15:01:28 <davdunc[m> #startmeeting fedora_cloud_meeting 15:01:29 <zodbot> Meeting started Thu Apr 13 15:01:28 2023 UTC. 15:01:29 <zodbot> This meeting is logged and archived in a public location. 15:01:29 <zodbot> The chair is davdunc[m. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 15:01:29 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:01:29 <zodbot> The meeting name has been set to 'fedora_cloud_meeting' 15:01:37 <davdunc[m> yaya! 15:01:46 <davdunc[m> #topic roll call 15:01:47 <davdunc[m> ! 15:01:56 <mhayden> .hello2 15:01:57 <zodbot> mhayden: mhayden 'Major Hayden' <mhayden@redhat.com> 15:02:08 <davdunc[m> did I do that right? 15:02:09 <miyunari[m]> π Hi all π 15:02:28 <davdunc[m> #chair miyunari 15:02:28 <zodbot> Current chairs: davdunc[m miyunari 15:02:33 <davdunc[m> #chair mhayden 15:02:33 <zodbot> Current chairs: davdunc[m mhayden miyunari 15:02:34 <mhayden> Felix Felix[m]123 and Nina miyunari[m] from my team are here today π 15:02:38 <dduffey> .here 15:02:58 <davdunc[m> #chair dduffey Felix 15:02:58 <zodbot> Current chairs: Felix davdunc[m dduffey mhayden miyunari 15:03:21 <davdunc[m> welcome Nina and Felix! 15:03:33 <davdunc[m> dduffey: can you ping themayor ? 15:03:42 <dduffey> yes 15:03:48 <davdunc[m> ty 15:04:05 <mhayden> davdunc[m: i think felix will miss this one, actually π₯² 15:04:27 <davdunc[m> well, Felix is still on the chair. 15:04:33 <davdunc[m> :) 15:04:46 <dduffey> davdunc[m, themayor is out today 15:04:53 <davdunc[m> dduffey: ack. 15:05:20 <davdunc[m> #topic action items from last meeting 15:05:45 <davdunc[m> that isn't changing the topic is it? 15:05:51 <spotz_> o/ 15:06:04 <mhayden> davdunc[m: i'm on IRC and i see the topic changed 15:06:15 <nzwulfin> same 15:06:32 <davdunc[m> oh good! 15:07:01 <davdunc[m> I'll open up irc so I can see it too. 15:07:10 * mhayden has not entered the matrix yet π 15:07:24 <davdunc[m> so I am going to open with the azure agent. 15:07:29 <spotz_> I'm there but IRCing today:) 15:07:42 <davdunc[m> #chair spotz_ 15:07:43 <zodbot> Current chairs: Felix davdunc[m dduffey mhayden miyunari spotz_ 15:07:59 <davdunc[m> spotz_: you are the irc champion. 15:08:00 * nzwulfin barely remebers hexchat 15:08:01 <davdunc[m> :) 15:08:42 <davdunc[m> we have a patch to support the work in the agent RPM for the council requirement to be environment aware 15:08:53 <davdunc[m> we just need to add the image to the build. 15:09:09 <mhayden> right, i was looking for that pagure issue for context 15:09:36 <davdunc[m> I have it on my action items to add it to fedimg rebuild, but, and this is going to be familiar, I just keep running out of time to work on it. 15:10:12 <davdunc[m> so I am going to just throw it out there that if you have time, I need assistance. 15:10:19 <mhayden> is it mainly just a pull request to the image build manifest and then an image test after that? 15:10:25 <dduffey> I just learned that Red Hat provides an Azure RHEL image builder (but is behind RHN account) 15:10:36 <davdunc[m> mhayden: yes. 15:10:57 <dduffey> cpatterson is also out of the office today 15:10:57 <davdunc[m> dduffey: we know about it, but it doesn't support btrfs. 15:11:36 <davdunc[m> so we can't build the image with a shrinkable root volume. 15:12:01 <davdunc[m> cpatterson++ 15:12:09 <davdunc[m> for doing the heavy lifting on the patch. 15:12:52 <mhayden> i assume we just need to add the walagent package to the cloud kickstarts then? nothing else special outside of that (other than uploading/testing) 15:13:05 <davdunc[m> mhayden - yes. 15:13:28 <davdunc[m> so if you can get that rolled in, that would be ideal. 15:13:31 <mhayden> davdunc[m: i think my team and i could have a looksee at that 15:13:44 <davdunc[m> mhayden: I will dance at your wedding. 15:13:50 <spotz_> hehehe 15:14:09 <mhayden> #action mhayden and team to look at getting walagent into the Fedora cloud kickstart 15:14:15 <mhayden> #link Fedora kickstarts repository: https://pagure.io/fedora-kickstarts/tree/main 15:14:26 <davdunc[m> mhayden thanks. 15:14:43 <mhayden> i'll need to ask someone how to run a build again from there -- it's been a while 15:14:43 <davdunc[m> okay. I am going to call that good. 15:14:53 <mhayden> if anyone has a link to docs for that, i'd be grateful 15:15:17 <davdunc[m> Conan Kudo: can tell you. After integrating kiwi with the CentOS Hyperscale work, he can speak volumes to it. 15:15:34 <mhayden> sweet π 15:15:41 <davdunc[m> #topic Change Proposals 15:16:18 <davdunc[m> we have three change proposals out for the AMIs. 15:16:33 <davdunc[m> one is support for the new uefi-preferred configuration. 15:17:02 * davdunc[m looks up the link. 15:18:00 <davdunc[m> I can't find it. 15:18:10 <davdunc[m> I'll add it after the fact in the followup email. 15:18:11 <mhayden> hah, i was looking, too. π₯² 15:18:23 <davdunc[m> itβs tested and it works when the images are booted UEFI. 15:18:56 <mhayden> well that's a plus 15:18:58 <davdunc[m> this change will boot uefi on all the instance types that will support it and then fall back to classic boot on the ones that don't. 15:19:22 <davdunc[m> we already did the work to support it in F35 F36 15:19:37 <mhayden> sounds like a plan -- has the proposal gone through FESCo yet? 15:19:39 <davdunc[m> So this is a great way to shave off some boot time. 15:20:03 <davdunc[m> mhayden: I don't think this one has. It might need to be wrangled yet. 15:20:12 <davdunc[m> The others have. 15:20:33 <mhayden> if you find the proposal, let me know and I can review it 15:20:50 <davdunc[m> okay. I'll pass it over after the meeting. 15:20:57 <davdunc[m> Here's the second one: 15:20:58 <mhayden> i see a couple of other cloud issues in pagure (such as removing the standard storage AMIs) that would make for a quick proposal 15:21:23 <mhayden> #action davdunc to dig up the proposal for UEFI boot in cloud images and send it to mhayden for review 15:21:29 <davdunc[m> yes. 15:22:01 <davdunc[m> that's perfect. It's set for F39, but we may want to respin the images for F38 to get the benefits. 15:22:23 <mhayden> sounds reasonable 15:22:23 <davdunc[m> So onto the second one: Change the default volume type to gp3 15:22:34 <davdunc[m> #link https://fedoraproject.org/wiki/Changes/CloudEC2gp3 15:23:09 <davdunc[m> This one is really huge change and we just have to change one little number in the config for the volume creation. :) 15:23:34 <mhayden> hah, well said 15:23:36 <davdunc[m> it decreases volume costs by around 10% and increases performance on snapshots and the like. 15:23:51 <mhayden> it has the green light from FESCo -- where do we make this change? 15:24:16 <nzwulfin> sounds like an easy first issue if it's really that simple 15:24:32 <davdunc[m> it's in the fedimg code. Someone from the infrastructure team has already gotten excited about doing it across the volumes for infrastructure. 15:24:59 <davdunc[m> nzwulfin: yes. it's a modification "ye olde fedimg" code. 15:25:07 <davdunc[m> modification to* 15:25:16 <mhayden> that sounds interesting 15:25:19 * nzwulfin pretends he knows what that means and nods 15:25:35 <davdunc[m> #chair nzwulfin 15:25:35 <zodbot> Current chairs: Felix davdunc[m dduffey mhayden miyunari nzwulfin spotz_ 15:25:59 <mhayden> is this the right code: https://github.com/fedora-infra/fedimg ? 15:26:26 <davdunc[m> we have a build script called fedimg and it builds the image - there is a DEVICE Definition that identifies the volume type as "gp2" 15:26:54 <davdunc[m> sorry, not DEVICE, it's a Disk Container 15:27:16 <davdunc[m> yes. mhayden that's the right code and can we get your team as owner? 15:27:38 <mhayden> https://github.com/fedora-infra/fedimg/blob/fcbface13726cd1cb09b007b510bc888116a56e0/fedimg-conf.toml.example#L13 15:28:04 <mhayden> davdunc: i wonder if i can get miyunari[m] to have a looksee at a PR π€π€π€π€π€ 15:28:17 <davdunc[m> While I appreciate that Sayan did a lot of work on it, he's not maintaining it for us any longer. 15:28:28 <nzwulfin> the changelog for hp2 is pre-PR tracking 15:28:35 <nzwulfin> gp2* 15:28:55 <davdunc[m> spotz_: I just realized that we are clobbering the centos-cloud meeting again. 15:29:02 <davdunc[m> will fix that. 15:29:21 <spotz_> Yeah, hence IRC vs Matrix:) It's just easier to see action and switch channels 15:29:30 <nzwulfin> @mhayden i might ping you offline on this one :) 15:29:39 <davdunc[m> nzwulfin: thanks. 15:29:41 <mhayden> nzwulfin: oh, the fedimg one? sure, let's do it 15:29:56 <davdunc[m> okay. While you're doing that one. 15:30:03 <davidduffey> What is your favorite / a good matrix client? (using hexchat at the moment) 15:30:12 <davdunc[m> have a look at this third one: 15:30:25 <mhayden> #action nzwulfin and mhayden to sync up on changing gp2 > gp3 for Fedora cloud image 15:30:30 <davdunc[m> dduffey: element. 15:30:40 <spotz_> davidduffey: For Matrix I use Element for IRC I use Limechat (I'm a bad Open Sourcer I'm on a MAc) 15:31:13 <spotz_> But Element can be used in a browser which is handy so not just an app 15:31:17 * mhayden uses irccloud 15:31:36 <mhayden> davdunc: what else on the AMI side? should i draft a proposal to remove standard AMIs? 15:31:37 * davdunc[m likes irccloud it's so convenient. 15:31:59 <davdunc[m> mhayden: I think that's appropriate. We don't need them any longer.. 15:32:22 <mhayden> nzwulfin: i just re-read the gp2>gp3 change proposal -- the approved proposal is a switch from 2 to 3, not adding 3 now and removing 2 later 15:32:25 <mhayden> i couldn't remember 15:32:29 <davdunc[m> The third one is moving to the Instance Metadata Service (IMDS) version 2 15:32:50 <mhayden> #action mhayden to draft a change proposal for removing standard storage images https://pagure.io/cloud-sig/issue/295 15:32:57 <davdunc[m> nzwulfin: it should just be switching from 2 to 3. 15:32:57 <nzwulfin> @mhayden: yep reading up in another window on the history 15:33:06 * nzwulfin nods 15:33:45 <davdunc[m> If a user decides to swap back to gp2 for $reasons then they can modify the config at instance creation. 15:34:02 <mhayden> oh i didn't know that 15:34:03 <davdunc[m> #link https://fedoraproject.org/wiki/Changes/CloudEC2IMDSv2Only 15:34:27 <davdunc[m> mhayden: it's super convenien. 15:34:30 <davdunc[m> t* 15:34:54 <davdunc[m> that dot was just a placeholder for the promise of a "t" 15:35:15 <mhayden> π 15:35:31 <davdunc[m> so the goal with this one is to tokenize the access to the IMDS service. 15:35:39 <mhayden> so what's the action required on the IMDSv2 from us? cloud-init already supports it 15:36:10 <davdunc[m> yes. cloud init supports it. 15:36:51 <davdunc[m> the cloud-init config that fredlef created tries the imdsv2 service first and then falls back to the imdsv1 support. 15:37:21 <davdunc[m> if we do this, then the images will support IMDSv2 only and that is the recommendation of the Amazon security team. 15:38:22 <davdunc[m> decreasing the max hops thwarts complications that happen as a result of misconfigured proxy servers or manually configured load balancing solutions. 15:39:11 <davdunc[m> So it's generally considered to be a better route to go. 15:40:09 <mhayden> so this one also has the FESCo green light π 15:40:12 <davdunc[m> it is supported in all of the regions that we support today. 15:40:32 <davdunc[m> awesome. it's a change that we can support in the libcloud definition too. 15:40:59 <davdunc[m> So while I move fedimg away from libcloud to the native tools, we can still effect the change today. 15:42:02 <davdunc[m> okay. That's the three changes there. 15:43:32 <davdunc[m> mhayden, we may want to talk to nirik about it. There was someone from the copr team iirc who was already working on some of these for their instances and willing to make the changes for the rest of us. 15:44:01 <nirik> yeah, see the infrastructure list/ansible commits. 15:44:21 <davdunc[m> I'll go dig that up. thanks nirik 15:45:23 <davdunc[m> #action davdunc chase down the modifications in the ansible and then modify the flag if mhayden and team don't beat him to it. 15:46:03 <davdunc[m> moving on. 15:46:10 <davdunc[m> #topic cloud directory 15:46:22 <davdunc[m> #topic cloud image directory 15:47:32 <davdunc[m> mhayden: do you have the link for that? 15:48:01 <mhayden> #link https://fedora.imagedirectory.cloud/ 15:48:19 <mhayden> this is totally a PoC / experiment because i don't know how to make wireframes π 15:48:39 <davdunc[m> well I am so excited for the work and the goals. 15:48:51 <mhayden> my team (miyunari[m] + Felix[m]123 and others) have been working on an image directory frontend/backend application that makes it easier to find images in public clouds 15:48:57 <davdunc[m> and excited that you and the rest of the team are working this! 15:49:05 <davdunc[m> miyunari++ 15:49:06 <zodbot> davdunc[m: Karma for miyunari changed to 1 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:49:07 <mhayden> for example, we have one for RHEL that we are developing, too: https://imagedirectory.cloud/ 15:49:12 <davdunc[m> Felix++ 15:49:27 <davdunc[m> Felix++ 15:49:32 <mhayden> *however*, this is just a POC. there's no accessibility handling and fedora branding or anything like that 15:49:44 <davdunc[m> we understand! 15:50:10 <mhayden> what i'd like to see is if we can somehow get this into the fedora website in a fedora-centric format that replaces the existing cloud image listings 15:50:31 <davdunc[m> So as this progresses, I, for one, and am excited to see it. I can't wait until we are talking about it in action and using it for deployments! 15:50:36 <mhayden> i have no idea how to propose something like that and would be glad to take advice on that 15:51:00 <mhayden> our teams' goal is to provide a frontend for clicking through the options as well as a JSON backend for tools and such 15:51:55 <davdunc[m> well we have some great help on the website side. As they go through their rebuild that will be great. They were previously depending on us to provide the images through the fedimg reporting and that was json. I haven't looked at what they are doing now. 15:52:26 <mhayden> i found the website content on gitlab recently so that could be a start 15:52:36 <mhayden> who do we talk to on the web team for a potential update there? 15:52:40 <dustymabe> lots of people in the meeting today! 15:52:43 <davdunc[m> I know it's different. This gives us a chance to get back to providing something reliable that they can depend on. 15:52:54 <davdunc[m> #chair dustymabe 15:52:54 <zodbot> Current chairs: Felix davdunc[m dduffey dustymabe mhayden miyunari nzwulfin spotz_ 15:53:03 <davdunc[m> happy to see you here too dustymabe 15:53:14 <davdunc[m> alright 15:53:24 <davdunc[m> #topic open_floor 15:53:46 <nzwulfin> o/ 15:53:53 <davdunc[m> We need to think about our scheduling for test days. 15:54:05 <mhayden> miyunari++ 15:54:05 <zodbot> mhayden: Karma for miyunari changed to 2 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:54:06 <davdunc[m> nzwulfin you have a topic? 15:54:15 <nzwulfin> i have a question really :) 15:54:23 <davdunc[m> well by all means! 15:55:01 <davdunc[m> and I forgot to throw some appreciation Stewart's way for working on the change proposals. 15:55:03 <nzwulfin> saw an email on the list about docs (that had little context), is there an official "improve our docs" project / ticket / effort? 15:56:10 <davdunc[m> nzwulfin: there is not an official project. we have a new contributor who was interested in diving in with us and I asked them to take up documentation for the work. 15:56:35 <nzwulfin> cool, i'll throw my hat on that pile in the list then 15:56:50 <davdunc[m> @mayura_andrew 15:57:02 <davdunc[m> We really need it. 15:57:34 <nzwulfin> i read your email and concur :D 15:57:36 <davdunc[m> nzwulfin, would you mind including @mayura_andrew in your efforts and help get them onboarded? 15:57:53 <nzwulfin> sure! i need to get me onboarded too 15:58:14 <davdunc[m> nzwulfin: well any help either of you need, we are here to help. 15:58:48 <davdunc[m> happy to have you jumping in. 15:59:36 <nzwulfin> got smart and made SIG work a part of my goals 15:59:42 <mhayden> woot! 15:59:55 <spotz_> hehe 16:00:07 <spotz_> I just make Open Source projects one of mine:) 16:00:34 <nzwulfin> bugging @mhayden as OKRs 16:00:46 <mhayden> π₯΅ 16:00:48 <spotz_> Don't forget blaming him as another:) 16:00:53 <mhayden> i think we're at time 16:00:57 <nzwulfin> lol 16:00:59 <davdunc[m> so the test days. We missed it in the F38 beta space, but we need to make sure we are working with Sumantro Mukherjee and his team to ensure that we are properly getting onto the schedule. He is doing everything he needs to do, but I haven't followed up. 16:01:01 <spotz_> hehe we are 16:01:02 <davdunc[m> okay. 16:01:11 <davdunc[m> well thanks for being here everyone. 16:01:18 <davdunc[m> We'll do it again soon! 16:01:25 <davdunc[m> #endmeeting