| *** harlowja has quit IRC | 00:26 | |
| tristanC | corvus: nice! | 01:31 |
|---|---|---|
| *** yolanda has quit IRC | 02:03 | |
| *** jhesketh_ is now known as jhesketh | 03:14 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul-jobs master: [WIP] kerberos & afs client roles https://review.openstack.org/589334 | 04:21 |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul-jobs master: [WIP] kerberos & afs client roles https://review.openstack.org/589334 | 04:51 |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul-jobs master: [WIP] kerberos & afs client roles https://review.openstack.org/589334 | 05:38 |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul-jobs master: [WIP] kerberos & afs client roles https://review.openstack.org/589334 | 05:44 |
| * SpamapS really wishes he'd had time to swap out the spinning disk on his new laptop for an SSD before he left for europe | 05:53 | |
| SpamapS | :-P | 05:53 |
| *** pcaruana has joined #zuul | 06:36 | |
| SpamapS | :( and now pbrx doesn't work | 06:37 |
| SpamapS | http://paste.openstack.org/show/727462/ | 06:37 |
| *** yolanda has joined #zuul | 06:41 | |
| SpamapS | OH | 06:43 |
| SpamapS | this doesn't work on a remote docker | 06:43 |
| SpamapS | :( | 06:43 |
| SpamapS | naughty pbrx needs to COPY rather than -v. :-/ | 06:44 |
| *** quiquell has joined #zuul | 06:46 | |
| quiquell | Hello there | 06:46 |
| quiquell | I am checking job's 'file' and 'irrelevant-files' option | 06:47 |
| quiquell | Is it possible to express them in a external file, so we can reuse the lists ? | 06:47 |
| SpamapS | quiquell: I don't think so. I was needing something similar the other day, wanting to build the list from some rules. I think it would need to be added as a new feature. | 06:54 |
| quiquell | SpamapS: Ok, going to check, thanks man. | 06:55 |
| SpamapS | mordred: the way you've used volumes makes it impossible to use a remote docker daemon. It's pretty common to do that, in my case, trying to simplify things by using minikube's docker so I don't have to figure out pushing to minikube's registry. :-P | 07:23 |
| SpamapS | I think we can do it all with docker cp though. | 07:23 |
| SpamapS | which is the CLI equivalent of COPY | 07:24 |
| *** jpena|off is now known as jpena | 07:35 | |
| *** electrofelix has joined #zuul | 08:33 | |
| *** panda|rover-ish is now known as panda|rover | 08:33 | |
| *** gtema has joined #zuul | 08:58 | |
| *** eumel8 has joined #zuul | 10:03 | |
| *** electrofelix has quit IRC | 10:18 | |
| *** electrofelix has joined #zuul | 10:26 | |
| *** goern has joined #zuul | 11:08 | |
| *** panda|rover is now known as panda|rover|lunc | 11:18 | |
| *** elyezer has quit IRC | 12:03 | |
| *** elyezer has joined #zuul | 12:06 | |
| *** chmouel has quit IRC | 12:15 | |
| *** jpena is now known as jpena|lunch | 12:15 | |
| *** panda|rover|lunc is now known as panda|rover | 12:16 | |
| mordred | SpamapS: yes - I was explicity avoiding copy | 12:20 |
| mordred | SpamapS: because copying that data in is just annoyingly inefficient | 12:21 |
| mordred | SpamapS: however, let's see if we can't work out a solution, cause I'd like for you to be able to use a remote docker if that's important | 12:22 |
| mordred | SpamapS: maybe a flag that'll use cp instead of volume for cases when someone is trying to do that? that seems potentially reasonable? | 12:23 |
| *** gtema has quit IRC | 12:47 | |
| *** sdoran_ has joined #zuul | 12:49 | |
| *** sdoran has quit IRC | 12:50 | |
| *** pbrobinson has quit IRC | 12:50 | |
| *** weshay has quit IRC | 12:50 | |
| *** sdoran_ is now known as sdoran | 12:50 | |
| *** samccann has joined #zuul | 12:51 | |
| Shrews | morning zuulers | 13:01 |
| *** jpena|lunch is now known as jpena | 13:30 | |
| AJaeger | mordred, tristanC , want to rebase https://review.openstack.org/#/c/550979/ ? | 13:45 |
| *** quiquell is now known as quiquell|off | 13:47 | |
| *** quiquell|off has quit IRC | 13:51 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/projects and /{tenant}/project/{project} routes https://review.openstack.org/550979 | 13:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/pipelines route https://review.openstack.org/541521 | 13:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: scheduler: add job's parent name to the rpc job_list method https://review.openstack.org/573473 | 13:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/labels route https://review.openstack.org/553979 | 13:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/nodes route https://review.openstack.org/553998 | 13:56 |
| mordred | AJaeger: done | 13:56 |
| mordred | Shrews: morning! | 13:56 |
| AJaeger | great, mordred | 13:56 |
| corvus | SpamapS: re quiquell's question -- you can use yaml anchors to re-use lists | 14:14 |
| *** yolanda has quit IRC | 14:31 | |
| openstackgerrit | Merged openstack-infra/nodepool master: Switch storyboard url to be by name https://review.openstack.org/588600 | 14:55 |
| *** panda|rover is now known as panda|backin2h | 15:01 | |
| *** pcaruana has quit IRC | 15:11 | |
| SpamapS | corvus: that's a good point, I hadn't thought of doing that. My need was to actually programmatically share a list with another system. | 15:24 |
| SpamapS | mordred: so, the use case is "I have git state over here in my dev workstation, and docker over there in my dev VM. I think it may be pretty common, especially for mac users. | 15:25 |
| mordred | SpamapS: sure - totally understand the existence of the use case | 15:26 |
| SpamapS | And in this case, I want to just build the containers in the docker that minikube uses directly to avoid a second step of pushing to the registry inside minikube. | 15:26 |
| mordred | SpamapS: mostly just want to make sure we solve it in a way that doesn't also make it more expensive for the other case | 15:26 |
| SpamapS | I wonder if there's another way though. I mean, we already have to put the source for these python programs into the container image at some point.. | 15:27 |
| mordred | we don't though | 15:27 |
| mordred | that's just the thing | 15:27 |
| SpamapS | maybe if we ship the bindep.txt we can just read it from the wheel? | 15:27 |
| SpamapS | oh, we don't? | 15:27 |
| mordred | we need to put the unpacked wheel into the container image | 15:27 |
| SpamapS | also side note.. is there a way to make all the moron random account msg spam go away? | 15:28 |
| mordred | the source code itself does not need to be in the image | 15:28 |
| SpamapS | Ok well I meant the same thing. :) | 15:28 |
| mordred | SpamapS: yes - /mode SpamapS +R | 15:28 |
| SpamapS | mordred: ty | 15:28 |
| mordred | SpamapS: I mean, the distinction is splitting hairs, as we DO need the source code to be in the builder image | 15:28 |
| mordred | or, more specifically, we need the source git repo to be in the builder image | 15:29 |
| mordred | because we need that to build the wheel | 15:29 |
| mordred | so copy vs. volume is a thing regardless | 15:29 |
| SpamapS | Right, just.. doing it with volumes instead of copies.. TBH.. I've been iterating on the aws provider and docker build using COPY and RUN ends up caching more aggressively and more efficiently than pbrx... so I'm not sure what the aversion to COPY is. | 15:31 |
| mordred | well - there's two sides to the coin | 15:31 |
| SpamapS | Maybe there's another volume we can mount in that has a wheel cache or something? | 15:31 |
| mordred | the first is just it drives me batty to have to copy an entire git repo to a second directory on my laptop to build a container image | 15:31 |
| mordred | it makes local iteration massively slower | 15:32 |
| mordred | BUT - the second is what you just mentioned ... | 15:32 |
| mordred | pulling out the wheel cache and the built artifacts | 15:32 |
| mordred | that said - I'm sure we can make a version of that that uses copy for all of those things - it shouldn't be super hard | 15:33 |
| SpamapS | mordred: yeah, I'm also sort of trying to figure out why it's so hard to just login and push to this silly minikube registry, which may be a better option ultimately. | 15:46 |
| mordred | SpamapS: :) | 15:46 |
| tristanC | SpamapS: maybe configure minikube to pull the image from the host docker? | 15:52 |
| SpamapS | tristanC: I hadn't thought of that. Seems complicated. | 15:55 |
| SpamapS | since the host docker isn't a registry | 15:55 |
| *** pbrobinson has joined #zuul | 15:59 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Use full gerrit changeid in HTTP report https://review.openstack.org/589530 | 16:04 |
| corvus | clarkb, mordred: could you review that with some urgency? | 16:05 |
| clarkb | yes | 16:05 |
| clarkb | corvus: does change.project need to be urlencoded? | 16:07 |
| corvus | clarkb: lemme check | 16:07 |
| corvus | yes | 16:08 |
| corvus | or at least, that's what gertty sends | 16:09 |
| clarkb | just noticing it is going into a path so we can't encode it later | 16:09 |
| corvus | yeah, gertty receives the value urlencoded and never decodes it, which is why i didn't notice that | 16:10 |
| corvus | but since we're generating it, we should do that in order to match | 16:10 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Use full gerrit changeid in HTTP report https://review.openstack.org/589530 | 16:15 |
| corvus | clarkb, mordred: ^ | 16:15 |
| corvus | fungi: ^ | 16:15 |
| fungi | oh, yeah i didn't even consider that | 16:16 |
| fungi | good eye! | 16:17 |
| fungi | change.project will certainly contain risky characters (at least / in the case of the openstack community deployment, but more generally who knows what other characters people might put in the names of their project in different gerrits) | 16:18 |
| clarkb | actually branches too right? | 16:18 |
| clarkb | stable/foo ? | 16:18 |
| fungi | yep! | 16:19 |
| clarkb | I just +2'd but now realize that branch might need the same treatment | 16:19 |
| * fungi was just thinking about that as well looking at the interdiff | 16:19 | |
| mordred | corvus: +2 from me | 16:19 |
| mordred | oh - except what clarkb and fungi said | 16:20 |
| *** jpena is now known as jpena|off | 16:21 | |
| corvus | i guess we can quote the whole thing? | 16:22 |
| corvus | i believe ~ symbols make it through unscathed | 16:22 |
| clarkb | corvus: ++ | 16:22 |
| corvus | though, hrm, i guess if ~ ever ended up in a project name... | 16:23 |
| corvus | i'll just quote each part | 16:23 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Use full gerrit changeid in HTTP report https://review.openstack.org/589530 | 16:24 |
| corvus | clarkb, fungi, mordred: third time's the charm | 16:24 |
| mordred | corvus: +2 | 16:24 |
| fungi | leaving the approval for clarkb | 16:25 |
| * clarkb looking now | 16:25 | |
| clarkb | approved | 16:26 |
| *** electrofelix has quit IRC | 16:29 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-website master: Add BMW case study https://review.openstack.org/589541 | 16:37 |
| corvus | tobiash: ^ | 16:37 |
| tobiash | :) | 16:40 |
| tobiash | corvus: I have a problem with a branching use case: a project has more than one branch (and the jobs need to be branched) and we want create releases in a tag triggered release pipeline | 16:43 |
| tobiash | but the jobs won't execute as we have jobs with branch matchers to the branches but they don't match the tag | 16:44 |
| tobiash | I didn't find a possibility to realize this use case so far | 16:44 |
| tobiash | my current thinking is that we might need to create a speculative config from the tag so the tag has its own set of jobs | 16:45 |
| clarkb | tobiash: I think you can match arbitrary refs including tags (but yo umay not know what those tags will be in advance?) | 16:45 |
| tobiash | clarkb: but that would require explicit branch matcher right? | 16:46 |
| clarkb | tobiash: yes | 16:46 |
| tobiash | which would be an overhead during every branch creation | 16:47 |
| fungi | wasn't there a patch to make it possible to guess an appropriate branch for a given tag? | 16:47 |
| tobiash | fungi: I don't know if any | 16:47 |
| * fungi can't recall whether that merged, but it was a potential useful feature for openstack's deployment too | 16:48 | |
| fungi | or maybe corvus had merely outlined the logic needed for doing that mapping | 16:48 |
| tobiash | if we create a dynamic layout for a ref triggered pipeline that might just work as expected | 16:49 |
| tobiash | fungi: oh I think you mean https://review.openstack.org/578557 ? | 16:51 |
| fungi | tobiash: yep, that's the one | 16:51 |
| * tobiash overlooked that one | 16:52 | |
| corvus | tobiash: yeah, i was holding out for more user feedback on whether that was a good idea before i proceeded. sounds like i just got some. :) | 16:53 |
| corvus | tobiash: let me know what you think | 16:54 |
| tobiash | corvus: I'm just thinking which approach is better | 16:54 |
| fungi | agreed, it seemed like a less interesting addition if it was just needed because of openstack community process weirdness and nobody else actually used git that way | 16:55 |
| corvus | tobiash: i considered the speculative config approach, but it's a *lot* of work to ask zuul to do | 16:56 |
| corvus | (most of which wouldn't actually end up being necessary) | 16:56 |
| tobiash | corvus: ok, so you did this analysis already :) | 16:57 |
| tobiash | corvus: I thought that would be an easy approach, but if not, the branch guessing is probably the easiest | 16:57 |
| tobiash | corvus: the branch guessing also has the advantage of not running speculative code which is especially with github probably more secure | 16:58 |
| tobiash | so I like that approach | 16:59 |
| *** panda|backin2h is now known as panda|rover | 17:08 | |
| tobiash | corvus: added a comment to 578557 | 17:10 |
| tobiash | corvus: but I think that looks good otherwise | 17:12 |
| *** harlowja has joined #zuul | 17:31 | |
| *** harlowja has quit IRC | 17:43 | |
| openstackgerrit | Merged openstack-infra/zuul master: Use full gerrit changeid in HTTP report https://review.openstack.org/589530 | 18:01 |
| openstackgerrit | Merged openstack-infra/zuul-website master: Add BMW case study https://review.openstack.org/589541 | 18:11 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: Bump minimum openstacksdk version to 0.17.2 https://review.openstack.org/589563 | 18:15 |
| Shrews | mordred: clarkb: ^^^ | 18:16 |
| Shrews | though, hrm... will that work if the tag is only on the rocky branch? | 18:16 |
| mordred | Shrews: yes | 18:17 |
| Shrews | wrt to tests, that is | 18:18 |
| mordred | it should - we don't follow upper-constraints | 18:18 |
| mordred | although - I think we are currently accidentally following upper-constraints in our nodepool/devstack jobs | 18:18 |
| mordred | so it might not | 18:18 |
| Shrews | i guess we'll find out soon enough | 18:19 |
| *** zxiiro has joined #zuul | 18:29 | |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Un-wip upload-logs-swift https://review.openstack.org/588677 | 18:34 |
| *** samccann has quit IRC | 18:38 | |
| *** jimi|ansible has joined #zuul | 18:46 | |
| *** openstackgerrit has quit IRC | 18:49 | |
| *** openstackgerrit has joined #zuul | 20:13 | |
| openstackgerrit | Merged openstack-infra/zuul master: Cache decryption results https://review.openstack.org/582794 | 20:13 |
| openstackgerrit | Merged openstack-infra/zuul master: Switch storyboard url to be by name https://review.openstack.org/588597 | 20:13 |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul-jobs master: [WIP] kerberos & afs client roles https://review.openstack.org/589334 | 20:27 |
| *** jimi|ansible has quit IRC | 20:40 | |
| -openstackstatus- NOTICE: Due to a bug, Zuul has been unable to report on cherry-picked changes over the last 24 hours. This has now been fixed; if you encounter a cherry-picked change missing its results (or was unable to merge), please recheck now. | 20:44 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: WIP Extract pep8 messages for inline comments https://review.openstack.org/589634 | 20:55 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: WIP Extract pep8 messages for inline comments https://review.openstack.org/589634 | 21:10 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: WIP Extract pep8 messages for inline comments https://review.openstack.org/589634 | 21:15 |
| *** goern has quit IRC | 21:21 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul-jobs master: [WIP] kerberos & afs client roles https://review.openstack.org/589334 | 22:38 |
| *** gothicmindfood has quit IRC | 22:42 | |
| corvus | tobiash, clarkb: i just attempted and gave up on 2 implementations of a fix for the branch cache problem. i think that the best solution is going to be to have the drivers handle branch caching internally and transparently, removing any attempts at caching in the configloader or model layers. i'll work on that tomorrow. | 23:07 |
| corvus | tristanC: openstack is having a problem with periodic jobs not running. i've confirmed that's because the timer trigger is never firing for them. i suspect a bug in this change: https://review.openstack.org/581553 but i haven't had time to dig further into it. | 23:09 |
| corvus | tristanC: just a guess: maybe we need to also look at project-templates? | 23:09 |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul-jobs master: [WIP] kerberos & afs client roles https://review.openstack.org/589334 | 23:22 |
| gundalow | Does Zuul set ENV for BRANCH/COMMIT/REPO/ETC? | 23:23 |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul-jobs master: [WIP] kerberos & afs client roles https://review.openstack.org/589334 | 23:36 |
| mordred | gundalow: zuul doesn't set env vars - it does provide a nice pile of info in ansible vars | 23:58 |
| mordred | gundalow: such as: http://logs.openstack.org/53/588553/3/check/tox-linters/df37fed/zuul-info/inventory.yaml | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!