| openstackgerrit | Merged zuul/zuul-jobs master: Fix python3 compat in tox siblings handling https://review.opendev.org/698335 | 00:00 |
|---|---|---|
| *** rf0lc0 has quit IRC | 00:10 | |
| *** igordc has quit IRC | 00:13 | |
| *** jamesmcarthur has joined #zuul | 00:21 | |
| openstackgerrit | Merged zuul/zuul master: Add upgrade note about ansible_python_interpreter https://review.opendev.org/698318 | 00:31 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: DNM test all the things on ansible 2.8 with python3 https://review.opendev.org/698344 | 00:34 |
| clarkb | corvus: pabelanger fungi ^ occurred to me that we can do that too | 00:34 |
| clarkb | an extra layer of sanity checking. Not sure how much coverage it will actually give us though | 00:34 |
| clarkb | oh wow that is a lot of jobs :) | 00:35 |
| *** jamesmcarthur has quit IRC | 00:39 | |
| corvus | clarkb: ha, that is many jobs | 01:06 |
| clarkb | looks like only gentoo multinode failed. Probably good to proceed with the plan | 01:30 |
| *** bhavikdbavishi has joined #zuul | 01:35 | |
| *** bhavikdbavishi1 has joined #zuul | 01:38 | |
| *** bhavikdbavishi has quit IRC | 01:40 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 01:40 | |
| *** rlandy|bbl has quit IRC | 01:49 | |
| *** sgw has joined #zuul | 01:50 | |
| *** sgw has quit IRC | 01:58 | |
| *** sgw has joined #zuul | 02:26 | |
| *** Petar_T has quit IRC | 02:29 | |
| *** sgw has quit IRC | 02:58 | |
| *** ianychoi has joined #zuul | 05:32 | |
| *** pcaruana has joined #zuul | 06:09 | |
| *** raukadah is now known as chkumar|ruck | 06:10 | |
| *** ianychoi_ has joined #zuul | 07:08 | |
| *** tosky has joined #zuul | 07:09 | |
| *** ianychoi has quit IRC | 07:11 | |
| *** jcapitao has joined #zuul | 07:50 | |
| *** hashar has joined #zuul | 07:51 | |
| *** sanjayu__ has joined #zuul | 07:57 | |
| *** pcaruana has quit IRC | 08:22 | |
| *** wznoinsk has quit IRC | 08:28 | |
| *** Petar_T has joined #zuul | 08:36 | |
| *** jpena|off is now known as jpena | 08:53 | |
| *** themroc has joined #zuul | 08:53 | |
| *** jangutter has quit IRC | 09:03 | |
| *** jangutter has joined #zuul | 09:04 | |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile: create APP_DIR https://review.opendev.org/693646 | 09:20 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile: install sudo for nodepool-builder https://review.opendev.org/694709 | 09:20 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Dockerfile: add DEBUG environment flag https://review.opendev.org/694845 | 09:20 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: Also build sibling container images https://review.opendev.org/697393 | 09:20 |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: [wip] move openstack testing to use containerised daemon https://review.opendev.org/693464 | 09:20 |
| *** jangutter_ has joined #zuul | 09:36 | |
| *** avass has joined #zuul | 09:37 | |
| *** jangutter has quit IRC | 09:39 | |
| *** ianychoi_ has quit IRC | 09:56 | |
| *** hashar has quit IRC | 10:15 | |
| *** arxcruz is now known as arxcruz|pto | 11:32 | |
| *** pcaruana has joined #zuul | 11:36 | |
| *** ianychoi has joined #zuul | 11:49 | |
| *** sshnaidm|afk is now known as sshnaidm | 11:52 | |
| *** jcapitao is now known as jcapitao|afk | 12:00 | |
| *** avass has quit IRC | 12:03 | |
| *** spsurya has joined #zuul | 12:05 | |
| *** rfolco has joined #zuul | 12:05 | |
| *** jpena is now known as jpena|lunch | 12:29 | |
| *** jangutter_ is now known as jangutter | 12:41 | |
| *** rlandy has joined #zuul | 13:19 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Add optional support for circular dependencies https://review.opendev.org/685354 | 13:20 |
| *** jcapitao|afk is now known as jcapitao | 13:20 | |
| *** jpena|lunch is now known as jpena | 13:30 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Add optional support for circular dependencies https://review.opendev.org/685354 | 13:33 |
| pabelanger | zuul-maint: https://review.opendev.org/698324/ is a trivial reno update, hopefully help others | 13:35 |
| *** sgw has joined #zuul | 14:02 | |
| *** sanjayu__ has quit IRC | 14:24 | |
| *** bhavikdbavishi has quit IRC | 14:28 | |
| *** bhavikdbavishi has joined #zuul | 14:29 | |
| *** jamesmcarthur has joined #zuul | 14:31 | |
| *** jamesmcarthur has quit IRC | 14:36 | |
| *** sanjayu__ has joined #zuul | 14:36 | |
| *** avass has joined #zuul | 14:38 | |
| *** jamesmcarthur has joined #zuul | 14:44 | |
| *** jlk has quit IRC | 14:52 | |
| *** jlk has joined #zuul | 14:52 | |
| *** jamesmcarthur has quit IRC | 15:06 | |
| *** Goneri has joined #zuul | 15:11 | |
| *** swest has quit IRC | 15:21 | |
| *** jamesmcarthur has joined #zuul | 15:25 | |
| *** sanjayu__ has quit IRC | 15:27 | |
| *** jamesmcarthur_ has joined #zuul | 15:27 | |
| *** jamesmcarthur has quit IRC | 15:31 | |
| *** bhavikdbavishi has quit IRC | 15:35 | |
| *** panda is now known as panda|bbl | 15:44 | |
| *** chkumar|ruck is now known as raukadah | 15:47 | |
| *** themroc has quit IRC | 16:20 | |
| clarkb | I've double checked that the jobs we were testing yesterday ran with ansible 2.8 and auto python interpreter | 16:27 |
| clarkb | I don't think I've got anything else to check, should be good to revert the 2.7 pin in !zuul and then tag release | 16:27 |
| *** jcapitao is now known as jcapitao|afk | 16:29 | |
| zbr | is there a way to use a single github pipeline that triggers some jobs on all PRs+comments but some jobs are triggered only by comments? | 16:39 |
| zbr | or the only solution is to split the pipeline in two? | 16:39 |
| zbr | current one looks like https://review.rdoproject.org/r/#/c/24032/2/zuul.d/github.yaml | 16:39 |
| zbr | and I want to enable triggering on PRs for some repos, but problably not for all. | 16:39 |
| corvus | zbr: yeah, 2 pipelines | 16:43 |
| zbr | corvus: ok, thanks. wanted to be sure i am not doing something stupid. | 16:46 |
| openstackgerrit | Merged zuul/zuul master: Add additional info for executor.merge_jobs release note https://review.opendev.org/698324 | 16:46 |
| *** avass has quit IRC | 17:12 | |
| *** panda|bbl is now known as panda | 17:20 | |
| *** rlandy is now known as rlandy|brb | 17:26 | |
| clarkb | corvus: let me know if I can help do anything else to make zuul release stuff go smoothly | 17:36 |
| corvus | clarkb: is there a change to unpin 2.7? | 17:37 |
| SpamapS | zbr: one trick I've used, btw, is to have a label that triggers each, and on success/fail use those labels to have one trigger the other. | 17:38 |
| clarkb | corvus: looking | 17:38 |
| clarkb | corvus: doesn't look like it. Let me push it up | 17:39 |
| SpamapS | I don't do this today, but I did play with a thing that would add the 'run-deep-checks' label to any PR that failed check, so that the dev would have more data when responding to the fails. | 17:39 |
| clarkb | corvus: do you think we should set the pin to 2.8 or just rely on zuul defaults? | 17:40 |
| corvus | clarkb: default | 17:40 |
| corvus | clarkb: can remove zuul 2.8 pin too | 17:40 |
| *** igordc has joined #zuul | 17:42 | |
| *** hashar has joined #zuul | 17:43 | |
| *** rlandy|brb is now known as rlandy | 17:50 | |
| *** jpena is now known as jpena|off | 18:06 | |
| *** reiterative has quit IRC | 18:13 | |
| *** igordc has quit IRC | 18:13 | |
| *** igordc has joined #zuul | 18:14 | |
| *** rlandy is now known as rlandy|brb | 18:18 | |
| *** igordc has quit IRC | 18:25 | |
| *** jamesmcarthur_ has quit IRC | 18:30 | |
| openstackgerrit | Matthieu Huin proposed zuul/zuul master: admin REST API: zuul-web integration https://review.opendev.org/643536 | 18:35 |
| *** mhu has joined #zuul | 18:35 | |
| clarkb | still waiting on ansible unpin to apply in production | 18:39 |
| clarkb | but I've been watching jobs going through waiting for that | 18:39 |
| *** tosky has quit IRC | 18:44 | |
| corvus | mordred, Shrews: the googlecloudapi client is not thread-safe, plus we sort of have this idea in nodepool that maybe we would like to serialize requests to be nice to providers. i feel like we should implement the taskmanager we once did with the openstack provider for the google provider. and moreover, why not for aws and azure and ... | 18:44 |
| corvus | mordred, Shrews: the obvious implecation of that is that openstack would be the odd one out since it's the only one with its own taskmanager | 18:46 |
| corvus | mordred, Shrews: do either of you have ideas on how to approach this? do you have any thoughts about how other drivers should handle this issue? | 18:46 |
| clarkb | the openstack driver could use a nodepool level taskmanager if we wanted to be like the others (as sdk/shade operate not taskmanaged by default iirc) | 18:47 |
| corvus | clarkb: yeah; i thought there was some benefit to handing that off to sdk though | 18:48 |
| fungi | i guess the discrepancy is that we abstracted the task management out into shade and eventually openstacksdk when all that logic was forklifted from nodepool? but that other providers don't have similar features in the sdks we use (or does nodepool just talk directly to rest apis for them?) | 18:49 |
| corvus | fungi: for both aws and google, there are sdks that are helpful (but probably not required) | 18:49 |
| tristanC | corvus: I'm not sure to understand the difference between threads and task manager | 18:49 |
| corvus | in particular, there's a bunch of auth+discovery stuff it does | 18:50 |
| corvus | tristanC: a taskmanager is a single thread which performs tasks on behalf of other threads. so 100 launch threads can submit a task saying "create an instance" and the single taskmanager thread executes those serially. | 18:50 |
| fungi | with caching | 18:50 |
| fungi | er, in the list of things shade's providing on the auth/discovery front i mean | 18:51 |
| corvus | tristanC: it solves both the problem of thread-safety in client libraries (if there is one) and the problem of rate-limiting to a single api endpoint across threads. | 18:51 |
| mordred | corvus: so - there's a tricky-ish part here | 18:51 |
| AJaeger | https://review.opendev.org/#/c/691114/ adds some go jobs and roles to zuul-jobs, anybody interested to review it besides mordred , please? | 18:51 |
| AJaeger | and https://review.opendev.org/695594 " replace /home/zuul by ansible_user_dir" needs a review as well, please | 18:52 |
| mordred | corvus: one of the reasons we kept pushing the taskmanager down into the sdk/ksa layers is to catch the rate-limiting requests for the actions that are ancillary | 18:52 |
| tristanC | corvus: i see, makes sense. thus could this be abstracted away? e.g. launch thread submit operations as lambda to a generic task manager? | 18:52 |
| corvus | tristanC: yes, that's more or less how it was implemented. | 18:53 |
| mordred | so, from nodepool, our action is "create_server" - but that involves one or more calls to keystone, one or more to nova and possibly some to cinder and neutron and glance - all abstracted away from the caller | 18:53 |
| mordred | so a high level taskmanager helps with the coordination of actions across threads - but not being in there there's holes in the rate limiting | 18:54 |
| corvus | mordred: yes, though i can think of some technical solutions to that problem | 18:54 |
| corvus | (which, i mean, would likely be necessary for client libraries which we don't plan to expand with this support) | 18:55 |
| mordred | yah | 18:55 |
| tristanC | mordred: for kubernetes, or more generally for rest, should an action be (verb, endpoint, data, callback)? | 18:55 |
| clarkb | fwiw I think many of the nodepool drivers are already distinct enough that havin different taskmanagement is probably not the end of the world | 18:55 |
| corvus | clarkb: i don't think they need to be that distinct | 18:55 |
| clarkb | that is fair | 18:56 |
| corvus | the framework is pretty close to "implement create; implement list; implement delete; you're done" | 18:56 |
| corvus | i'm basically there, after 2 hours of work, for the google api client. but then i run into thread safety issues. | 18:56 |
| tristanC | mordred: oh nevermind, create/list/delete sounds more formal | 18:56 |
| corvus | so i'm thinking, i like the idea that one can implement a basic driver in 2 hours. we did a pretty good job there. but the thread-safety/rate-limit aspect of it is not genericised. so i'm thinking, maybe we should do that, and end up still with the idea that it takes about 2 hours to implement a driver. | 18:57 |
| mordred | corvus: if we moved back to a nodepool level taskmanager - it might then also make sense to have list/filter be in the nodepool layer too, yes? | 18:57 |
| corvus | mordred: instances or other things? | 18:58 |
| AJaeger | corvus: left a comment on https://review.opendev.org/#/c/696939/14 - want to fix with a followup? Then I change my -1 to a +2A | 18:58 |
| mordred | corvus: instances | 18:59 |
| mordred | corvus: like - that seems like a pattern that you'd want in the generic parts | 18:59 |
| corvus | AJaeger: replied | 18:59 |
| mordred | from the sdk side - the task management is pretty easy for nodepool to just stop using if it added its own taskmanager back | 19:00 |
| AJaeger | corvus: thanks, I'm convinced ;) | 19:00 |
| corvus | mordred, clarkb, fungi, tristanC: okay, i'll dig deeper into this. i'll see if i can quickly come up with a reasonable proposal for a driver-generic taskmanager. if not, i'll do something quick for the google driver and we can come back later and refactor. | 19:03 |
| corvus | (but in any case, it sounds like the answer is not "do not make a taskmanager for the google driver" :) | 19:03 |
| tristanC | corvus: thanks, that sounds like a good thing to have. i'd be happy to review it | 19:03 |
| mordred | corvus: ++ | 19:04 |
| tristanC | corvus: one things that is also difficult to manage is the quota and respecting max-servers... would be nice if the manager could do that too | 19:04 |
| corvus | tristanC: agreed | 19:05 |
| clarkb | I thought max-servers was already handled in a single thread | 19:07 |
| clarkb | the provider manager that spawns the launch threads does that accounting iirc | 19:07 |
| clarkb | (I might have the names wrong) | 19:07 |
| mordred | max-servers is - but requests-per-second is handled down in the ksa Adapter layer (in sdk) | 19:10 |
| *** rlandy|brb is now known as rlandy | 19:11 | |
| clarkb | mordred: ya just responding to tristanC's feature request. I think that feature is already complete (though maybe needs improving?) | 19:11 |
| mordred | yah | 19:11 |
| Shrews | i just returned from my last ever doc appt for my shoulder (yay \o/), but i agree with the things mordred said | 19:13 |
| fungi | last ever? are you hanging up the snowboard? ;) | 19:13 |
| Shrews | fungi: yup. sold my equipment a few weeks ago | 19:13 |
| fungi | wow | 19:14 |
| mordred | https://opendev.org/openstack/keystoneauth/src/branch/master/keystoneauth1/session.py#L1002-L1007 <-- that's the current state of rate limit support in the ksa layer - along with https://opendev.org/openstack/keystoneauth/src/branch/master/keystoneauth1/adapter.py#L171-L178 | 19:14 |
| clarkb | Shrews: wow | 19:14 |
| mordred | Shrews: \o/ | 19:14 |
| *** spsurya has quit IRC | 19:14 | |
| corvus | Shrews: yay! | 19:14 |
| Shrews | i figure 2 injuries requiring 3 surgeries is enough | 19:14 |
| fungi | you can come out here and take up kitesurfing | 19:14 |
| Shrews | imma chase turtles with mordred | 19:15 |
| fungi | or that, sure | 19:15 |
| mordred | Shrews: it's much less likely to injure your shoulder for sure | 19:15 |
| fungi | worst case you'll need your shoulder treated for sharkbites | 19:15 |
| fungi | just bring a harpoon and you'll be safe | 19:16 |
| clarkb | coral is harder than snow though | 19:17 |
| openstackgerrit | Merged zuul/zuul-jobs master: openshift speculative containers https://review.opendev.org/696939 | 19:17 |
| mordred | clarkb: yes. but you're not supposed to touch the coral | 19:17 |
| Shrews | i'm assuming "harpoon" is the beer | 19:17 |
| Shrews | not the pointy stick thing | 19:17 |
| mordred | also - sharks really aren't interested in biting scuba divers - we're big and scary with all of our super loud bubbles | 19:18 |
| clarkb | mordred: ya be nice to the coral. But wave and current conditions can be terrible sometimes | 19:18 |
| mordred | it's the silly surface swimmers who look like tasty tasty morsels backlit against teh sun | 19:18 |
| tristanC | clarkb: i meant the other max- limit, for openshift it is done by the handler, and when multiple threads are creating pods, then the provider may go over the limit | 19:18 |
| mordred | clarkb: this is true | 19:18 |
| clarkb | (probably avoid getting in the water in the first place) | 19:18 |
| mordred | yeah | 19:18 |
| mordred | like - just don't dive when it's like that | 19:18 |
| clarkb | tristanC: seems like a bug in that provider? | 19:18 |
| clarkb | tristanC: I think we can fix that the same way openstack handles it? | 19:18 |
| *** jamesmcarthur has joined #zuul | 19:20 | |
| tristanC | clarkb: probably yes, a bug that can be fixed. but if there is a generic manager, then perhaps this can be done automagically? | 19:20 |
| clarkb | tristanC: maybe. I guess my point is that the proposed task manager isn't necessary and avoiding adding extra functionality into the task manager might make the abstraction simpler | 19:21 |
| tristanC | well i don't know how the existing task manager operate, but if it could do such thing (handle over-quota issue) it would further reduce the complexcity of driver implementation | 19:25 |
| *** themroc has joined #zuul | 19:27 | |
| tristanC | e.g., https://opendev.org/zuul/nodepool/src/branch/master/nodepool/driver/openstack/handler.py#L247-L265 and https://opendev.org/zuul/nodepool/src/branch/master/nodepool/driver/openshift/handler.py#L59-L80 are handling quotas | 19:27 |
| clarkb | tristanC: to me quota handling in particular is going to be driver specific | 19:27 |
| clarkb | because there are many axis on which you can apply quotas | 19:27 |
| tristanC | openstack invalidate it's cache, and openshift just keep on retrying | 19:27 |
| tristanC | aws and kubernetes doesn't check for quota issue | 19:28 |
| clarkb | ya | 19:28 |
| tristanC | and they all have that while attemp <= retry loop. | 19:28 |
| *** mhu has quit IRC | 19:29 | |
| ianw | mordred / corvus : https://zuul.opendev.org/t/zuul/build/0aebe58151d245ad835506d716c6f28a <- sibling-ised, containerised nodepool+nodepool-builder+nodepool-launcher functional test that worked | 19:29 |
| clarkb | though I guess if we don't actually care about the axes in the taskmanager and simply a boolean flag returned by the driver that may be simple enough | 19:29 |
| ianw | need to get it reviewable, but the basics are there | 19:29 |
| mordred | ianw: zomg! that's so awesome | 19:30 |
| tristanC | clarkb: yes, e.g. the taskmanager could have a 'is_exception_a_quota_exceeded_error' callback, then perhaps it could delay all create request until the exception is gone? | 19:32 |
| *** jamesmcarthur has quit IRC | 19:34 | |
| *** jamesmcarthur has joined #zuul | 19:34 | |
| tristanC | or the otherway around, e.g. driver.create can raise quota issue | 19:35 |
| tristanC | as this "while attempts <= retries" loop is implemented by all the drivers (except static)... | 19:37 |
| Shrews | we have API calls for the driver to report quota issues since the driver is the only source of knowing its quota. I don't feel like we could (or should) make that generic | 19:38 |
| clarkb | fwiw ansible 2.8 seems to be running stable on opendev after unpinning now | 19:39 |
| Shrews | if there's a deficiency around the existing api's, we should fix that | 19:39 |
| clarkb | all of the job failures I've looked at were failures in the job unrelated to ansible version | 19:39 |
| *** mhu has joined #zuul | 19:41 | |
| Shrews | corvus: oh, also, 2 hours for the gce driver is awesome. i looked at possibly implementing that a few weeks back (just on a whim) but didn't like the trial period stuff (and the resulting sales calls that I'm sure would have occurred) | 19:42 |
| corvus | Shrews: i used my redhat account and redhat office address | 19:43 |
| corvus | it doesn't seem like a terrible api | 19:44 |
| *** themroc has quit IRC | 20:49 | |
| *** hashar has quit IRC | 21:19 | |
| *** hashar has joined #zuul | 21:22 | |
| *** mattw4 has joined #zuul | 21:36 | |
| *** pcaruana has quit IRC | 21:38 | |
| jamesmcarthur | Hi y'all. I've been working on a Zuul page for wikipedia. I basically copied what was on the wikimedia page | 21:40 |
| jamesmcarthur | But there is a ton of stuff that's WMF specific. Plus, we have an opportunity here to put our keywords into the page (like open source CI) | 21:41 |
| jamesmcarthur | If you have a chance, could you all take a look: https://en.wikipedia.org/wiki/Draft:Zuul_CI | 21:41 |
| jamesmcarthur | The page isn't scheduled to be reviewed for several months, so we have some time to get it right | 21:41 |
| jamesmcarthur | However, I'd love to get a jump on it. Please feel free to suggest edits directly or just email them to me directly (jimmy@openstack.org) | 21:42 |
| *** sgw has quit IRC | 21:50 | |
| jamesmcarthur | There is some stuff that I'm pretty sure is off/old - like the references to Jenkins ( It listens to Gerrit stream-events feed and trigger jobs function registered by Jenkins using the Jenkins Gearman plugin.) | 21:50 |
| jamesmcarthur | So not only should we update this stuff, but also the wikimedia page as well | 21:51 |
| jamesmcarthur | Any/all help appreciated | 21:51 |
| corvus | jamesmcarthur: thanks! yeah, most of the info on that page is outdated or is very specific to wmf, so i think it'll be a good chunk of work. it's a good outline though, to make sure we answer all those questions. | 21:52 |
| corvus | i'm not sure the wmf page itself needs updating though; they may still be running on a fork of zuul v2. | 21:52 |
| jamesmcarthur | corvus: ok - got it | 21:52 |
| jamesmcarthur | So it turns our Robert Cathey can actually publish it for us. | 21:53 |
| jamesmcarthur | So basically, as soon as we get it ready we can get it live. This will be a huge SEO boost for us. | 21:53 |
| *** jamesmcarthur has quit IRC | 21:56 | |
| *** jamesmcarthur has joined #zuul | 21:58 | |
| corvus | i will try to give it some significant time by the end of the week; hopefully someone else will have time before then :) | 21:58 |
| jamesmcarthur | corvus: Happy to post this to the ML as well, if you think that's appropriate? | 21:59 |
| jamesmcarthur | I know it's just before the holidays, so no worries. | 21:59 |
| corvus | also, should the title be something like "Zuul (something to disambiguate)" ? | 21:59 |
| corvus | "Zuul (Project Gating System)" "Zuul (CI/CD Software)" "Zuul (CI Software)"... | 22:00 |
| jamesmcarthur | I did Zuul CI since that's basically our URL | 22:00 |
| corvus | it's our url, but not our name | 22:00 |
| jamesmcarthur | Excellent point | 22:00 |
| corvus | [and that only because we can't afford our name as a url :( ] | 22:01 |
| jamesmcarthur | lol | 22:01 |
| corvus | course, with the .org thing, maybe nobody will be able to | 22:01 |
| jamesmcarthur | Ergh. That irritates me so much. | 22:02 |
| jamesmcarthur | Re: the name - why don't you put some thought into it. I think we have to make a new entry, which is just fine. | 22:02 |
| jamesmcarthur | B/c you can't change the name after you've suggested it. | 22:02 |
| jamesmcarthur | Personally, I would go with Zuul (Open source CI) if we're really looking to get maximum SEO juice | 22:03 |
| jamesmcarthur | Otherwise (CI/CD software) | 22:03 |
| jamesmcarthur | I don't think the term "project gating" is easily or widely understood. HOWEVER, it's an amazing opportunity to define the term | 22:04 |
| corvus | i don't know enough about wikipedia norms to know what the best disambiguator would be; i'm open to about anything, but i do think it would be good to indicate that the name is Zuul | 22:04 |
| jamesmcarthur | And in fact, we could also put together a Project Gating page | 22:04 |
| jamesmcarthur | to explain the concept | 22:04 |
| corvus | jamesmcarthur: there's this page, all of whose citations post-date the creation of zuul: https://en.wikipedia.org/wiki/Gated_commit | 22:05 |
| corvus | once the zuul page exists, we should update that page to link to it, and dig up earlier citations | 22:05 |
| *** sgw has joined #zuul | 22:05 | |
| jamesmcarthur | ok - it looks like these guys (https://en.wikipedia.org/wiki/TeamCity) basically created that page for the same reason | 22:06 |
| clarkb | typically it would be Zuul (Software) I think. But there is multiple zuul softwarews | 22:07 |
| clarkb | jamesmcarthur: I think we should remove operational and configuration details and add history/evolution? | 22:07 |
| jamesmcarthur | Interesting that it's from JetBrains. I used to use a lot of their SQL analysis tools for MS SQL. | 22:07 |
| clarkb | and a bit on gating | 22:08 |
| jamesmcarthur | clarkb: Agreed. History, evolution, versions/release info, maybe plugins? | 22:08 |
| clarkb | jamesmcarthur: ya supported plugins/drivers is another good one | 22:09 |
| jamesmcarthur | If you all want to create a github account, you can just edit the page. | 22:09 |
| jamesmcarthur | Once we're all in agreement, I can create a new one with the correct page name | 22:09 |
| clarkb | github? | 22:09 |
| jamesmcarthur | and voila | 22:09 |
| clarkb | or wikipedia | 22:09 |
| jamesmcarthur | uhh | 22:09 |
| jamesmcarthur | wikipedia | 22:09 |
| fungi | i'm also working on an article about zuul for opensource.com (i'll seek some editors from this crowd once i have the draft ready) which could serve as another citation in the wiki article | 22:10 |
| ianw | jamesmcarthur: i agree that "project gating" isn't really an understood term ... i wrote https://review.opendev.org/#/c/683085/1 (which has a lot of great comments from Jan i should get back to) when i thought about this very thing | 22:11 |
| fungi | but i think we should feel free to cite recorded conference talks and things like that too | 22:12 |
| jamesmcarthur | oh - this is great ianw :) | 22:12 |
| jamesmcarthur | fungi: absolutely - the more citations we can do the better | 22:12 |
| clarkb | ianw: jamesmcarthur there is actually a wikipedia article on project gating | 22:12 |
| clarkb | https://en.wikipedia.org/wiki/Gated_commit | 22:13 |
| corvus | i don't think it's recorded, but here's a slide presentation from 2011 about gated commits: https://docs.openstack.org/infra/publications/2011-uds_p-launchpad/#(3) | 22:14 |
| corvus | here's a recorded talk from 2012: https://www.youtube.com/watch?v=1uzTKKxnYHs slides https://docs.openstack.org/infra/publications/2012-lca-overview/#(9) | 22:15 |
| *** tosky has joined #zuul | 22:15 | |
| corvus | please don't watch that | 22:15 |
| corvus | you can tell it's old because the aspect ratio is 4:3 | 22:16 |
| clarkb | wow this is really difficult | 22:20 |
| clarkb | https://etherpad.openstack.org/p/zuul-wikipedia | 22:20 |
| clarkb | it is easier for me to start putting edit ideas in there than make a new account >_> | 22:20 |
| clarkb | corvus: we good to make those releases? | 22:21 |
| clarkb | its been about 1.5 hours since ansible 2.8 took effect in opendev and there hasn't been any issues that I have seen | 22:21 |
| corvus | clarkb: sounds good, i'll get started | 22:22 |
| corvus | commit a7fa150e216d3eec9bcba974ef27349a1fd4d742 (HEAD -> master, tag: 3.13.0, origin/master, origin/HEAD) | 22:23 |
| corvus | clarkb: ^ look right? | 22:23 |
| corvus | relnotes are at https://zuul-ci.org/docs/zuul/releasenotes.html | 22:23 |
| clarkb | yes, that commit looks right. Also the delete 2.5 change hasn't merged yet so we won't accidentally tag that one | 22:24 |
| corvus | pushed tag | 22:25 |
| *** rfolco has quit IRC | 22:25 | |
| *** sshnaidm is now known as sshnaidm|afk | 22:33 | |
| clarkb | oh heh I've just realized corvus linked to the gated commit url before I did | 22:37 |
| clarkb | also this is actually really difficult | 22:38 |
| clarkb | what I've got on that etherpad is more of a rich outline than article content | 22:38 |
| clarkb | feel free to stick it in the wikipedia draft or edit it | 22:39 |
| *** Petar_T has quit IRC | 22:39 | |
| *** rfolco has joined #zuul | 22:45 | |
| *** hashar has quit IRC | 22:52 | |
| *** jcapitao|afk has quit IRC | 22:56 | |
| *** jamesmcarthur has quit IRC | 22:56 | |
| *** jamesmcarthur has joined #zuul | 23:03 | |
| *** jamesmcarthur has quit IRC | 23:03 | |
| corvus | sent the 3.13.0 announcement | 23:12 |
| *** rlandy is now known as rlandy|bbl | 23:27 | |
| *** sgw has quit IRC | 23:34 | |
| *** tosky has quit IRC | 23:40 | |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: WIP: Add Google Cloud provider https://review.opendev.org/698342 | 23:57 |
| corvus | clarkb, mordred, Shrews, tristanC: ^ that's actually in worse shape than yesterday, and i hesitate to suggest anyone look at it. *but* if you want to take a look at drivers/gcloud/adapter.py, you can see what i think a driver impl will look like with this | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!