[07:11:44] 10Continuous-Integration-Config, 10PHP 8.1 support, 10Patch-For-Review: Make PHP 8.1 voting on MW master - https://phabricator.wikimedia.org/T316078 (10tstarling) Is this fixed? It looks fixed. [07:18:43] (03PS2) 10Hashar: zuul: archive labs/striker/{deploy,staticfiles,wheels} [integration/config] - 10https://gerrit.wikimedia.org/r/848514 (https://phabricator.wikimedia.org/T317378) (owner: 10BryanDavis) [07:22:22] (03CR) 10Hashar: [C: 03+2] "Thanks for the patch! I have amended it to mark the repositories with the archived template and moving their definitions at the bottom of " [integration/config] - 10https://gerrit.wikimedia.org/r/848514 (https://phabricator.wikimedia.org/T317378) (owner: 10BryanDavis) [07:24:13] (03Merged) 10jenkins-bot: zuul: archive labs/striker/{deploy,staticfiles,wheels} [integration/config] - 10https://gerrit.wikimedia.org/r/848514 (https://phabricator.wikimedia.org/T317378) (owner: 10BryanDavis) [07:24:54] !log Reloaded Zuul for https://gerrit.wikimedia.org/r/c/integration/config/+/848514 # T317378 [07:24:57] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [07:24:57] T317378: Deactivate unused Striker scap3 deployment repos - https://phabricator.wikimedia.org/T317378 [07:40:06] 10Continuous-Integration-Config, 10Release-Engineering-Team: Speed up integration-config-shellcheck-docker job - https://phabricator.wikimedia.org/T321536 (10hashar) [08:50:54] (03CR) 10Hashar: [C: 03+2] "I guess we could later add an entry to allow anyone from miraheze.org? There are some wildcard entries at the top of the list for eg wikim" [integration/config] - 10https://gerrit.wikimedia.org/r/848549 (owner: 10Zoranzoki21) [08:52:44] (03Merged) 10jenkins-bot: Zuul: Add Owen from Miraheze to the CI allowlist [integration/config] - 10https://gerrit.wikimedia.org/r/848549 (owner: 10Zoranzoki21) [08:54:27] (03CR) 10Hashar: [C: 03+2] "That looks correct (short of our system requiring copy pasting of definitions :D ). I am +2ing, CI will check whether all triggered jobs " [integration/config] - 10https://gerrit.wikimedia.org/r/848553 (owner: 10Tim Starling) [08:56:43] (03Merged) 10jenkins-bot: Only run PHP 8.1+ tests on the EventSimulator extension [integration/config] - 10https://gerrit.wikimedia.org/r/848553 (owner: 10Tim Starling) [08:56:59] (03CR) 10Hashar: [C: 03+2] "Deployed" [integration/config] - 10https://gerrit.wikimedia.org/r/848549 (owner: 10Zoranzoki21) [08:59:33] til php 8 has Fibers ( https://www.php.net/manual/en/language.fibers.php ) [09:06:06] yeah, the RFC has some good details on the motivation https://wiki.php.net/rfc/fibers [09:19:07] NodeJS has https://github.com/laverdet/node-fibers#readme (which is optionally(?) used by webdriver.io) but that got broken with NodeJS 16. The author states Fibers are no more needed after NodeJS learned about async/await, Promises and Generator [09:19:22] anyway, I will read the RFC ;) [09:31:26] I watched a talk about it a while ago, it seemed like some people had built up a pretty impressive-sounding ecosystem around fibers back when they were still an extension [09:31:54] but in my notes I only linked to the RFC so I assume the RFC links to whatever else they presented ;) [09:33:27] and their FAQ has `Why not add an event loop and async/await API to core?` which is the question I had in mind ;] [09:34:02] :P [09:34:06] that would apparently be a lot more work since builtin would have to be made async as well [09:34:38] whereas fibers seems to me a low level brick which is fairly isolated from the rest of php [09:34:45] but at least lets others build on top of it [09:34:49] yeah [09:35:03] one sure thing, it is good to see PHP evolving ;-] [09:35:39] and from yesterday, a Go php app server https://frankenphp.dev/ ;D [09:36:31] a php sapi with a php webserver having TLS cert generation built-in https://caddyserver.com/ [09:50:12] “☢️ FrankenPHP is very experimental, don't use it in production yet” good thing they make that so clear on the landing page (: [09:57:03] NO WARRANTY! [09:57:12] but that caddy server is interesting [12:32:33] 10Continuous-Integration-Infrastructure, 10MediaWiki-Core-Tests, 10Browser-Tests, 10Patch-For-Review, 10User-zeljkofilipin: ffmpeg stderr Capture area 1920x1080 at position 0.0 outside the screen size 1280x1024 - https://phabricator.wikimedia.org/T317879 (10zeljkofilipin) >>! In T317879#8240762, @hashar... [12:35:27] 10Continuous-Integration-Infrastructure, 10MediaWiki-Core-Tests, 10Browser-Tests, 10Patch-For-Review, 10User-zeljkofilipin: ffmpeg stderr Capture area 1920x1080 at position 0.0 outside the screen size 1280x1024 - https://phabricator.wikimedia.org/T317879 (10zeljkofilipin) Looks like [[ https://codesearch... [12:57:15] (03PS1) 10Jaime Nuche: perform-release: reset patch version when suggesting new version [tools/scap] - 10https://gerrit.wikimedia.org/r/849091 [13:44:41] Hi, we've already done a couple of ad-hoc deployments on restbase. Since there is some development happening lately, can we reuse the deployment window we have for content transform node services to also deploy restbase ? [13:47:04] 10Continuous-Integration-Infrastructure, 10MediaWiki-Core-Tests, 10Browser-Tests, 10User-zeljkofilipin: ffmpeg stderr Capture area 1920x1080 at position 0.0 outside the screen size 1280x1024 - https://phabricator.wikimedia.org/T317879 (10zeljkofilipin) 05In progress→03Resolved [14:04:07] 10Continuous-Integration-Config, 10PHP 8.1 support, 10Patch-For-Review: Make PHP 8.1 voting on MW master - https://phabricator.wikimedia.org/T316078 (10Jdforrester-WMF) >>! In T316078#8280879, @Tacsipacsi wrote: >>>! In T316078#8280802, @Jdforrester-WMF wrote: >> THat's how CI has been configured for years.... [14:08:05] 10Continuous-Integration-Config, 10PHP 8.1 support, 10Patch-For-Review: Make PHP 8.1 voting on MW master - https://phabricator.wikimedia.org/T316078 (10Jdforrester-WMF) >>! In T316078#8340645, @tstarling wrote: > Is this fixed? It looks fixed. No, not yet. CI has 8.1 voting for MediaWiki core itself and Me... [14:20:59] (03CR) 10Zoranzoki21: "Thanks for deploying this quickly! 😊" [integration/config] - 10https://gerrit.wikimedia.org/r/848549 (owner: 10Zoranzoki21) [14:26:53] maintenance-disconnect-full-disks build 433193 integration-agent-docker-1024 (/: 27%, /srv: 100%, /var/lib/docker: 30%): OFFLINE due to disk space [14:30:40] !log Manually cleaned /srv/jenkins/workspace on integration-agent-docker-1024 [14:30:41] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [14:35:38] maintenance-disconnect-full-disks build 433195 integration-agent-docker-1024 (/: 27%, /srv: 53%, /var/lib/docker: 30%): RECOVERY disk space OK [14:43:16] 10GitLab (Project Migration), 10Phabricator, 10Release-Engineering-Team (Bonus Level 🕹️), 10serviceops-collab, and 3 others: Migrate active repositories in Phabricator Differential to GitLab - https://phabricator.wikimedia.org/T191182 (10hashar) 05Resolved→03Open There are still active repositories in... [14:49:19] (03CR) 10Ahmon Dancy: [C: 03+2] "Thanks!" [tools/scap] - 10https://gerrit.wikimedia.org/r/849091 (owner: 10Jaime Nuche) [14:56:23] (03Merged) 10jenkins-bot: perform-release: reset patch version when suggesting new version [tools/scap] - 10https://gerrit.wikimedia.org/r/849091 (owner: 10Jaime Nuche) [15:08:00] 10Continuous-Integration-Config, 10PHP 8.1 support, 10Patch-For-Review: Make PHP 8.1 voting on MW master - https://phabricator.wikimedia.org/T316078 (10hashar) @Jdforrester-WMF Tim had the use case for a PHP 8.1 only extension (due to using Fibers). As a result the Zuul config now has a `extension-quibble-ph... [15:16:16] (03PS1) 10Hashar: Add RESTBase to Content Transform Team window [tools/release] - 10https://gerrit.wikimedia.org/r/849112 [15:18:32] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap: Set up equivalent of the scap-tox-doc-docker job and scap-tox-publish in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321223 (10dancy) [15:21:40] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap: Set up equivalent of the scap-tox-doc-docker job and scap-tox-publish in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321223 (10dancy) a:05hashar→03dancy [15:36:41] !log Tag Quibble 1.4.7 @ f838a24cc2 # T320935 T318029 [15:36:44] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [15:36:44] T318029: Quibble fails with UnicodeDecodeError when test output has invalid unicode - https://phabricator.wikimedia.org/T318029 [15:36:45] T320935: api-testing tests break for GrowthExperiments - https://phabricator.wikimedia.org/T320935 [15:42:45] (03PS1) 10Hashar: dockerfiles: update to Quibble 1.4.7 [integration/config] - 10https://gerrit.wikimedia.org/r/849116 (https://phabricator.wikimedia.org/T320935) [15:47:12] (03PS1) 10Hashar: jjb: update to Quibble 1.4.7 [integration/config] - 10https://gerrit.wikimedia.org/r/849118 [15:47:30] (03CR) 10Hashar: [C: 03+2] dockerfiles: update to Quibble 1.4.7 [integration/config] - 10https://gerrit.wikimedia.org/r/849116 (https://phabricator.wikimedia.org/T320935) (owner: 10Hashar) [15:49:21] (03Merged) 10jenkins-bot: dockerfiles: update to Quibble 1.4.7 [integration/config] - 10https://gerrit.wikimedia.org/r/849116 (https://phabricator.wikimedia.org/T320935) (owner: 10Hashar) [15:54:03] 10Release-Engineering-Team (Priority Backlog 📥), 10Release, 10Train Deployments: 1.40.0-wmf.8 deployment blockers - https://phabricator.wikimedia.org/T320513 (10thcipriani) p:05Triage→03Medium a:03jeena [16:10:30] hashar: When you have a moment I need help debugging a tox problem. [16:10:44] sure! [16:10:55] The symptom is visible at the bottom of https://gitlab.wikimedia.org/repos/releng/scap/-/jobs/27142 [16:11:24] AttributeError: module 'py' has no attribute 'io' [16:11:25] fun [16:11:51] oh that is to run `tox -e doc` from gitlab for scap isn't it ? :-\ [16:11:52] And this is what I'm trying to do: https://gitlab.wikimedia.org/repos/releng/scap/-/commit/fed894a26992c73093dfa9ba9b59a0c7ff90419f [16:11:55] yes [16:14:02] I tried adding python3-py to the package list but that didn't help [16:14:36] That package provides /usr/lib/python3/dist-packages/py/_io/terminalwriter.py [16:15:11] I would assume some of the dependencies had a recent release causing the break [16:17:03] https://pypi.org/project/py/#history has 1.11.0 from Nov 4th 2021 [16:20:07] fun it works for me locally with a fresh env [16:20:37] I can reproduce the problem on my machine (when using blubber to build the image) [16:21:15] If I run tox outside of a container it works [16:23:14] the gitlab build has a lot more dependencies than I would expect [16:23:43] including `pytest` and there is a report from a few hours ago with the same issue for pytest https://github.com/pytest-dev/pytest/issues/10420 [16:23:54] 7.2.0 got released 8 hours ago https://pypi.org/project/pytest/#history [16:24:00] so I am guessing that is the sameissue [16:24:30] ooh thanks Hashar! [16:24:34] for building the doc I think we need the urntime requirements (no pytest needed) [16:24:46] then invoke tox to install the latest libs defined in doc-requirements.txt [16:24:54] this way pytest (from test-requirements.txt) would not be installed [16:24:57] which work around the bug [16:25:18] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Projects-Cleanup, 10Release Pipeline (Blubber): Archive the Blubber gerrit repo - https://phabricator.wikimedia.org/T317820 (10thcipriani) >>! In T317820#8333031, @dduvall wrote: > I would think it would be on a repo by repo basis in the settings. Wh... [16:25:20] hmm. ok I'll do some experimentation [16:25:22] (03CR) 10Hashar: [C: 03+2] "Successfully published image docker-registry.discovery.wmnet/releng/quibble-buster-php80:1.4.7" [integration/config] - 10https://gerrit.wikimedia.org/r/849116 (https://phabricator.wikimedia.org/T320935) (owner: 10Hashar) [16:25:40] dancy: sorry I could not get to that kiriku build system again :-\ [16:28:20] hashar: Your suggestion did the trick! Thanks! [16:32:22] dancy: https://gitlab.wikimedia.org/repos/releng/scap/-/commit/fed894a26992c73093dfa9ba9b59a0c7ff90419f#note_12324 ;) [16:32:36] so yeah I think just using reuqirements.txt and doc-requirements.txt is good enough [16:33:43] !log Updating Jenkins jobs for Quibble 1.4.7 # T320935 T318029 [16:33:47] Logged the message at https://wikitech.wikimedia.org/wiki/Release_Engineering/SAL [16:33:47] T318029: Quibble fails with UnicodeDecodeError when test output has invalid unicode - https://phabricator.wikimedia.org/T318029 [16:33:47] T320935: api-testing tests break for GrowthExperiments - https://phabricator.wikimedia.org/T320935 [16:34:01] kostajh: I am switching CI to Quibble 1.4.7 [16:57:09] and one day I will have to add the remote tracking branch and remote name in my git prompt [16:57:41] 10GitLab, 10Wikimedia-GitHub: Mirror repositories hosted on our GitLab to GitHub - https://phabricator.wikimedia.org/T321597 (10taavi) [16:58:14] yeah the double -vv is the magic (the first gives you the status (ahead/behind) the second -v gives the tracking branch [16:58:18] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Projects-Cleanup, 10Release Pipeline (Blubber): Archive the Blubber gerrit repo - https://phabricator.wikimedia.org/T317820 (10bd808) I made {T321597} as a new home for my thread jack of this ticket. [16:58:29] Great info. Thanks for the help again! [16:58:58] 10GitLab (Integrations), 10Release-Engineering-Team, 10Wikimedia-GitHub: Mirror repositories hosted on our GitLab to GitHub - https://phabricator.wikimedia.org/T321597 (10brennen) [16:59:55] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap: Set up equivalent of the scap-tox-doc-docker job and scap-tox-publish in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321223 (10dancy) https://gitlab.wikimedia.org/repos/releng/scap/-/merge_requests/3 [17:03:29] Awaiting your approval. [17:03:44] afk in the meanttime [17:09:27] dancy: I think I have merged it [17:09:32] Changes merged into master with a2b3268a. [17:11:07] looks like gitlab forces a merge commit which refers to the merge request, the message has `See merge request repos/releng/scap!3` [17:11:55] ah yeah https://docs.gitlab.com/ee/user/project/merge_requests/methods/ [17:12:03] 10GitLab (Integrations), 10Release-Engineering-Team, 10Wikimedia-GitHub: Mirror repositories hosted on our GitLab to GitHub - https://phabricator.wikimedia.org/T321597 (10bd808) Having the ability to inherit project config from the parent project (https://gitlab.com/gitlab-org/gitlab/-/issues/28593) would ma... [17:16:08] hashar: you can configure a fast-forward merge as the behavior on a per-project level. I expect that some folks will have very strong opinions about enabling that for various repos to avoid complex ancestry. [17:18:12] Nod. Same as Gerrit [17:24:18] well I am off diner time! [17:25:44] (03CR) 10Hashar: [C: 03+1] "I have deployed Quibble 1.4.7 which should now set `QUIBBLE_APACHE=1` for the api-testing phase. We can deploy and check this tomorrow." [integration/config] - 10https://gerrit.wikimedia.org/r/843421 (https://phabricator.wikimedia.org/T320935) (owner: 10Kosta Harlan) [17:25:46] GitHub<->GitLab comparisons get pretty tiring, but I did find it confusing that GitLab only allows one merge type per project where GitHub lets you choose between merge, squash, and rebase per PR (by default) [17:45:44] bd808: I moderately prefer the per-repo setting TBF. But yes, the feature divergence is odd. [17:57:13] you can allow choosing at merge time... right? [18:00:26] brennen: I think you're thinking of "squash commits when merging" [18:01:08] or are you referring to "When there is a merge conflict, the user is given the option to rebase" under "Fast-forward merge"? [18:01:54] ah, yeah, i think i'm thinking of the options under "squash commits when merging" [18:57:48] 10Continuous-Integration-Config, 10WikiLambda, 10Abstract Wikipedia team (Phase κ – Clean-up), 10MW-1.39-notes (1.39.0-wmf.22; 2022-07-25): Run some full end-to-end browser integration tests on Beta Cluster Wikifunctions now it exists - https://phabricator.wikimedia.org/T294388 (10Jdforrester-WMF) [19:23:49] 10Project-Admins: Requests for addition to the #acl*Project-Admins group (in comments) - https://phabricator.wikimedia.org/T706 (10JArguello-WMF) Hello @Aklapper! @VirginiaPoundstone is a product manager working with the API value stream team, and she needs to create subprojects and milestones for every sprint.... [19:45:41] 10Continuous-Integration-Config, 10Quibble: Fully remove deprecated --commands - https://phabricator.wikimedia.org/T321617 (10hashar) [20:05:54] 10GitLab (Project Migration), 10Phabricator, 10Release-Engineering-Team (Bonus Level 🕹️), 10serviceops-collab, and 3 others: Migrate active repositories in Phabricator Differential to GitLab - https://phabricator.wikimedia.org/T191182 (10Dzahn) arr. ACK. "Query Overheated. This query took too long, so onl... [20:19:17] 10GitLab (Project Migration), 10Phabricator, 10Release-Engineering-Team (Bonus Level 🕹️), 10serviceops-collab, and 3 others: Migrate active repositories in Phabricator Differential to GitLab - https://phabricator.wikimedia.org/T191182 (10thcipriani) >>! In T191182#8343504, @Dzahn wrote: > arr. ACK. "Query... [20:19:36] 10Project-Admins: Requests for addition to the #acl*Project-Admins group (in comments) - https://phabricator.wikimedia.org/T706 (10Ladsgroup) >>! In T706#8343275, @JArguello-WMF wrote: > Hello @Aklapper! @VirginiaPoundstone is a product manager working with the API value stream team, and she needs to create subp... [20:35:27] /25/7 [20:46:10] (03PS1) 10Stef Dunlap: Use a more portable shebang in build scripts [tools/release] - 10https://gerrit.wikimedia.org/r/849172 [20:49:29] (03CR) 10Dzahn: "I wonder if this changes how/if CI runs commands like "shellcheck" or pylint. I think if scripts have file extensions and/or a known sheba" [tools/release] - 10https://gerrit.wikimedia.org/r/849172 (owner: 10Stef Dunlap) [21:10:47] (03PS2) 10Stef Dunlap: Use a more portable shebang in build scripts [tools/release] - 10https://gerrit.wikimedia.org/r/849172 [21:15:29] 10Phabricator, 10Striker, 10cloud-services-team (Kanban): Unable to mirror repository from git.legoktm.com into diffusion - https://phabricator.wikimedia.org/T143969 (10sbassett) [21:17:26] 10GitLab (Project Migration), 10Phabricator, 10Release-Engineering-Team (Bonus Level 🕹️), 10serviceops-collab, and 3 others: Migrate active repositories in Phabricator Differential to GitLab - https://phabricator.wikimedia.org/T191182 (10hashar) Awesome! When looking at phabricator translations (T321350)... [21:18:26] mutante: thcipriani: thanks for the investigation on Differential repos ;) [21:21:44] ACK, hashar. I disabled a bunch, all striker repos are gone but there is always something left.. plus "overheated query". thanks as well for taking phabricator/deployment [21:22:17] scap extension framework :p heh [21:22:28] some are probably easy, such as 3d2png which iirc got started around the time we wanted to move to Differential [21:22:34] 14 is not too many :) [21:22:44] and got written by gilles who came from DevianArt and was a former colleague of Mukunda [21:22:51] so I guess he started the project with Differential [21:23:04] well, I could import them all to gitlab but then I dont know where they are supposed to go, gerrit or gitlab [21:23:06] then looking at the config it seems Phabricator replicates it to Github [21:23:14] so well it is not so easy ;-] [21:23:32] my rule of thumb is probably Gerrit first cause that one is stable [21:23:33] I wish github was not also part of this discussion.. but that is too late [21:23:42] and move to gitlab after [21:23:47] well that depends really [21:24:08] for some third parties standalone repositories I guess we can reach to the users and get them straight to gitlab [21:24:23] the decision is probably going to be on a case by case basis [21:24:32] yea, that happened before for other repos [21:24:41] github mirroring has always been a trouble to me [21:24:43] I mean [21:24:53] the cost of maintaining the mirroring [21:24:56] yea, the "case by case" part is the problem I have :p [21:25:02] then it is done more or less on a best effort basis I guess [21:25:50] that old task T191182 was about to phase out Differential etnirely [21:25:51] T191182: Migrate active repositories in Phabricator Differential to GitLab - https://phabricator.wikimedia.org/T191182 [21:25:57] and I would be more than happy to complete that now [21:26:03] I don't really see "should it mirror to github" as related to "it should move away from phab" [21:26:19] since having four different git review/hosting systems is a bit of a pain [21:26:31] I think Striker / toolforge was the largest / most complicated one to achieve [21:26:51] and Phabricator does mirroring as well :-\ [21:26:58] yea, my part was to shut down git-ssh LVS service [21:27:04] I left "push over https" as an option [21:27:08] so have some nice stuff like gerrit > phabricator > github [21:27:10] for remaining users who might complain [21:27:16] or github > phabricator > gerrit [21:27:18] etc [21:27:20] beyond that I don't really know [21:27:34] especially the github part [21:27:49] https://phabricator.wikimedia.org/differential/query/all/ ;) [21:28:15] github -> phabricator -> gerrit = not really allowed, afaict [21:28:15] differential is barely active at least [21:28:19] ooh, both the people I want to see. [21:30:03] hashar: For T321223 we need a gitlab-deploy ssh keypair analogous to jenkins-deploy. [21:30:04] hashar: so if your default is to move to gerrit, and gerrit can sync to github then there should not be a problem to keep it as is for them..right [21:30:04] T321223: Set up equivalent of the scap-tox-doc-docker job and scap-tox-publish in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321223 [21:31:01] mutante: I think that is the easiest path yeah [21:31:08] hashar: ack, good [21:31:11] mutante: at the expense of later having to migrate the repo from gerrit to gitla [21:31:12] b [21:31:36] I don't care, as long as we can move it off of phab :) [21:31:39] but if we can strike out Differential, that would be already an achievement of some sort [21:31:49] agreed [21:32:33] dancy: I get the ssh key pair generated by SRE and stored in the puppet private repo then they get available from Puppet via `secret()` [21:32:37] hashar: can you guys kill releng-secrets [21:33:01] dancy: then I would assume Gitlab has a way to hold secrets? [21:34:14] Gitlab was supposed to copy Gerrit and Gerrit has always been decidedly "we dont want to trust it with secrets" [21:34:23] mutante: yeah definitely. Tyler told me yesterday about moving to 1password.com , I was supposed to write something about it for ^demon since he is interested. that is https://phabricator.wikimedia.org/T290337 [21:34:52] hashar: ACK, thanks [21:35:29] Gitlab has "protected variables". [21:35:43] dancy: I think T321223 can be marked solved and some other task filed to figure out how we can publish from GitLab CI toward doc.wikimedia.org [21:35:44] T321223: Set up equivalent of the scap-tox-doc-docker job and scap-tox-publish in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321223 [21:35:52] reasonable [21:35:54] or well reuse it but the scope is wider [21:35:59] I'll do that. [21:36:09] and kudos on implementing tox -e doc :) [21:36:15] haha thank you. [21:36:21] sorry I delayed that one to a point you had to do it yourself :-\\\ [21:36:43] no problem. [21:37:02] https://www.mediawiki.org/wiki/Continuous_integration/Documentation_generation has some probably obsolete documentation [21:37:18] https://wikitech.wikimedia.org/wiki/Doc.wikimedia.org has some other [21:37:32] I just traced through the job definition in integration/config. Ultimately rsync. [21:37:33] and I am pretty sure at some point I did an architecture diagram of how stuff get published [21:37:51] so hopefully soon you can congratulate me for implementing rsync /src /dest/ [21:38:00] ;D [21:38:38] there is a project floating on to phase out doc.wm.o entirely [21:38:41] well [21:38:42] hmm [21:38:49] Don't tell me about it! [21:38:52] baby steps. [21:38:57] dish out the current infra and instead have it hosted on k8s [21:39:20] but we need a place on which to store the generated documentation (currently that is the rsync dance and stuff being on the doc host) [21:39:37] but we should get a shared Ceph cluster at some point in the next few months [21:40:03] so theorically, we could get Gitlab to publish to that Ceph cluster under some namespace / dir it has access to [21:40:07] but none of that is any ready [21:43:07] I think the way doc is published is: [21:43:08] job runs on some CI agent to generate the doc then triggers `publish-to-doc` passing as parameter the ip of the agent [21:43:43] the publish-to-doc job runs on contint machine which has the ssh credentials to reach out to the doc server [21:44:26] it rsync from the agent to the contint host and then rsync again from contint to the doc host [21:44:35] so agent --> contint --> doc [21:45:21] I think it is done this way to have the ssh key pair credentials to only be exposed to the contint machine and to that specific job [21:46:47] anyway sleep time :-] [21:47:57] mutante: hashar said SRE generated the keypair (for jenkins-deploy). Does that mean you can get the private key to me? [21:48:33] 10GitLab (Project Migration), 10Phabricator, 10Release-Engineering-Team (Bonus Level 🕹️), 10serviceops-collab, and 3 others: Migrate active repositories in Phabricator Differential to GitLab - https://phabricator.wikimedia.org/T191182 (10Dzahn) p:05Low→03Medium [21:52:41] dancy: sounds like this is about adding a secret to a private repo, so probably yea. I am happy to do that via the ticket. I can't offer real time right this moment on IRC [21:52:56] OK! [21:53:03] I'll file something now. [21:53:34] feel free to tag it with serviceops-collab [22:11:42] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap, 10serviceops-collab: Set up equivalent of the scap-tox-publish job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321629 (10dancy) [22:22:44] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap, 10serviceops-collab: Set up equivalent of the scap-tox-publish job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321629 (10dancy) [22:26:11] (03CR) 10Subramanya Sastry: [C: 04-1] Use sup.reference instead of sup.mw-ref (courtesy T265930) [integration/visualdiff] - 10https://gerrit.wikimedia.org/r/848554 (owner: 10Subramanya Sastry) [22:30:43] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap, 10serviceops-collab: Set up equivalent of the scap-tox-publish job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321629 (10dancy) [22:34:43] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap, 10serviceops-collab: Set up equivalent of the scap-tox-publish job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321629 (10Dzahn) quick questions for now. Is it really rsync:// or is it rsync-over-ssh? We do have an rsyncd off... [22:35:28] 10Release-Engineering-Team (Deployment Training Requests): Deployment training request for Sergio Gimeno - https://phabricator.wikimedia.org/T320722 (10thcipriani) 👋 @Sgs — I added you to the invite for the Oct 27th training, will you be able to make it that day? [22:36:04] 10Release-Engineering-Team (Deployment Training Requests): Deployment training request for mfossati - https://phabricator.wikimedia.org/T313812 (10thcipriani) >>! In T313812#8335404, @mfossati wrote: > @thcipriani , thanks for the heads up! I've just rescheduled, ~~would you be so kind to send me the calendar in... [22:37:38] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap, 10serviceops-collab: Set up equivalent of the scap-tox-publish job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321629 (10Dzahn) edit: I saw your edits after making this comment. If it's that then I think the hosts would have t... [22:43:23] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap: Set up equivalent of the scap-tox-doc-docker job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321223 (10dancy) 05Open→03Resolved [22:43:25] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap: Get scap CI working on GitLab - https://phabricator.wikimedia.org/T319335 (10dancy) [22:44:00] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap: Set up equivalent of the scap-tox-doc-docker job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321223 (10dancy) [22:44:02] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap, 10serviceops-collab: Set up equivalent of the scap-tox-publish job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321629 (10dancy) [22:51:06] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap, 10serviceops-collab: Set up equivalent of the scap-tox-publish job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321629 (10dancy) [22:51:08] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap: Set up equivalent of the scap-tox-doc-docker job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321223 (10dancy) [22:51:10] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap: Get scap CI working on GitLab - https://phabricator.wikimedia.org/T319335 (10dancy) [22:53:11] 10Release-Engineering-Team (GitLab II: Wrath of Kahn 👾), 10Scap, 10serviceops-collab: Set up equivalent of the scap-tox-publish job in scap's .gitlab-ci.yml - https://phabricator.wikimedia.org/T321629 (10dancy) Thank you @Dzahn I could use some assistance managing the access list in a way that doesn't amou... [23:13:05] 10Release-Engineering-Team (Bonus Level 🕹️), 10Patch-For-Review, 10Release, 10Train Deployments: 1.39.0-wmf.27 deployment blockers - https://phabricator.wikimedia.org/T314188 (10Etonkovidova)