15:00:51 #startmeeting fedora_cloud_meeting 15:00:51 Meeting started Thu Nov 11 15:00:51 2021 UTC. 15:00:51 This meeting is logged and archived in a public location. 15:00:51 The chair is davdunc. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 15:00:51 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:51 The meeting name has been set to 'fedora_cloud_meeting' 15:01:05 #topic roll call 15:01:21 .hi 15:01:23 dustymabe: Something blew up, please try again 15:01:26 dustymabe: An error has occurred and has been logged. Please contact this bot's administrator for more information. 15:01:26 .hi 15:01:29 dustymabe: Something blew up, please try again 15:01:32 dustymabe: An error has occurred and has been logged. Please contact this bot's administrator for more information. 15:01:41 .hi 15:01:42 mhayden: Something blew up, please try again 15:01:45 mhayden: An error has occurred and has been logged. Please contact this bot's administrator for more information. 15:01:46 great zodbot - nb seen this before? 15:01:49 πŸ”₯ 15:01:56 looks like the bots are striking today 15:01:58 hi my name is: Dusty Mabe (dusty@dustymabe.com) 15:03:01 oh no. 15:03:32 .hello mhayden 15:03:33 mhayden: Something blew up, please try again 15:03:36 mhayden: An error has occurred and has been logged. Please contact this bot's administrator for more information. 15:03:39 πŸ‘‹ 15:03:43 πŸ€”πŸ€·πŸ»β€β™‚οΈ 15:03:45 #chair mhayden 15:03:45 Current chairs: davdunc mhayden 15:04:00 #chair dustymabe Eighth_Doctor 15:04:00 Current chairs: Eighth_Doctor davdunc dustymabe mhayden 15:04:15 #chair Eighth_Doctor 15:04:15 Current chairs: Eighth_Doctor davdunc dustymabe mhayden 15:04:18 perhaps the cloud has not been friendly to zodbot 15:04:30 hopefully at least the meeting logs still work 15:04:31 we love zodbot 15:04:57 bless it's heart. 15:05:14 #chair dcavalca 15:05:14 Current chairs: Eighth_Doctor davdunc dcavalca dustymabe mhayden 15:05:17 πŸ‘‹ 15:05:27 #chair michel 15:05:27 Current chairs: Eighth_Doctor davdunc dcavalca dustymabe mhayden michel 15:06:07 well, I didn't get a direct "blew up" message, so I am going to continue as if the logging is working. 15:06:23 if not, it won't be the first time I recreated the meeting minutes by hand. 15:06:37 okay. 15:06:52 #topick Action items from last meeting 15:07:01 #topic Action items from last meeting 15:07:06 without the "K" 15:07:33 no action items from last week. 15:07:44 last *meeting 15:08:20 grabbing what i think should be first. 15:08:38 #topic Cloud Working Group Correction Of Errors Report 15:08:49 #link https://pagure.io/cloud-sig/issue/352 15:09:17 So I'll start with the big question 15:09:26 Should we focus on being an edition again? 15:09:53 seems like nothing is happening? 15:10:04 Eighth_Doctor: 15:10:17 yes, I think we should 15:10:21 I think that we should 15:10:37 seems like Fedora Cloud isn't really an edition, but more like an opinionated installation that is packaged in unusual ways (relative to other editions) 15:10:54 similar to Fedora in containers 15:10:59 mhayden: that's kinda the definition of an edition though :) 15:11:01 who's chairing this? (been away for too long) 15:11:12 I have heard that argument before mhayden. 15:11:15 dcavalca: point taken πŸ˜‰ 15:11:29 I'll note that practically speaking being an edition is really only a promotion to toplevel display on fedora websites and such, which I think is useful 15:11:34 michel: michel I think I am. 15:11:47 dustymabe: top level display is definitely worth it 15:11:49 #chair michel 15:11:49 Current chairs: Eighth_Doctor davdunc dcavalca dustymabe mhayden michel 15:11:57 the reason this hasn't come up for so long is because practically speaking, nothing changed with the Fedora Cloud image - people could still use it as they had before 15:11:57 s/worth it/worthwhile/ 15:12:08 mhayden: I don't disagree. I agree in fact! 15:12:34 just stating that the reason we let it "slide into the shadows" is because technically speaking nothing really changed 15:12:39 dustymabe: I think that' we already have a lot of groundwork here too. 15:12:43 so is publicity/awareness the main benefit from becoming an edition? 15:12:59 mhayden: I think that's the primary benefit (from my perspective) 15:13:16 There's probably other benefits (like being considered "blocking") 15:13:23 mhayden: I think it's a commitment to actively participate in the goals related to the community and marketing and product management too. 15:13:51 beyond the marketing stuff (which is valuable), it's also awareness within Fedora itself 15:13:54 so blog posts and solutions efforts, along with commitments to meet deadlines. 15:13:59 davdunc should be 15:14:01 editions at the end of the day are the deliverables of the project 15:14:11 sure - I think all of these points are valid 15:14:17 +1 dcavalca 15:14:19 but what I will say... 15:14:26 i'm definitely in support of the additional edition work (pardon the pun) 15:14:30 is that none of this stuff will happen without us making it happen 15:14:38 dustymabe is right. 15:14:48 we have to be willing to sweat a little. 15:14:53 if we get the "edition" stamp and then do what we've done for the past 5 years, nothing will change 15:15:01 and take responsibility. 15:15:10 right, it's "become addition and... " 15:15:17 edition 15:15:20 πŸ€¦πŸ»β€β™‚οΈ 15:15:25 the main short-term impact I can see is getting included in the website redesign that's being discussed 15:15:31 similarly, if we don't get the edition stamp, but still put in the sweat equity - we'd probably be at about the same place we would be WITH the edition stamp 15:15:48 what i'm mostly stating is that our destiny is mostly controlled by us, with or without the edition stamp 15:15:49 dcavalca: we were on-track for that too. I was the representative. 15:16:24 .hi salimma 15:16:25 michel: Something blew up, please try again 15:16:27 dustymabe: we made what I think are some big advances in f35. 15:16:28 michel: An error has occurred and has been logged. Please contact this bot's administrator for more information. 15:16:46 * dustymabe would also like to say that he's very happy to see everyone here today and passionate about pushing Fedora Cloud forward 15:17:00 .hi salimma 15:17:01 michel-slm: Something blew up, please try again 15:17:04 michel-slm: An error has occurred and has been logged. Please contact this bot's administrator for more information. 15:17:09 sigh 15:17:16 dustymabe: I suspect we're all violently agreeing here :) 15:17:20 welcome michel-slm 15:17:22 oh yea. dustymabe you made this a great group for us to participate in. 15:17:24 dcavalca: indeed 15:17:33 thanks dustymabe 15:17:34 #chair michel-slm 15:17:34 Current chairs: Eighth_Doctor davdunc dcavalca dustymabe mhayden michel michel-slm 15:17:46 we'll get you in there michel-slm 15:17:57 * mhayden will commit time to doing the needful to get fedora cloud as a full edition 15:18:02 switching back to IRC, the timestamps in Matrix are weird. Got a heart attack when I saw davdunc and dustymabe having messages supposedly from 15 mins ago 15:18:02 cool. 15:18:45 so let's just have a vote for the sake of having it resolved amongst ourselves. 15:18:46 mhayden: is that an action item? 15:19:04 if becoming an edition is step 1, we could then move on to making fedora cloud a better cloud citizen. better cloud integration/functionality 15:19:11 dustymabe: i think so 15:19:32 .hello ngompa 15:19:33 Eighth_Doctor: Something blew up, please try again 15:19:36 Eighth_Doctor: An error has occurred and has been logged. Please contact this bot's administrator for more information. 15:19:40 i got an email from someone this morning who wants to work on more cloud tooling + SDKs so fedora can manage the cloud that is running underneath it 15:19:53 so we have room to grow and mature πŸŽ‰ 15:20:02 nice 15:20:07 davdunc: +1 for vote 15:20:27 I don't remember the meetbot vote function 15:20:32 * mhayden is happy to make an action item but doesn't want to step on davdunc's toes 🦢🏻 15:20:41 just say #proposed this is my proposal 15:20:45 got it. 15:20:46 then people +1 or -1 15:20:53 then say #agreed this is my proposal 15:21:26 #proposed as a group we will commit time and resources to making Cloud Base image an Edition once again. 15:21:42 some people summarize the vote count but that's also manual right, the bot doesn't help 15:22:05 +1 - though i would like to say I can' 15:22:10 +1 15:22:11 can't put significant work into it 15:22:12 +1 15:22:18 +1 15:22:39 dustymabe: we don't expect that. we know you are trying to spin this out. We need you as an advisor. 15:23:01 +1 15:23:16 Eighth_Doctor: you sorting it out over there? 15:23:45 though your voice is clear in the Fedora Blockers log. 15:24:15 #agreed We will work to be added back as an edition 15:24:31 #chair Eighth_Doctor 15:24:31 Current chairs: Eighth_Doctor davdunc dcavalca dustymabe mhayden michel michel-slm 15:25:05 okay then. . . I think that one is done. 15:25:13 #action mhayden to create a tracking ticket (if one doesn't already exist) and lead up efforts to re-instate Cloud as an edition 15:25:13 yes for sure 15:25:19 we should be an edition 15:25:46 cool. 15:25:52 More on this in later topics too. 15:25:59 moving away for a sec. 15:26:05 #topic default user for qcow2 base cloud image 15:26:15 #link https://pagure.io/cloud-sig/issue/348 15:26:47 I saw this come up in the issues and I wondered if we wanted to look into the request and why it was complicated for the packer scenario. 15:26:56 That was unexpected. 15:27:37 packer seems to work well with cloud-init, so I wasn't sure what the ask was here and it we needed to publish something for guidance. 15:27:55 if* we needed to publish something. 15:27:55 * mhayden reads the issue 15:28:38 this feels like a docs/awareness issue 15:28:41 seems like a documentation request, since OP is reporting two different problems 15:29:00 yea. That's what I thought michel-slm 15:29:02 i always have to scratch my head on the user to use for fedora/rhel/ubuntu/debian in clouds πŸ™ƒ 15:29:37 mhayden: i did for a while, but now having the distro as the default user makes plenty of sense to me. 15:29:46 mhayden: your config could enforce a user 15:30:01 to make matters worse, some clouds make their own fedora images and then use root πŸ€¦πŸ»β€β™‚οΈ (but that's another topic for another time) 15:30:23 dustymabe: my "default user is davdunc typically" no matter wahat. 15:30:45 mhayden: we should try to help them along in those cases, but another day. 15:30:59 for this particular problem I don't know enough about packer to help them 15:31:17 "starting automated install..." sounds like it's trying to run anaconda 15:31:23 but I really have no idea 15:31:25 this seemed like a place where we could make a big difference in the use by just documenting a path forward for using the cloud-init defaults. 15:31:45 davdunc: documentation would be really nice 15:31:52 Eighth_Doctor: helped him by pointing that out. 15:32:08 something similar to https://docs.fedoraproject.org/en-US/fedora-coreos/ 15:32:34 yes dustymabe very much like that. 15:32:45 and in asciidoc too . 15:33:09 which is another thing we should cover in the move to an edition (not today though) 15:33:21 okay 15:33:37 any takers on researching the issue with the user? 15:34:37 #action davdunc reproduce the issue and outline documentation for Issue 348 15:34:46 I'll take it. 15:36:05 anyone else have anything for this? 15:36:50 #topic Add Ignition support to Fedora Cloud 15:37:03 #link https://pagure.io/cloud-sig/issue/329 15:37:11 I think this was something maybe jdoss was interested in 15:37:20 I could see that. 15:37:52 basically making a decision to use cloud-init if it is yaml or script? 15:38:24 yeah, i mean I can't imagine us dropping cloud-init - so we'd probably have to make it either/or 15:38:41 Eighth_Doctor mhayden how much lift do you see here? 15:38:48 switching to ignition for good would allow us to remove a lot of python-related deps from the image, but it would make fedora an oddball in clouds 15:38:57 yea. 15:39:01 agreed, I think we'll have to keep both 15:39:11 would be sweet to offer both potentially 15:39:15 A lot of documentation from providers would just fail. 15:39:36 it seems like a significant amount of effort though. 15:39:39 quite true, and that's the last thing we would want to do 15:40:07 Can we look at making this a change proposal? 15:40:11 there's also the option of, down the line, offering separate images for both 15:40:21 but that's probably overkill 15:40:48 michel-slm: that might be easier in the short-term than creating a failover mechanism. 15:40:57 unless there is already one in ignition. 15:41:29 * mhayden is glad we 'ignited' this conversation at least 🀭 15:41:42 * davdunc as well! 15:42:11 hm? 15:42:13 Not sure what ignition does right now.. probably barfs because the config isn't valid Ignition config, but we could probably change it to ignore 15:42:15 * Eighth_Doctor needs to switch computers 15:43:12 yea. if it is just a change in the load order, that's no big deal. 15:43:19 #chair King_InuYasha 15:43:19 Current chairs: Eighth_Doctor King_InuYasha davdunc dcavalca dustymabe mhayden michel michel-slm 15:43:39 so can someone ask the question again? 15:43:51 this room was out of sync on matrix and I just noticed that, hence switching machine 15:43:54 King_InuYasha: https://pagure.io/cloud-sig/issue/329 Add ignition support 15:44:36 King_InuYasha: was wondering if that's why you've been so quiet :) 15:44:46 hah, it's because none of the messages were making it through 15:45:23 I don't know much about ignition+cloud-init on the same image 15:45:30 King_InuYasha: yeah, I mentioned that in our group chat but I wondered if you don't see that either 15:45:37 there needs to be some logic to one over the other and cloud-init is slow enough as it is. 15:45:41 yeah 15:47:38 I think the right way would be to check if ignition config is provided on boot, and if it isn't then ignition queues a file to be written to disable itself for cloud-init 15:47:38 but I see a lot of value in the discussion and finding the non-convergent use-cases. 15:47:38 since ignition doesn't run on every boot, we can eat that cost on the first boot 15:47:38 i don't really think boot will be any slower 15:47:38 well 15:47:38 i take that back 15:47:38 dustymabe: we'll find out, but I suspect you are right. 15:47:38 ignition operates in the initramfs - and thus needs networking to be brought up there 15:47:38 it will take some decision time. 15:47:40 so that's the biggest "slowdown" that I can think of 15:48:03 but once it's up it then will see the config isn't for ignition and will NOP 15:48:46 well, it doesn't need networking to determine that 15:48:50 then cloud-init will see the default config and act without regard for ignition. 15:49:01 it ddoes King_InuYasha 15:49:12 for ec2 metadata there has to be link local 15:49:12 how come? 15:49:16 for ignition? 15:49:22 yes. 15:49:28 King_InuYasha: sure it does need networking to determine that 15:49:50 needs to contact a metadata service (depending on the cloud) to get the config 15:50:15 some are direct device, but there are a bunch that rely on that eucalyptus-ago model. 15:51:09 so we need a change proposal and then we can work out the details. 15:51:35 well.. i'd say first we need a volunteer to drive the work 15:51:41 without that it's dead in the water anyway 15:51:58 the ignition thing is jdoss' baby 15:52:04 dustymabe: I believe you are correct. 15:52:34 well, do we want to see if jdoss is interested in joining us on this journey? I would enjoy that very much. 15:52:35 yeah, if jdoss wants to drive it +1 15:53:33 okay. I'll take an action item to work with jdoss and find out if there is bandwidth and if there is sufficient interest in making it work. 15:54:34 #action davdunc to follow up with jdoss on driving the adoption of ignition along with cloud-init 15:55:15 one more. 15:55:25 actually. 15:55:33 I think we are down to time. 15:55:45 let's go to open floor. 15:55:56 #topic open floor 15:56:09 Anyone have anything they want to cover specifically. 15:56:10 ? 15:56:32 Not I. Thanks for running the meeting. 15:57:07 King_InuYasha: brought up edk2 state in fedora this morning. That's falling behind RHEL even. We should keep our eye on that. 15:57:20 yeah 15:57:25 the PMBR bug isn't fixed in Fedora 15:57:31 it's fixed in RHEL, SLE, and openSUSE 15:57:41 thanks davdunc 15:57:44 Fedora Release party is tomorrow everyone. 15:57:59 are you doing something for the release party? 15:58:03 is edk2 maintained by a different person in Fedora vs RHEL? 15:58:21 King_InuYasha: I am giving a talk on the State of the cloud image. 15:58:22 maybe we should ping both sides and see if there's ever an attempt to merge changes back, because this is weird 15:58:24 King_InuYasha: do you have a link? 15:58:41 michel-slm: it is. 15:58:45 * michel-slm didn't see any schedule for the release party in hopin, is it elsewhere? 15:58:52 dustymabe: https://bugzilla.redhat.com/show_bug.cgi?id=1978966 15:59:06 another member of the community wants to work on packaging more cloud tools in Fedora -- hopefully I can report back on that soon 15:59:08 he suggested a SIG 15:59:21 mhayden: i'm all for that existing! 15:59:24 thanks for your hard work 15:59:27 mhayden: bring them here! 15:59:32 well, cloud tools would go in this WG, wouldn't it? 15:59:34 will do 15:59:43 King_InuYasha: I think so too. 15:59:47 that's kinda our thing after all 15:59:52 for sure. 16:00:07 alrighty, we should get out of the way for the next group. 16:00:14 thanks everyone for attending! 16:00:26 King_InuYasha: sounds like there's not much for us to do but wait on the package maintainer? 16:00:34 probably, yeah 16:00:41 I think that's right. 16:00:45 it's just concerning because it prevents us from dropping hacks in our images 16:00:54 #link https://bugzilla.redhat.com/show_bug.cgi?id=1978966 16:01:04 yes. 16:01:23 okay. we'll carry over to #fedora-cloud now. 16:01:31 #endmeeting