 And thanks for joining the OKD working group in the chat. I put the link to the attendee list. So if you could add your name in there, that would be wonderful while we're waiting for everyone to join us. All right. So I don't see Christian, though I was chatting with him just a few minutes ago. So I just wanted to get started on time today, if we could, and respect everybody's time. And we've got a few people here, which is perfect. All right. Let's just see if I can get to the working group meeting. All right. To kick things off, a little bit of interesting news. We are shifting the OpenShift Commons gathering in Amsterdam on which Christian was going to speak. And we're all going to meet face to face in Amsterdam into a digital conference format rather than a live event thing. And we're moving the date to March 27. So we're going to do the virtual event so it doesn't collide with anything on day 0 and doing all of the main stage talks that we were planning on doing. But it is now, and you can read more about it on the OpenShift blog. And if you were planning on coming and bought a ticket, you can get a refund. And the virtual one will be a free event. And we'll just make that happen instead. And I hope everybody stays healthy and happy. But please do note that KubeCon EU and Amsterdam is still a go. They're still planning on hosting it. And there's a link here for finding out more details about how they're managing their novel coronavirus updates and their strategy for hosting. So I think it's going to be an interesting meeting, though a lot of folks are starting to decide not to come based on their company's preferences. So that's been sort of top of mind, I know, for a lot of people. We still have a room for an OKD working group meeting there. I booked a room at the Apollo Hotel across the way from the conference center for the week to hold working group and SIG meetings. So once we're a little closer to the event, I'll send around a note and we'll see who's actually coming. And if we can get together at least for a small face to face at that event. So let me just check here and see if Christian's joined yet. I'm here. Hello, everybody. Hi again. Hi there. So yeah, so Christian, we're hoping that you on the 27th, at the same time you are on the main stage schedule anyways, EU time can deliver your talk. We'll train you up on the platform on how to use it and everything. And then there's live Q&A and interaction with the platform that. The 27th is a Friday, right? Yeah, I think that's right. Yeah, so then that would be great if other people can join us too, because it's going to be a much more, I think, global and interactive audience for this. But we're planning on doing it on EU time. So everybody can chat in the chat and Q&A and ask questions and answer questions. It's a pretty cool platform called Intrata that I have not used before, but I will be trained on by then. So with that in mind, we'll go back to the meeting planning here. So I figure the next thing to do is to get an update Christian from you on the timelines for things and an update, perhaps, on what's going on with OKD at the moment. Hello. Welcome, Hey. Yeah, sure. So the timeline hasn't really changed. So we're still kind of on track to release the beta by KubeCon. By the time KubeCon starts, hopefully. So in like three weeks, we hopefully have everything done by then. There's still like the blockers from last time haven't really changed, but we've made some progress on most of them, I'd say. And also, we've been preparing another rebase of the F course branches for Installer and MCO. That is currently still blocked by the release of a stable ignition 3.1 specification that we will need for that. Well, we want to use that. We don't want to use the experimental one just to make things easier and more easily migratable as well. The main difference is that, well, between 3.0 and 3.1 is that HTTP headers support was added or will be added. That's been added into the spec 2.x series already. And now we're just waiting to get that into 3.0 as well. Yeah, we've been, as I said, working on all of the blockers. There is still the gray cut issue that I know Joseph is waiting for to get resolved that is currently blocking the Azure DNS from working correctly. We do have one good, I can actually paste a link, that we have set up a Fedora namespace in Google. And apparently Google is giving us credits or is giving the Fedora community credits to host images. So we will have Fedora Core OS images uploaded to the GCP, the Google container platform, very soon. And then the installer should be working there. Yeah. All right, so Vadim, do you have anything else to add to the update? Just a follow up for the Christian's news about the Google support. I verified that GCP is installable. We just need the location once Fedora Core OS was uploaded to the old place. We would be able to test GCP support and officially announce that it's ready. Other platforms still have small various issues, most notably OpenStack and AllWeird needs a few changes in MCO to land. And those would be ready to be added. And the problem with the full secret change would be resolved by MCO rebase. So we are pretty close. So by pretty close, is that in the next week or two weeks or pre-Coupcon? For Coupcon. For Coupcon. So could we do what I'm just trying to think how to make some noise and publicity about the GCP stuff? Could we record a short OpenShift Commons briefing walking through how to deploy on GCP? OK, deep. We could probably do that. Yeah, I hope that. So there is, especially the Azure issue has been so we've said it's a blocker for beta. And I know there's people working on that. This sprint and also next sprint, which should be closing before Coupcon. But there isn't really a guarantee that that'll be fixed. So I would even say we will release a beta before Coupcon the way. And then the beta may not include all the platforms, but we will definitely have some of the platforms supported there. They'll work in the beta because otherwise. So a sprint is three weeks. So the next sprint starts next week. And yeah, so they'll have three more weeks to get that done. It's one sprint. It's scheduled for the DRACUT team to work on that. I know that. But yeah, I can't really tell whether they'll finish it in that time. I do hope so, very much. So as soon as you think you're ready to walk through that one deploying on GCP, let me know. And we'll find the time. And we can just record the video. And I can edit it into an actual commons briefing type style thing with some OKD stuff on it. Is there someone who we should have on that recording from the F-COS team who could speak to that side of things? Would you add? Unless Dusty, maybe? Yeah, sure. But I'm probably Dusty. But I don't really know because we probably don't want to convolute things with F-COS. Well, of course, it's using F-COS as a base. But we do configure F-COS in a way that is not standard F-COS. Well, I'm not sure because the installer will do everything automatically. There won't be any special. F-COS becomes an implementation detail of OKD, essentially like R-COS is an implementation detail of OCP. But sure, I think Dusty May is probably a good person to have in that meeting anyway. Just because one, giving a shout out to Google or giving us, allowing us to host, which is a nice thing to do, but just basically having a video that shows it going all the way through deploying on Google. So it's not IBM Cloud. It's not Red Hat's hosted on AWS. It's something different and new. If we could have that ready and a blog post about it, then I can put out for KubeCon. That would be awesome. So Vadim, what I need from you is as soon as it works all the way through, ping me. And we'll schedule that recording. And we'll work out what the messaging should be as well. So that's the release timeline. Someone will have to tell me pretty quick who they are. Who KubeWay at Gmail is. Is that someone on this call or is it not? Would this email address ask for access to the Fedora Magazine article draft? They were probably deeply disappointed that there wasn't more there there to review and edit, but it was a good reminder for me that I'm going to try and get this out. And I think having the Google story, the GCP deployment story would be a good thing to push us on. So Neil, look for me to start writing this now that I'll get some more content in here. But I don't know who the KubeWay person is. I did give them access to it. So anyways, that's the status of the Fedora Magazine article. It's a non-status update. We still haven't gotten anything there. I'm just looking through the open engineering tests. What else should we be covering here? Is there anything around documentation update? Vadim? Yeah, I submitted a pull request for them. When we were in contact with Michael from Docs Team, I will keep on pushing him because the changes look pretty trivial. We just need to hide a few sections which don't apply to OCD or Fedora CoS, and then start writing them. That would be like, I think we counted five sections out of something like 50, so the majority still stays. Can you toss the pull request that you made so that I can track it as well? Sure. Pass that in the BlueJeans chat, and I'll do that. So there's nothing else on the update side of things. You want to open it up for conversation here and how people are doing with their testing, or if people have paused on their testing for another release cycle. Sounds good to me. Well, I'm getting close to getting to a position where I can start doing maybe some open-stack-based testing that might be happening in the next couple of weeks. Not exactly sure yet, but that is the plan to do that soon. But I haven't gotten to it yet. It sounds like the same status as my Fedora Magazine article. So we'll get pretty close to that here, and I'll watch this. All right, anyone else? Vadim and I have been working together with the bare metal team the past few days to resolve one of the last, probably the last, outstanding issue for OpenShift and bare metal, which is they ship a script, which is a Python script in the MCO, and we can't run that on F-Cos because we don't ship Python in F-Cos. And they're going to actually translate that to go. And yeah, so that's... The bad news is that it's also a blocker for OpenStack and OvirDeploy, because these are using the scripts. But we're pushing them, and hopefully a proper rebase would be merged in F-Cos branch of MCO soon because we can do that by passing the OCP process. I also found out that CentOS CoreOS kind of works. Thanks to James Castle, who built a lot of RPMs. We assembled the image for CentOS CoreOS and managed to make Fedora CoreOS update to CentOS CoreOS as a part of OKD cluster. The installation didn't proceed because of... I think multis doesn't know which binaries to use for CentOS CoreOS, but that's a minor thing. And the bulk of the job to actually make instances boot and make instances upgrade from one operating system to the other has been fixed. So we'll take it from here. Looking through the other one, so that's a lot of the blocker there. We did have a discussion last time around Minikube, we had a talk with CodeReady containers people. They want to use OKD as a basis for them. However, because of a few changes in 4.4, they cannot use it, they cannot use a single master deployment just yet. Needs a few fixes. But once they would be able to use OCP as a basis, we will ask them to evaluate OKD, and perhaps official CRC would be using OKD as well. Awesome. So the GEP won't be a CRC based on OCP anymore? That's an option we haven't discussed that far, but they are interested to stay as close to communities as possible. I'm okay with that idea entirely. So it would be post-4.4, so when we get to the 4.5 release, they would be able to use OKD. Is that basically the assessment if they had resources to do it? Probably, maybe the fixes will end in 4.4. It's still in a lot of things during the year. Okay, I'm just going to say working towards using OKD around the 4.4 release. That's as vague as I can be. I'm not going to change this from mini OKD to flash CRC using OKD. Let's save the notes so next time I know. So are there any other updates out there in the chat? And folks, James is always, I think you don't have a microphone pausing here and reading the chatter here. James Castle is saying it should be easy to fake F-cost at CD Red Hat release to work with the Maltis issue with Sentos until the underlying issue is fixed. It's currently listed as a hack to get it working on F-cost already. You know, saying it's using at CD OS release, not Red Hat release. Vadim is suggesting we can just work Maltis repo. And someone is asking that one. What is Maltis? What is Maltis? So now that we get down to the real meat, how about someone explains what Maltis is for fill up? Maltis is a container networking interface implementation. Basically, it's the low level TNI we use. And on top of it, we run either OVN or OpenShift SDN. The bonus for that is that Maltis can run multiple implementations of various TNI. So you can have two or three TNI plugins running in your cluster and just they're taking care of different things. That's the basis how we can use hardware implementations and, for instance, OVN. It's just a matter of placing, of teaching it, which binaries for OVN or OpenShift SDN to place and it's really a terrible hack on their side. So we should either quickly work in or make them do the right thing and detect the operating system in a better way so that those for OS would be supported as well. So I noticed that, I'm gonna say your name wrong Friak. You're here from Proton. I'm wondering what your plans are for possibly deploying OKD over there at Proton and if you've been doing any testing. Yeah, that's the way we tested CentOS CoreOS. We started with the proper Fedora CoreOS image but the update, which is included in our release image was in fact CentOS CoreOS RPMs and OS3 commit and it worked until the point of Maltis but once we have all things in place, we will know which repo should be updated and we can contemplate a plan. Meanwhile, CentOS folks could work on making this whole system official so that we would know if it's feasible or things. There is some problem with architecture which won't let it happen but I don't think it's, I don't see any issues with making a CentOS CoreOS at all and make OKD run on top of it. We're all on agreement with that. So I've been Friak, I'm sort of chatting with Friak from Proton in chat here. Are you guys running Proton on CentOS? What I think I remember? Unmute yourself if you like. That's actually smarter. There you go. We're running both CentOS and REL between both Enterprise and OpenVbox which is pretty much still stuck in the 3.11 for the moment, YouTube time constraints. Okay, once we have a beta or a GA, will you be trying to test the, on a small cluster? I will be trying to do that, yes. I'm not sure if we're going that direction with the company but I will definitely. Okay, well, good to know and welcome. Nice to have you here. Yeah. We also have some, maybe talk to us on the side as well about migrating from 3.11 to four. There are some helper tools out there as well. But anyone else on the call want to talk about where they're at with their testing? I think most of it is, we're still awaiting a few blockers to get the open stack and over and bare metals almost there. Not, Vadim and Christian, do you have anything else that you want to ask of people? Our demo on KubeCon.eu has been accepted. So me and Christian would be on Redhead's booth on KubeCon.eu and we'll be showing how to make your own OCD based release. Basically what release payloads are, how do we test it, which things are easy to modify, which are more complex and so on and so forth. So if you have ideas what to show, we're open for suggestions I guess. We have a small plan, which things to show but if you have some more please shout it out. So I would like to record that as a briefing prior to KubeCon, just so that we have it captured. And so let's along with the Google one, let's also record that once we figure out what it is, what that is going to look like. So let's get both of those things recorded as soon as, and it doesn't have to be super professional or anything because it's on the community side, but capturing that will allow me to have some content to use during the week of KubeCon to promote OCD. So there's one other, Joseph is giving a quick update here in chat on vSphere. I don't know if you want to unmute yourself, I'll read what you've written that vSphere looks good with a few issues on corporate proxy but it should be addressed by the next rebates. And that's good news. And also someone is asking Joseph again, if we're building our own OCD I'll be doing it with all of the images, all the images for all of the operators. I'm reading these out loud. It's going to be the same installation profile as the default because right now we don't have a way to sort of disable operators and not install them. So it's either everything or nothing and we'll do everything. So in the chat, there's a little bit of a conversation that there's about 100 plus images to do the build. Yeah, we're not rebuilding all of them but we're definitely taking them all for the install, I would say, but anymore, did you have something else in mind? Oh, that's correct. That's just not efficient to rebuild everything from scratch so that you would have the very same result. It's also pointless from the release, from the license point of view because you can use UBI 7, which is a basis for all of our images, everywhere. It's not like in three X days where you had to use CentOS base instead of the REL 7 base. What we will be showing Christian is correct. We would be slowly replacing our release image one, image by one, showing the difference, showing how you can upgrade from your own custom system to your other custom system and what do those images mean, where to look for source and how to play with them? It's essentially also useful for development if you want to test your own change and for example, let's say any operator you have made a change in, you will build that image and create a new release payload with your custom changed image in it and then you can upgrade to that. That's essentially also how we test things in OCP and OKD, so yeah, creating your own release payload from customized images but not all of them have to be customized. You can replace as many as you want. Usually you'll just replace one at a time to really see the failure and have a place to look for but you can definitely also replace more than just one. Are there any other questions in the chat? This may be our shortest meeting yet, which is not a bad thing. So again, KubeCon is still a go, so we'll all be there and I will send around a note to see who is coming and find a time that works for all of us to have a face-to-face meeting there and then we can get together there and as well there'll be people in the booth at certain times and I'll let everybody know what those times are, you guys get assigned and if anybody has other content that they want shared out around OKD, maybe that's a good push to get the Fedora article up and out so that we have all of that around KubeCon but we also have OpenStack, not open set, Red Hat Summit coming up in April as well. So that's it and I don't know when the next flock is but we should have a presence at that. So I have a large group of people coming with me to Red Hat Summit this year so and we're all gonna be at the OpenShift at the OpenShift Common Gathering. So there's that and a flock is this August 6th through 9th in Detroit. So... Detroit, okay. So I don't know if the CFP has opened up yet for flock but if it hasn't, if it has, then it probably, no, it hasn't opened yet but when it does, it'd probably be a good idea to get some proposals in for OKD stuff. Yeah. We will definitely do that from the Red Hat side, from my side and buddy, probably we can do that together again as well, I guess. And yeah, but yeah, we'll definitely do that from our side and everybody else is definitely also welcome to submit things and by then hopefully there is a stable version out. Yeah, cool. Oh yeah, there better be one by then. Let me just tell you. Yeah, they're grilling. It's gotta be existing by that point. Otherwise this will be just a miserable set of doubt expectations. Yeah, we'll get there. It's just, so the Ignition 3.1, we should be tracking that as well. I'm just gonna add that link in the notes. Maltis, all good. So I think, folks, that is all we have for today unless someone has something else that they wanna do. So the next, say that again. DevConf US is this chamber in Fringham, Massachusetts. Oh, that's right, that's right. And the CFP opened up. And the CFP opened up so somebody should start proposing. Anybody on this call lived near Fringham. And that's an, is that? I don't know where that is. It's just outside of Boston Mass. What month was it again, Neil? September, late September. I'm gonna say Boston area, Mass area. Yeah, those are all things that we should do. So I will endeavor to put a little bit more detail around all of these and hit the mailing list and remind people to submit. It's even better if it's from external folks at DevConf and Flock. I'm happy to help push you right abstracts with you and get those in if you're feeling the love for going to any of those places and have the travel budget and that will be great. All right, I am going to turn the record button off unless people have other things to say going once, going to.