 Ok, bonjour tout le monde et bienvenue à la rencontre de l'infrastructure du week-end. Nous sommes le 11 octobre 2022, donc aujourd'hui, à la table, nous avons moi-même Damien du Portal, Bruno Verrarten, Hervé Le Meur, Marquise Hof, Stéphane Merle et Hello Bob, Kevin Martins et Bob from Salette.com. Ok. Hello Bob, you might want to get started. I wasn't sure. Yeah, thank you. Popping in unannounced, it's been some time since I was in one of these meetings and thought I'd just swing by to see where we left it last time. You guys were going to think about the management of secrets, whether there was a way that we could support Jenkins with our infrastructure. So for those who haven't heard, we're building a managed container service, kind of the equivalent to what you get in the major cloud providers. We're a new startup and kind of looking for jobs to support. So there's obviously a lot more context I can give, but I don't want to take too much time. So we first connected five or six weeks ago at one of these meetings. So I thought I'd just pop in to see what the latest was, if you guys had had a chance to consider Salette. So we didn't have heard back from the security team on the topic. And so we did not have time to spend more on that topic. So sorry for that. No problem. We have a full agenda. Right now we have priorities around one of our major sponsors, qui est un compositeuriste for the artifact that we produce, which is a lot on a lot of bandwidth. And that one is threatening the existence itself of the project. So we are absolutely focused on that one. I will say that infrastructure cost is a minor ions for us, because we're able to keep it at stake right now. Extending could be interesting to have to provide more builds and more build capability for our users. But right now, that's not on the top priority list. That's why we didn't come back to you earlier. No worries. No problem, not a problem. Hopefully, it sounds like you guys have got some fires to fight. So maybe I'll check back in, perhaps early next year, and see if we can be useful. I still have my question about the future request that we had last time about if there is a way for you to provide indicators of the energy that was consumed for a given task. Because we are currently studying the different cloud provider. And we are trying to correlate both the costs of a given Jenkins build, which is hard to get because it involves different resources, but also the electric impact of that build from end to end. So these are two, let's say important metrics that are going to gain more and more priority in the upcoming months and years. So those two elements, the cost per task and the consumption per task will be really interesting to get that topic that I think is going to grow and that will be important decision for us. Let's say, for instance, in AWS. We are not migrating yet. But if you want to, to us, the S3 storage, for instance, one of the best solution is to use a Stockholm data center, because it's one of the most efficient energy data center in the world. Close to French OVH data centers. So there are these maps and this kind of crowdsourcing information from cloud providers that helps to know, okay, what will be the, the most green way to run task or store data. Got it. I've taken that note down. And we'll put that one in the, in the feature list here at Salad. So, no, thanks, thanks for that insight. And yeah, thank you all for, for your time. And maybe I'll, I'll peel off and let you get, get back to business. But hopefully I'll touch base again in a few months. Cool. I think November should be a good time to get back to us. October is absolutely full because of October fest and the overhead it create for us with combines to the actual dumpster fire that we have to fight. So that will be worth it to check middle of November if it's okay for you. Perfect. No, it sounds great. Thanks everyone. Cool. Thank you. Have a nice day. Take care. Likewise. Cheers. Bye bye. So, first, thanks a lot, Bob. Okay. Let's continue to the announcement. If anyone can help me take notes of what we just said, that will help. No question about what Bob checked. I think at least a fan and RV were there last time. Salad.com is a new service. Which is a kind of crowd distributed computer. Everyone bring their own personal computer open. It's a kind of that the same idea as like city at home. Except this time, people run cloud resources and it run on the distributed Kubernetes across cloud machines. To be trusted on their system, you need to mine some whatever coins or whatever, I forgot the name of the system. So there is a blockchain system and the more you mine, the more you gain interest. And of course it uses GPUs. Initially it was a gamer network and it provides kind of distributed Kubernetes. So the challenge for us is as I said, the energy consumption. If we have to multiply by 50 or 100 times electricity consumed for a given plugin. I personally have an issue with that kind of matter. If it's only twice, we can discuss, but that's, yeah, that's kind of issue. The cost is absolutely decreased because yeah, the more people use their personal machine, the more we can crowdsource. But there is also safety and security issue. We run our workloads on someone else's computer basically. Okay, I know cloud is someone else's computer already. But in that case, there is, it's not a virtual machine, it's container that run on people machine. So the isolation is close to zero. So if we have technical opportunity that will be for CI Jenkins IO builds, assuming there is no credential at stake. The thing is that even cloning a git repository require a credential. And I don't want anyone on the world to have access to the GitHub credential use for that. The question has been asked by the security team. Haven't had a strong official answer yet. Okay, let's go proceed to announcements. So weekly release. I haven't checked yet. I saw some pull requests. Stéphane, did you saw the Docker image pop up in the pull request? I haven't checked anything. I'm sorry. No, no problem. Ah, what did I do? Okay, so let's check directly. Okay, I don't understand what just happened on my machine. I've lost my notes. Yeah, there. Do you have a keyboard? Can someone of you look at the weekly release if it's valid on Jenkins IO? Let me try. It's off until October. Happy redays. Just a reminder that Mark won't be available until the end of the month. I will have to prioritize something. It's the elections. So I'm, I'm the person in the community who will drive the election. We have proposed a topic. So Kevin, I might have to synchronize with you if it's okay. About, I might need help on that on that topic. But I just wanted to say I will have a decreased bandwidth on the upcoming this week. And we need to sync. No need for action right now. We'll sync after the meeting if it's okay. I see a raise down. Yeah, go ahead, Stephen. In the download section of Jenkins IO, I can see the new version for the weekly 2.373. 3.373. Okay, cool. The new one because on the intro, I see the one that we use right now is 2.372. I don't, I don't always responsible for the checklist, but I assume, I assume team will make it is the release officer. So either delegated to someone looks good for 2.7333. Is that correct, Stephen? 2.373. Yes. So I will focus on the elections. We have a proposal timeline for candidates. So again, if you will have to the idea is like, we'll do the same thing as last year. With some slight changes. So election will be still on the condo system because it's an outside of the table system, which itself service. We will have a first period of one week where we can send nomination. So everyone considered as concern the nomination. We validate for everyone sending a nomination if they are contributor contributor is broad. If you discussed on an issue on a community topic, you are considered contributor. You are not required to have written code documentation. Opening issues, raising bugs, helping others in the community are considering contribution. So you nominate people for the different roles and board if I'm not mistaken. That's the idea. So after a first period of people sending nomination. We aggregate and contact a different person and ask them to make a presentation and message for each. We publish this. Then people register for voting inside a community. We will ask this year for people to fill either their Github account or a Jenkins account so we can identify them more closely than only an account on discourse. And then there will be a voting period in November. The target is the end of all term, the term of all mandates is second of December. So that's when we plan to close the vote one or two days before. I think the 30th of November slightly and then result will be officially published the third of December. The first day of the new mandates. That's the overall idea. I'm now going into the details. I have to rush the first step for people being able to nominate candidates. So you will have the four of you to candidate to nominate candidates and to push anyone you know that have contributed even a little bit to Jenkins to nominate candidates and register for voting. That's really important to have a voting process. More detail as soon as get some that will be a topic on community Jenkins. I will create a case log where I will log every actions and decision on that part for people will do that next year. As information, since I'm someone managing that part as infrastructure officer, but I'm also planning to present myself and nominate myself. For that case, that means they could be a bias. So that's why we are, we have Olivier Verna, the former officer will run the election the past three years. That is a secondary person there as a fallback. So if anything goes wrong on anyone dot dot me, no problem. There is a second person and Mark wait will be as well. Since he's not planning to candidates to any he's not, he's not planning to nominate himself for any officer board. So I don't know for the board part, but I think that's only officers. So that's why the idea is to have ideally purse people that doesn't want to to proceed so we report to the board. If you have an issue, don't hesitate to ask for the board to tell me, maybe we can switch roles. I don't mind because I'm still a bit annoyed by conducting and managing the election smells like a word democracy system. Olivier did that years and no one complained. That's not a reason to keep continuing doing that. Someone did a message. Yes. Correct. Any question or other announcement. You need to explain what you just read because if someone watched the recording, you will not see what you just read. Sorry. That's a message about an internal club is like a channel that we can use to also speak about the election because we have some contributor there to gain some water because we have a lot of legit. So please, if at your own company, you got some contributors, please send them a message. Any other announcement or question on these topics. 123. No. Okay. Upcoming calendar next weekly. Next Tuesday, as usual. Next. I don't remember. Let me open Jenkins.io events. Okay. So here is the Jenkins calendar. We have a release candidates. The 19 of October. Next. I'm absolutely confused brain confused by the US dates time. My brain. Next. The S will be second of December. So we have some time. Next security release. Nothing official as far as I know. So no dates. Next major events. October fest. For the question for the calendar. Let's proceed. So what are the tasks that we did last week. So Bruno, welcome to the Jenkins organization. You're crazy. You don't know what you are capable of. Bruno is going to for people following the recording. Thanks. Alex on team for managing the get a permission for the free. Request. Stefaan was able to update the CRL for the VPN. So CRL certificate revocation list. Certificate chaine, that has to be renew. Every six months that we use if you want to. Si on veut de l'explicitement déterminer la certification, la certification de la VPN de quelqu'un et c'est le système utilisé pour générer la liste de répétitions qui doit être renouvelée chaque six mois en soi-même, sinon c'est déterminé et on ne peut pas choisir la VPN. Merci Stéphane. Donc il y a comme usual un nouveau événement dans le calendarié, dans six mois, qui va tomber trois semaines avant, donc on ne va pas tuer la VPN. Il y a été un upgrade de l'LTS, je ne sais pas si c'était la semaine dernière ou deux semaines auparavant et ça a été déployé pour tous les systèmes que nous avons déterminés, le contrôleur de l'LTS. Le Gravview plan est maintenant installé sur la release de l'IJN Kinsayo. Donc c'est la nouvelle visualisation dans le classique UI de l'IJN Kinsayo, plutôt que de l'IJN Kinsayo. Il y a aussi le Gravview et des nouvelles features. C'est vraiment cool, ça va aider à retirer l'IJN Kinsayo, quand le temps sera là. Service pour redirecter, recréer les réquestes, pour exporter des contenus de conflètes, ou pour rechercher, je n'ai aucune idée de ce que c'est. Je pense que sur le link, je l'ai installé dans le meeting notes. Oh, c'est Halloween, ok. Ce n'est pas seulement Halloween. CSS update. Bonne job, CSS. Je pense que le contexte... Spooky! Merci Service. Il y a deux modifications ici. La première, c'est la fixation pour les CSS. Elle a été servie comme un texte en plein, plutôt qu'un type de CSS. Et puis j'ai installé un bonheur de réplication, sur le dessus. C'est cool. C'est d'appeler ce schade. Bonne job. J'ai aussi vu des réquestes de drafts de Gavine, pour avoir le classique banner. C'est un webcomponent. Et aussi Adrien est regardé pour peut-être adapter ce webcomponent pour le score des plugins, comme c'est maintenant un interface en made. Qu'est-ce que c'est? plugins-lscore. Plugins with ls. Plugins... without... lscore. plugins-lscore. No. plugins-lscore. .jankins.io. Ok, je l'ai déjà ajouté. Ok, donc pour avoir le banner webcomponent sur le dessus, c'est cool. Bonne job. On va continuer dans le progrès de travail. La priorité importante, je suis suivi par la note. Je vais réorder la note. Pour avoir la priorité. La priorité la plus importante, c'est l'artificat de cacher Proxy. Réintroduisez-vous l'artificat de cacher Proxy? Est-ce que vous pouvez nous faire une petite statue sur cette? Où veux-tu que je le fais? J'ai maintenant un ECS cluster travaillant presque, je pense. J'ai la preuve de balance avec le contrôleur d'AWS, installé. J'ai l'impression que j'ai un ingress travaillant avec un ingress ou avec le balanceur d'AWS. Je ne suis pas sûr. Il n'y a pas de choix. Nous utilisons un GenX pour un ingress. Il n'y a pas d'alternatives. C'est un constrain que j'ai utilisé personnellement pour plusieurs raisons techniques et sauvages. Mais aussi pour la capacité d'utiliser. Amazon fait beaucoup de push-ups pour utiliser tout leur service stack. Nous n'avons pas. Et l'ALB n'est pas faible pour les futurs que nous avons besoin. L'axe de collection metriques n'est pas faible. Nous n'avons pas pu l'utiliser avec un agent data. Tout est fait pour vous. Je pense que nous ne devons pas utiliser le contrôleur LB. Maintenant, le travail est fait. Il est mieux de le faire. Mais il n'y a pas de raison pour ne pas utiliser un GenX. En progressant, merci beaucoup, Hervé. Il y avait une challenge dans l'environnement AWS. C'est un tout rapide. La prochaine étape, puis... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... qui est la proximité à opt-out, plutôt que opt-in, comme je l'ai planifié. Switch to opt-out avec la communication sur la liste de mail. Avant de faire ça, à peu près un jour, les développeurs, si ils ont un problème avec le caching, sont autonomes pour opt-out sur leur plug-in, et c'est temps pour nous fixer l'issue. C'est bien, à peu près là-bas. La prochaine issue de Twitter, GeneKins release, il y a beaucoup de week-end, donc... J'ai commencé à tester. Oui, depuis que les développeurs ont été consultés, nous n'avons pas rencontré un espionnage qui a été rencontré avec KOSUK. J'ai regardé l'automatique sur la publication de KOSUK. J'ai trouvé une grande représentation. J'ai testé localement pendant le jour de l'organisation. Je n'ai pas eu l'issue de la question de l'assistance quand il a évalué ce programme. J'ai créé une carte pour déployer cette carte dans un de nos clusters. Probablement pour les publics test. Cool. Une petite question, depuis que le public s'est étendu. Je l'ai ouvert, mais j'assume que cette application est un système de bot, mais il n'y a pas besoin de recevoir une question d'assistance. Il va émettre une question du cluster à l'extérieur. Oui. Je ne suis pas sûr qu'elle sera la place correcte. Il n'y a pas besoin. Non, mais le public public est la location de D-Facto parce que c'est le cluster de production. C'est la seule cluster de production que nous avons. Juste un souvenir pour vous, pouvez-vous ajouter une commande quand vous ajoutez l'installation dans la liste de file de l'aile. Une commande qui dit que ceci doit être émettue sur des tests privés dans le futur. Une fois qu'on a un cluster privé de production. C'est absolument minor, et si vous n'oubliez pas, ce sera quand nous devons migrer. Cool, merci Hervé. La prochaine question, ce n'est pas sur le link pour les réquestes. Nous n'avons pas fait rien et nous devons garder ça pour l'avenir milestone. Si quelqu'un l'a regardé et non? Marc voulait le faire, mais je ne sais pas si nous devons attendre 2 semaines. Oui. Si nous devons... J'ai élevé un CNI pour Marc. L'un d'entre nous va le prendre cette semaine. Est-ce que c'est bon pour vous? Oui. Vous pouvez le changer. La prochaine est déjà existée. Mais oui, le changer plus tard. Il serait facile de suivre maintenant. Et movez-les après. Ok. Movez, thanks Hervé. Changez le milestone. Nous devons les changer plus tard, parce que nous allons aller à la prochaine milestone. J'ai ajouté des soucis dans la prochaine milestone. Donc, si vous voulez les voir dans... Ok. Et c'est facile de faire avec la boxe de check. Ok. La prochaine est de perdre l'accès pour les plug-ins. Donc, Marc peut... Donc, comme nous avons discuté la dernière semaine, cette personne n'a pas suivi les choses en cryptique. Et bien sûr, ils ont utilisé les passwords plaintex avec Skr, qui n'est absolument pas le même système de authentication que Mayven. Donc, ça ne peut pas marcher tout à l'heure. Et l'end point ici est bizarre. Parce que vous avez besoin d'un code CSRF sur cette web-service. C'est pourquoi c'est un 503. Je peux prendre, si c'est ok pour vous, je peux le prendre là-bas, parce que j'ai la permission de l'administration sur le repo, si quelqu'un veut le prendre. Je ne sais pas. C'est là-bas. Donc, let's keep for next release. Realign repo Jenkins org mission. N'importe quoi. Comme je vous l'ai remarqué, j'étais malade dans ces jours, toute la semaine, donc je n'ai pas eu le temps de passer sur cette. Cette est très importante, mais oui, c'est un top priorité aussi, mais j'ai besoin d'une validation. Je pense que je vais demander l'aide de Basile et Daniel sur cette. Parce que ce sera un topic même spécifique. L'aide de Basile. Le next topic est CID Jenkins. Stefan, pouvez-vous donner un statut sur cette? Oui, j'ai l'impression que j'ai été en train de s'occuper de ce type de data depuis ce matin. Et maintenant, j'ai besoin de focus sur le dashboard. Mais la partie de la part de l'initiation de data est faite sur Linux et Windows. Ce sont des metrics de data. C'est un bon point. Le next step, le dashboard ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ce qui sera dans la variable complète, qui est la main de ce field de recherche. Pour maintenant, ce n'est pas clair. Ok. Ça pourrait être quelque chose. Stéphane, juste une note. Depuis que nous sommes des employeurs de CloudBiz, je sais que nos collègues de Philo utilisent ça aussi. Ils peuvent vous donner des insights, donc n'hésitez pas à les contacter. Il y a des français qui parlent aussi. Si c'est pour vous, c'est pour vous. N'hésitez pas à contacter nos collègues. Je vais vous donner le nom privilécié. Oui. Il y a quelques noms. C'est génial. Merci. Si quelqu'un avec l'expérience de Datadoc, c'est bienvenu. Stéphane. Tu as mon email. Pas de question sur le part de Datadoc. N'oubliez pas d'y arriver, jusqu'à la fin de l'année. Je veux qu'il y ait quelque chose que je peux vous montrer. Ok. La prochaine question. Fini du cleanup de MirrorBrain. C'est la même chose. J'étais malade. J'étais sur la table. 2 avances sur cette question. N'importe quoi. La prochaine question. Publiquée par Stéphane. J'ai fait un travail sur cette question. Réminder. Les ingénielles sont toutes les IOTI. The website. J'en ai tá. Là, on m'a donné le message. C'est la Première question. Pour y mettre en place. Je vais vous montrer son site. Ceci là, c'est si c'est la première question. Anant. Le point d'enfer, c'est que nous avons deux pipelines. on va pouvoir récupérer ces artefacts publics. Le but c'est de changer ces jobs à l'infrascié dans une zone privée. On va pouvoir publier ces artefacts sur le système public report. Parce que nous avons déjà des reports.genkins.io, qui est un bucket Azure avec des files publics accessibles. Ce sont des reports temporaires que tout le monde peut atteindre. Donc, nous pouvons migrer ces jobs à l'infrascié. La chose est que maintenant, nous sommes en face, c'est la première fois. Nous n'avons pas les mêmes features entre c.i.genkins.io, un c.i. en une main et l'infrascié à l'autre. Parce que ces contrôles ont profité différents goals. Donc, maintenant, les procédures sont pour assurer que tout ce qu'on utilise à l'intérieur de ces deux pipelines, la définition des tools, la définition du label et l'usage de Maven et Java, que nous habituellement avons à l'arrivée sur le c.i.g. Nous avons besoin d'un c.i.g. pour avoir les mêmes features. Donc, c'est la prochaine étape où je suis bloqué. C'est juste pour ajouter l'application correcte et ce sera bien. Mais pour ce qu'il faut pour ces jobs, je ne vais pas faire un c.i.g. mais je le ferai opportunement. Il n'y a pas quelqu'un pour qui j'ai plus de temps. Ils ont besoin d'un c.i.g. label. Nous avons une discrétion sur cet area que nous devrions résoudre par les contrôles de tous les contrôles dans le futur. Nous devrions avoir un seul moyen de générer les contrôles. Parce que maintenant, nous avons un c.i.g. et un c.i.g. sur l'alm charte. Nous devrions donc avoir tout sur le même système, l'un ou l'autre. C'est pourquoi je ne veux pas couvrir les autres sets de features si nous n'avons pas besoin. Réminder, c'est un bloc pour bouger chenkins.io, le site chinois de Trusted et le site d'organisation sur la machine virtual de Windows. Je pense que c'est un bloc. Non, c'est le top 1. Vous pouvez le faire et il y a un checkbox. Non, je n'ai pas. Je vais vous couvrir ce bloc. Je ne vais pas vous couvrir ce bloc. Je vais vous couvrir dans la section de la ferme et je vais vous couvrir. Et enfin, un account créé d'email recevait. Alors, quelqu'un créé un account qui nous a donné un email de username. Mais la personne n'a pas été l'answer, c'était 19 jours avant. Est-ce qu'il y a une question qui s'est annoncée pour moi, si la personne est ok, elle peut s'ouvrir. Je vais avoir un message et je vais ouvrir cette question. Ça vous semble bien ? Ok, je vais vous avancer ce bloc. Je ne vais pas avoir l'answer. Ok, c'est tout pour le travail progress. Je vais vous avancer. Je veux juste couvrir et discuter sur une autre location. Le premier. Remouvrez des ressources privées d'infra. C'est un nouveau issue, comme je l'ai remis. Élevé par Alex, pas mon faute. Je vais vous avancer les ressources locales. Je vais vous avancer cette question. Pour l'infra-ci, c'est la question parce qu'on l'a utilisé. J'ai eu le lien avec la preuve qu'on a vu, pour la génération reportée. On a besoin de cette plug-in. Je peux comprendre pourquoi on veut retirer une plug-in de relise. Je n'ai pas d'outre-relise, ce n'est pas utilisé. Mais je ne comprends pas pourquoi on veut retirer une plug-in, qui est considérée stable. Je vais aussi demander à Alex d'expliquer un peu plus pourquoi. Est-ce qu'il y a une raison ? Peut-être qu'il y a une issue de sécurité qu'on ne sait pas encore. On va le demander directement, mais je ne le vois pas. Maintenant, j'ai l'impression que je vais être incliné de refuser cette issue. Je ne vois pas le point qu'il y a beaucoup de travail. La surprise de retirer un contrôleur et d'attendre un travail pour fail parce que le logement est faible. Donc oui. Eventuellement, pour la relise, ce sera bien. Pour l'autre contrôleur, je ne sais pas. Je n'aurais pas essayé de toucher une certaine plug-in. La team de sécurité sait ce qu'ils font. J'ai éventuellement tristé, je suis sûr qu'il y en a. Est-ce qu'il y a quelqu'un qui n'a pas dit « non » à Alex ? Parce que si vous vous souvenez de l'âge, « non » est temporairement « oui » c'est définitive. Ok. Si c'est ok, je le prends. La prochaine est 31.66 sur IRC Bot. Comment on dit celui-là ? Oui. Donc, je vais juste attendre pour la page pour être discrédite. Oui. Currently IRC Bot n'est pas intégrée dans tout le CI. Ce n'est pas faux. Il s'est demandé d'attendre le CI pour toucher une certaine plug-in dans l'infra-folder dans le job folder. Il a also opened several requests on the report using command where I think they shouldn't be. So I want to ask him or propose him we re-integrate it in infra.ci. touchenkins.io first then in CI and the end slide could update this app and the dependency. Ok. What do you think about following the same pattern as what we did for Adrian on the plug-in health scoring. CI.genkins.io a docker image without deploying anything except the Jenkins for a valid image if they want to test it. As the sub-appline is already doing that it's using more parallel docker on an updately function. So it will work as you want. We still have an issue about docker. We have an error when sub-appline try to push to Jenkins for a valid registry. Ok. So that mean they might have issues for the deploy port as we can. The second request the second request is one I don't really I think it's not the technical it's to have a summary. Ok. Do you want to continue working on that or do you want someone, is there someone else volunteering should I take it? I don't mind taking it. The idea of me taking a lot of minor issues to be sure that you are focused on the artifact caching things. Ok. Like plug-in health, not from LVA. There is an issue on ci.g when deploying Jenkins for eval docker images. One note about key clock performance or if it's Stefan I will I think we will work on the two of us on pair on this one. So key clock is slow for numerous reasons but one for sure is because the database is on AWS while the application is on Azure. So the latency is terrible that might not solve all the issues of course but migrating last RDS manage database from AWS to Azure like we did for other services a few months ago the two of us that should help a lot. So I propose that we put that and we work on this this week. So it sounds good for you Stefan? Yes, that's perfect. Ok Also we add two issues so I'm looking at the last issues. There is an update center json returning 404 I'm adding that by default to track it so I'm going to ask help to Daniel for history reason because these dynamic update center were a thing of the past and I'm not sure they are maintained anymore but if they have to we need to find where is it documented so we need to dig up knowledge because we don't have the knowledge right now sounds like this is for the latest LTS so is there a process somewhere that should be fixed? I don't know but that one has to be covered on the upcoming Mark gave short term fixes for the end user so they will be un blocked because it looks like they are using that for whatever things they are doing but it provides them value so we need to be careful if these URLs are deprecated we need that, especially because we are still planning to migrate updates to oracle so if we break these elements some people will complain so if we have to deprecate these URLs better to know it now Yeah, don't think that the migration did that because it's already done we check with Daniel for history what other new issues did we have? I think that's okay so windows agents are slow there has been a long upcoming discussion and miscommunication and a lot of technical element as well so as Basile reminded us because we should have talked about that issue last week already and we forgot sorry for that so as reminded by Basile thanks for that we at least should check for enabling disk caching or ensuring it's enabled on the windows virtual machines and windows container that are used on CI Jenkins.io got some indication from James Nord that could be cloud level that could be packer image level there are multiple entry points here so if it's okay I propose that we move that issue also on the next on the next milestone Stéphane, if it's okay for you I will search for the information with James and once we know if there are changes to do either packer images or CI Jenkins.io then I will delegate these changes to you on the issue sounds good to you? Yes, thank you there are other discussion looks like related to the Jenkins Core maintainer so we eat one of the limitation of using GitHub for us and Jira for others so there are advantages there are pro and cons for each of the solutions I'm not personally a Jira user and Jira is soon to be doomed so I'm not sure how to under that one reminder when there are such discussion please take the due diligence to open an issue on Jira on the Jenkins project so that the core developer can have a triage on their own because they don't follow up on this element as far as I understand so if we need help or actions from the core developer such as in that case retry to be added on Jenkins file if it's something that you need help on please open an issue so they can try it and link it to our help desk issues that's the most minimum due diligence that we have to Are there any question on that topic? To check windows caching and then define to implement it if needed there were some fixes on the java doc generation so as part of the usual run books that I still have to publish on pull request for you folks I got a daily routine on trusted CI to check the state of the jobs 2 or 3 weeks ago I saw that the java doc was failing to build on the principal branch I tried to read the log and I wasn't able at first and second site to understand the issue so Bezel gave us some pointer it was middle of DevOps world so I wasn't able to fix that however Bezel already fixed that on core taglib report generator and so as a courtesy to the team he also fixed that a few days ago so the fix is because we weren't using LTS like Jenkins version with 3 digits but it was using a 2 digit version which means the weekly that was used as base for the LTS when we have the LTS selection we select one of the weekly as base and then we fork the LTS branch from that base and the script was trying to generate java doc for that weekly version while it should have on the LTS version if I'm not mistaken I took the time because that's pretty complicated to me so maybe I'm not brain enough reminder it looks like there is an expectation at least from Bezel and team that we maintain this repository so in that case never hesitate to ask for help and as I said I'm actually opening an issue on the Gera Jenkins project notification to someone else outside here there was also timing failure because when this happened I should have broke that on the team team meeting don't remember it was when did I open it so I failed and I'm sorry for that I failed to bring that topic on our meetings I should have done that that would have gained more traction and more coverage so no problem to open issues but we'll try to be a bit more diligent on that in the future so I'm adding this any question on this topic or being maintained by the infra team ok other topic Kubernetes 1.33 so as AirV created the issue earlier today and of October Digital Ocean will drop Kubernetes 1.22 so we have to upgrade to Kubernetes 1.23 proposal is that we start working on this one at least this week at least the cubcutter port looks good for everyone and I propose the following timeline upgrading at least Digital Ocean clusters next milestone and then the others are right after who is who want to lead the subjects lead doesn't mean doing I mean leading just go Stéphane go no I think you're better to lead than me to try to do you are raising your hand I was but that's before I explained exactly what that meant I don't care let me get my my benevolent dictator hat AirV, you're leading and Stéphane you will do that's exactly what I meant because AirV is good to point stuff so it's good which means right now AirV led by creating the issue so Stéphane you have to do the cubcutter upgrade this week by yourself looks good for you that's perfect can I also drop some stuff to Bruno for him not to be just watching us now no problem so that mean AirV does not have to think about that topic this week but you have, that's the goal I don't want him to be disturbed on the artifact caching that's exactly what I had in my mind he will hate us because he will spend all his time on that no because I know AirV will work on the EKS upgrade since he has his deep dive on these clusters ok that's already a lot so I propose that we stop right there and we keep walking I still have my little on the side issue about the acceptance test check availability that I still try to what's the issue reference I have no clue I think there is none so that's not that's not a priority no new test and check availability agent there is 20 days ago see this one ok so it's not I will have to fulfill better better stuff even the amount of task may I ask you to not spend any time this week unless I missed something there is already a lot of things can we defer that to later I will try because I really need you on other tasks I understand the dashboard this one is not priority it's useful but it's not priority and we have more prior issues we can spend time let's say next Monday or actually Tuesday but end of deterioration to brainstorm on that but I really need you to forget about this one for one week other topics folks nope ok so I'm stopping recording