | 21:03 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: fetch-logs-openshift: fix miss when replacing item with loop_var: zj_ https://review.opendev.org/725392 | 21:04 |
| avass | zuul-jobs-maint: ^ | 21:05 |
| y2kenny | should be zj_pod right? | 21:06 |
| y2kenny | not zj_output? | 21:06 |
| clarkb | y2kenny: no the loop var below is zj_output | 21:06 |
| y2kenny | but zj_output only have src and dst? | 21:06 |
| clarkb | hrm ya | 21:07 |
| y2kenny | the namespace and context is from the loop in main.yaml? | 21:07 |
| clarkb | I think you are right | 21:07 |
| clarkb | y2kenny: updated my review. Thank you for pointing that out | 21:08 |
| avass | oh, yeah you're right | 21:08 |
| y2kenny | I would've comment directly but I couldn't find my login for the moment | 21:08 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: fetch-logs-openshift: fix miss when replacing item with loop_var: zj_ https://review.opendev.org/725392 | 21:08 |
| avass | try number 2 then | 21:08 |
| avass | y2kenny: thanks! | 21:09 |
| y2kenny | avass: does this go live immediately on submit? | 21:10 |
| avass | y2kenny: once it's merged yup | 21:10 |
| y2kenny | ok thanks | 21:11 |
| avass | clarkb: can make sure that change is merged? It's getting late here | 21:13 |
| openstackgerrit | Merged zuul/zuul master: Don't retry after Gerrit HTTP 409s https://review.opendev.org/723764 | 21:13 |
| clarkb | avass: I can try! | 21:13 |
| avass | clarkb: thanks :) | 21:14 |
| *** Goneri has quit IRC | 21:25 | |
| *** rfolco has quit IRC | 21:27 | |
| *** jamesmcarthur has joined #zuul | 21:30 | |
| jamesmcarthur | Hi Zuul team. Just want to give you a heads up re: our user survey results. We had 18 very solid responses, but we're still planning to do anotehr marketing push to grab more. | 21:30 |
| jamesmcarthur | The great news is, these are excellent companies, many who have opted to share their results publicly. | 21:31 |
| jamesmcarthur | I just sent an email following up with each of the "public" group (10 total that are new) and asked if they could submit a bit more info for a use case. | 21:31 |
| jamesmcarthur | Once we comb through those responses, we'll do follow ups as necessary and plan to publish use cases on the Zuul website via PR. | 21:32 |
| jamesmcarthur | Really excited about this :) And if anyone is interested, here are some of the big new names: oVirt, Wazo, T-Systems, CESNET, Good Money, and Just Eat | 21:33 |
| jamesmcarthur | Congrats to you all on making compelling and usable software! | 21:33 |
| *** sanjayu_ has quit IRC | 21:37 | |
| clarkb | mordred: corvus fungi ianw if any of you are still around can you look at https://review.opendev.org/#/c/725392/2 to fix an issue y2kenny found? | 21:41 |
| fungi | thanks for the deets jamesmcarthur! | 21:54 |
| fungi | clarkb: sure, i can take a peek here in a bit | 21:54 |
| clarkb | fungi: thanks! | 21:54 |
| *** y2kenny has quit IRC | 21:55 | |
| *** rfolco has joined #zuul | 22:03 | |
| corvus | clarkb: done | 22:07 |
| corvus | jamesmcarthur: awesome, thanks! | 22:08 |
| jamesmcarthur | yes! Thank you all :) | 22:08 |
| corvus | jamesmcarthur: i just thought of something -- maybe we should put a link to it at the end of the zuul quickstart | 22:10 |
| corvus | jamesmcarthur: so it's like "you just got your first zuul up and running! congrats! we'd like to hear from you" | 22:10 |
| jamesmcarthur | Oh... I love that. | 22:11 |
| jamesmcarthur | great idea corvus | 22:11 |
| *** tobiash has quit IRC | 22:13 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add a link to the user survey to the end of the quickstart. https://review.opendev.org/725415 | 22:15 |
| corvus | hah that is not the url | 22:15 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add a link to the user survey to the end of the quickstart. https://review.opendev.org/725415 | 22:15 |
| corvus | jamesmcarthur: ^ there we go | 22:15 |
| jamesmcarthur | :) | 22:17 |
| openstackgerrit | Merged zuul/zuul-jobs master: fetch-logs-openshift: fix miss when replacing item with loop_var: zj_ https://review.opendev.org/725392 | 22:20 |
| mordred | corvus: https://review.opendev.org/#/c/722339/ has 2x+2 but clarkb left it open in case other people wanted to review | 22:26 |
| *** avass has quit IRC | 22:35 | |
| *** hashar has quit IRC | 22:42 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: use python2-pip when running under Python 2 https://review.opendev.org/724777 | 22:54 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: Install backported pip for Xenial https://review.opendev.org/724788 | 22:54 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Add plain nodes to testing https://review.opendev.org/724776 | 22:54 |
| openstackgerrit | Merged zuul/nodepool master: Handle ipv6 literal zookeeper addresses https://review.opendev.org/725157 | 23:10 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: use python2-pip when running under Python 2 https://review.opendev.org/724777 | 23:10 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: Install backported pip for Xenial https://review.opendev.org/724788 | 23:10 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Add plain nodes to testing https://review.opendev.org/724776 | 23:10 |
| *** tosky has quit IRC | 23:28 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: use python2-pip when running under Python 2 https://review.opendev.org/724777 | 23:38 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: Install backported pip for Xenial https://review.opendev.org/724788 | 23:38 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Add plain nodes to testing https://review.opendev.org/724776 | 23:38 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!