 Hi, everybody. Welcome for this new Jenkins from meeting. Before we discuss about infrastructure on the peak, I just want to remind people that in today's, this is the deadline for the Jenkins officer and board position. So you still have two days to to eliminate people. So yeah. So then let's talk about the first thing. So the Docker Hub policy change. I sent an email to the CDF they accepted to sponsor the project. So this is a really great news. On the other side, I discovered at the same time that Docker Hub offer a free plan for open source projects. So I applied to see what's, if we could, I mean, if Docker Hub could sponsor the project, that would be really nice. We still have a strong deadline, which is the first of November. So if we cannot manage to get it done by that time, I guess we'll have to ask the CDF to pay for the bill. Anyway, the bill is not a big amount. We should be around 50, 50, 50, or up a month. So yeah, I still, I still hope that Docker Hub will accept to sponsor us. Any question or suggestion on this specific topic? No, one time, two times. So then Mark, did you want to share your screen? You're muted by my mark. I was trying to have my keyboard sounds not disturbed, but I'm happy to share my screen. Here we go. I like following on. Right. Notes. Here we go. And now I am going to mute myself because my clattering keyboard will disrupt it. So the next topic that I want to discuss is regarding the Windows Docker images. I haven't actively followed on that thing. So maybe Garrett, he can explain. I think you worked a little bit on those. You made good progress, but I'm just wondering what's missing right now. So we have the, the base VM images built by Packer. I think they're, they now should be up to date. There are a few issues with them getting the builds through. It's timing out in a few places. So I'm just watching those see if I can make sure that those go, get sorted. And then we should be able to recreate the VMs in Jenkins. I think after that, we, do we need some different images as well, rather than just the host that it runs on? Is that the idea? What do you mean by different images? Once, once you have, once you have the Windows machines running, we can just use them to build a Docker image, but I don't think that you have this, the version inside of the current image type. Okay. And I thought we used tagging of some form to identify the, the version number that previously the 1809 will be the, is the new version and 1809 will just, will continue to exist, but no longer be usable. That's the wrong way to say it. Continue to exist, but not be actively maintained. Right? But then what's the point to keep building images for that old version? If you're not, if the version is not maintained anyway. Right. We should stop building 1809 images that I agree wholeheartedly. I think we shouldn't bother building something that's on an operating system platform that's no longer receiving security updates after the last release in November. Because, yeah, because another point is when you go to let's say Docker Hub and you see an image, if you see that image was built, let's say, within the last 24 hours, you have the feeling that that image is still maintained, even if it's not, so we should definitely stop building for that specific version. So we may, so to answer the Garrett question, we may have to change that to the way we did those images. So we, I mean, it's not like we only need the virtual machine. We also have to change some stuff. So, Gareth, did I correctly capture this? The Packer definition is available, but the Packer image builds are not yet running successfully to completion. So we've got some work there before we can even Yeah, this, I think, yeah, most of them seem to be in the build queue. So they'll get processed at some point soon. That build queue's still about 60 in terms of where we're up to. But we have been having some timeouts or what look like, or connection closed midway through downloading some pieces. So locally just rerunning it seems to resolve those. But I'm just looking to try and see if there's a better way that we can download these dependencies. Are you using the EC2 instances for this, or is it also working on Azure? For the Packer image? So it's a Docker and Linux host that's does a Packer call to Azure or to AWS. So whatever is Azure is doing internally. But when you run it locally, it does the same thing. Yeah, I think. Okay, thanks. Thanks a lot. So no, forgive my ignorance. I thought that Packer generated an image that was specific to Azure or to EC2. Did I misunderstand that? No, that's correct. That's correct. But it actually does the building on Azure or EC2. Yeah, because the Docker image that run Packer is running from a container. But then you have specific script for Amazon or for Azure. Then you publish the reason. You can move it to infra.ci, that you can style it as well. And infra has access. Well, but I thought that oh, and infra shares the account because don't these images have to be published so that they're available to ci.jenkins.io to use them? And so I had assumed that. So in the case of Packer, it's not a major issue to use ci.jenkins.io because the account that pushed the image can only push images in a specific location. So I think it's fine to run it from ci.jenkins.io. So we could move that to infra.ci. But in the end, if the account used to push images is compromised, the only thing that we would be able is to change the image. I mean, it's not a risk for the whole account. But yeah, I think it would make sense to move those. But yeah, it's not needed for the work here. It could be a nice improvement for the future. Any other question? Nope. Then I propose to quickly move to the next one, which is the repo, the ci.jenkins.io.org. So the work being done by GFrog, I don't think we have any news here, Mark. Have you received an email from GFrog? I have not, and I've not seen any mention from Daniel Beck either of anything from them. Okay, so let's move to the next topic. Regarding the JIRA upgrade, we had a meeting last week with Mark with the Linux Foundation. We identified a few things that we have to work. So the different thing that we did was to share a backup of our instances of issues at jenkins.io. So we gave access to one person from the Linux Foundation to the JIRA machine so that person could retrieve the backup. Another topic that we discussed last week was, so the Linux Foundation used Let's Encrypt to Generate Certificates and they only used DNS methods. So we needed a way to share access to our zone. I didn't know at that time, but I discovered that on Azure, you can delegate permission only on one specific record. So I just created the record that are supposed to be created by Let's Encrypt and then create a specific user who can modify those records and share that information with Linux Foundation folks. So now, as far as I know, they are currently testing to restore and they are basically testing that they can restore the machine on their sites and they should be able to also generate JIRA certificates on their sites. So as far as I know, that's a current state. Maybe Mark has more context than me. Other than next week is our test week. They agreed in the meeting last week that they're okay with us using that as our scheduled test week. Next week to test a sample of the upgraded JIRA, we will discard the test, throw it away completely after we're done with it. It's just an experiment. I've had no subscribers yet, no volunteers to join the test. So let me ask directly in this meeting, Gareth, would you like to volunteer to be a tester? Tim, would you like to volunteer to be a tester? Olivier, would you like to volunteer? I have three volunteers now. Thank you. That's perfect. Just regarding JIRA, another thing that I discussed with them was to instead of using LDAP, we could use KickLoc as the source of identity. They are more concerned about the fact that they will have to restore the instance and so they fear that it may be more complicated for them to restore the backup on their site. So they're also testing to see if it's easy for them to use an auto token instead of an LDAP user to synchronize the user. Yeah, I thought the sims I had from their conversations was that they were concerned that the switch to OATOC and that they've done in the past caused them to do additional migration work. Did I understand them correctly, Olivier? Yep. Okay, so that's a risk and they'll want to control that risk. Yeah, another thing that I did was to share the license with them. So basically Atlassian is sponsoring license for a new open source project, so you can request any of them. So if in the future we are interested by a specific plugin or whatever we want to use, you can just request for it. But at least for now I just had to renew the different licenses, upgrade the plugin and then now we should be ready to migrate. So as a reminder, the first thing they will do is they will first restore our instance in their infrastructure and then they will do the upgrade from Gira7 to Gira8. So any questions or any misinformation here? Sounds like we are good. So I propose just to quickly discuss about the mirroring infrastructure happening at the moment. So just before the meeting we were just discussing about how we can have more mirrors and also how we deal with the fight that sometimes people from a country are not allowed to use mirrors from a different country. More specifically in this case it was people from India where they are not allowed to use mirrors from China. As far as I know, we do not support such things and I don't think that we should care about the different countries limitation except trying to have more mirrors so people rely only on mirrors in their own country. So we still have to promote the work that we did here and I still have to create a Gira8 pick with all the different tasks that we need to do more specifically on the mirroring infrastructure but I think that for now the migration is quite successful so I'm really happy with that. Anything to add on this topic? We have an action item previously to do a blog post announcing the new mirrors. I haven't done that yet. I think it was me or was that you Olivier that was going to do it? I don't think we decided who would work on that. I can try to find some time to work on this but I'm not sure right now but yeah this is something that I would like to do anyway. I really would like to have a blog post for that. I'm just more concerned about the time. Great and if one of us gets to it and submits it to the poll request, others can review it and correct an update as needed. But something that is really important to have in the blog post is a way for them to explain for the people how to report issues or things that they would like to have in the mirror infrastructure. The example is to have mirrors for a specific country for example. If someone is interested I think we should also have a way to explain how to hide a new mirror because it's not a big deal but yeah. Maybe if people realize how easy it is to maintain your own mirror. I'm not talking about the network bandwidth and stuff like that but yeah just configuring the machine is just really easy to do. Yeah I think Gunther he said he's running everything in containers so maybe we could ask him to share some of what he's doing. Well is that one where we want well should we do a Jenkins online meetup to highlight mirrors? That would be a good idea. That would be a good idea but I really think that the blog post should be part of the big picture about how we promote using mirror. Okay yeah it can be useful for March and especially to get some of those ones that map as mentioned would be good. Do you have access to a git repository? Tim? Which one? The one with the current manage to maintain a mirror. Have I seen one? No you don't. We would just have to ask Gunther whoever it was in Europe there that was maintaining as a... No yeah he's just sitting running all in the time as they didn't elaborate. Thank you but yeah anyway. But yeah if someone else is interested to work on this with Mark and me yeah feel free to manifest. I'll try to find some time to work on this. Those are the most important information right now so we still have nine minutes left so either we stop the meeting here or there isn't something else that you want to discuss. I propose to count three times one two three then I propose to stop the meeting here and continue on our scene. Thanks everybody. Thanks. Bye. Bye.