15:00:45 <davdunc> #startmeeting fedora_cloud_meeting
15:00:45 <zodbot> Meeting started Thu Jan 20 15:00:45 2022 UTC.
15:00:45 <zodbot> This meeting is logged and archived in a public location.
15:00:45 <zodbot> The chair is davdunc. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
15:00:45 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:45 <zodbot> The meeting name has been set to 'fedora_cloud_meeting'
15:00:50 <dustymabe> .hi
15:00:51 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dusty@dustymabe.com>
15:00:58 <davdunc> #topic roll call
15:01:11 <davdunc> chair dustymabe  Eighth_Doctor
15:01:22 <davdunc> #chair dustymabe Eighth_Doctor
15:01:22 <zodbot> Current chairs: Eighth_Doctor davdunc dustymabe
15:01:30 <davdunc> gotta make it work.
15:01:37 <Eighth_Doctor> .hello ngompa
15:01:40 <zodbot> Eighth_Doctor: ngompa 'Neal Gompa' <ngompa13@gmail.com>
15:02:03 <davdunc> #chair ngompa
15:02:03 <zodbot> Current chairs: Eighth_Doctor davdunc dustymabe ngompa
15:02:11 <mhayden> .hello mhayden
15:02:12 <zodbot> mhayden: mhayden 'Major Hayden' <mhayden@redhat.com>
15:02:18 <davdunc> #chair mhayden
15:02:18 <zodbot> Current chairs: Eighth_Doctor davdunc dustymabe mhayden ngompa
15:02:48 <davdunc> did I miss anyone?
15:03:00 <davdunc> gonna give it another minute.
15:04:19 <davdunc> #topic Action items from last week
15:05:21 <davdunc> davdunc to  create a release checklist
15:06:08 <davdunc> well. I got caught up in doing all the writing for the edition effort.
15:06:16 <davdunc> So it's still on for this week.
15:06:28 <davdunc> I'll have a draft by Monday for everyone to review.
15:07:03 <dustymabe> +1 - sorry for not following up with you on that link from earlier in the week - been slammed over here :(
15:07:26 <davdunc> yea. I see your workload dustymabe
15:08:05 <davdunc> #action davdunc to create the wiki checklist for launch
15:08:08 <davdunc> :)
15:08:21 <davdunc> I'll follow up with you on that dustymabe
15:08:36 <davdunc> * mhayden to talk to ericedens about packaging the google-cloud-sdk 📦
15:08:55 <davdunc> mhayden: any progress there?
15:09:02 <dustymabe> thank you davdunc for the work you are doing here!
15:09:04 <mhayden> i got referred to someone who knows the components well, and there's an uphill battle...
15:09:14 <dustymabe> uphill battle.... being?
15:09:22 <mhayden> Google's CLI utilities require Python versions from 2.7.5 to 3.10 🙃
15:09:32 <davdunc> oh my.
15:09:40 <dustymabe> i.e. they need different versions installed at the same time?
15:09:51 <mhayden> there is some bundled stuff, which could be made to work, but i got some tips and links to stuff they publish for packagers
15:09:56 <mhayden> dustymabe: yes
15:10:11 <mhayden> it looks like the storage utility (gsutil) might be easiest to do, but gcloud is pretty wild
15:10:32 <davdunc> ah.
15:10:36 <mhayden> apparently it's being worked on at google, but no ETA currently
15:10:47 <davdunc> They have various cli tools not all built from a central team?
15:10:48 <dustymabe> I supposed they don't have concrete plans for updating everything (assuming they want to update and don't want to stay on old python)
15:10:50 <mhayden> so i'll see if maybe gsutil would be easy to get going
15:11:06 <davdunc> mhayden: that's great!
15:11:08 <mhayden> davdunc: it appears the case from just examining the code, but i am not 100% sure
15:11:27 <davdunc> yea. there are always dependencies. :D
15:11:30 <mhayden> also the scripts/spec files they use for making the SDK zips/rpms are proprietary/confidential 😭
15:11:36 <dustymabe> mhayden: are there pieces that aren't open source (I thought that wasa part of the problem)
15:11:41 <mhayden> i'll keep poking at it
15:11:48 <mhayden> dustymabe: indeed
15:11:59 <mhayden> so i wasn't bad at searching after all 😆
15:12:00 <Eighth_Doctor> mhayden: I usually ignore their spec files anyway, since they're horrible
15:12:40 <mhayden> on a related note, i made an error on one of the Azure SDK packages that broke the azure-cli updates for F35 -- working on fixing that up today
15:12:54 <davdunc> awesome!
15:13:07 <mhayden> #action mhayden to attempt packaging of gsutil and re-evaluate gcloud after
15:13:12 <davdunc> if there is software, there will be bugs.
15:13:24 <davdunc> that's a great outcome.
15:13:25 <dustymabe> mhayden: so there are pieces that aren't open source? won't that block us?
15:13:27 <davdunc> Thanks mhayden
15:13:49 <mhayden> dustymabe: well, the method they use to assemble it all is closed, but the little pieces are all open source -- that might be enough
15:14:44 <davdunc> mhayden: if you need eyes over your shoulder to spur you on and ask quesitons, we are here for you!
15:15:07 <mhayden> thanks!
15:15:49 <davdunc> * davdunc to follow up with jdoss on driving the adoption of ignition
15:16:04 <dustymabe> mhayden: I see
15:16:10 <davdunc> I haven't spoken to jdoss for more than a minute or two this past week.
15:16:13 <dustymabe> so maybe we can stitch it together ourselves, +1
15:16:22 <davdunc> dustymabe has been keeping him busy.
15:16:28 <davdunc> :)
15:16:32 <dustymabe> ha - I haven't talked to him either
15:17:00 <dustymabe> we miss him
15:17:15 <davdunc> but adding that back to the actions to discuss ignition this week!
15:17:48 <davdunc> yes! we miss his influence and effort!
15:18:19 <davdunc> #action to follow up with jdoss on driving the adoption of ignition
15:18:36 <davdunc> #action davdunc to follow up with jdoss on driving the adoption of ignition
15:18:50 <davdunc> there we go. "to" is not going to follow up on anything.
15:19:07 <davdunc> #action davdunc to look at image upload using libcloud
15:19:17 <Eighth_Doctor> I think jdoss changed jobs recently
15:19:21 <davdunc> I did look into it a bit, but there is more to unravel.
15:19:43 <davdunc> ah. well, that early period usually takes a lot of focus.
15:19:54 <davdunc> we'll look forward to having him back when he is ready.
15:20:47 <davdunc> mhayden: can you add the tracking ticket for the edition while we are talking about it?
15:20:59 <mhayden> aaaack, i meant to do that
15:20:59 <mhayden> yes
15:21:18 <davdunc> #topic Reestablishing Fedora Cloud Base as an edition
15:21:31 <davdunc> I got the proposal in too late.
15:21:59 <davdunc> I didn't think that we would be making a site-wide proposal since it was a reestablishment
15:22:34 <davdunc> but... bcotton informed me that there was not a second policy and that effectively it is defined as site-wide.
15:22:52 <davdunc> so we have a target of F37
15:23:45 <davdunc> I am sorry that I didn't make that assessment. I could have gotten the paperwork in on time. I just didn't prioritize it for site-wide change.
15:24:06 <davdunc> So now we have a technical specification
15:24:25 <dustymabe> davdunc: no worrie
15:24:34 <dustymabe> i missed the change deadline for a lot of FCOS things too
15:24:38 <davdunc> and we still need to update the PRD
15:24:49 <Eighth_Doctor> can we at least use this a thing to stall out that stupid bug about changing the os-release for F36
15:24:50 <dustymabe> I think the best way to not miss the next deadline is to just submit the change for F37 ASAP
15:24:50 <davdunc> #link https://fedoraproject.org/wiki/Cloud/Cloud_PRD?rd=Cloud_PRD
15:24:50 <Eighth_Doctor> ?
15:25:04 <Eighth_Doctor> it'd be pointless to change it in F36 to change it back in F37
15:25:12 <davdunc> dustymabe: it is submitted.
15:25:39 <davdunc> grabbing the link
15:26:03 <davdunc> #link https://fedoraproject.org/wiki/Changes/RestoreCloudEdition
15:26:12 <mhayden> #link cloud-sig tracker ticket for re-establishing Fedora cloud as a Fedora edition https://pagure.io/cloud-sig/issue/364
15:26:32 <davdunc> I would like to have everyone's contribution if they see fit.
15:27:17 <davdunc> I am interested in scheduling a PRD working meeting if others would like to execute on the review together.
15:27:36 <davdunc> in fact, I will do that in early February.
15:28:44 <davdunc> #action davdunc to schedule PRD review session.
15:28:55 <dustymabe> davdunc: I'll try to make it
15:29:02 <davdunc> we should be able to get that rolling after devconf is over.
15:29:03 <dustymabe> maybe we should get mattdm and bcotton involved
15:29:13 <davdunc> that would be ideal.
15:29:22 <dustymabe> i.e. advice on what we should include in the PRD for effectiveness
15:29:52 <davdunc> bcotton already gave me some great advice for updating the proposal and the PRD. I look forward to their further review.
15:30:07 <davdunc> The tech spec needs some forward looking moments too.
15:30:16 <davdunc> that's a living document in a major way.
15:31:08 <davdunc> Eighth_Doctor: You were saying that the os-release change is coming?
15:31:12 <bcotton> davdunc: i'd be happy to sit in on that PRD review session if i'm available
15:31:29 <davdunc> bcotton: you will most defiinitely be invited.
15:31:47 <Eighth_Doctor> davdunc: no, there's just a bug where adamw poked us about fedora-release-cloud os-release file saying "Cloud Edition"
15:32:01 <davdunc> aaah.
15:32:28 <Eighth_Doctor> it currently calls itself that, and I don't particularly want to undo it for F36 for because of process stuff that will ultimately get done "in F37"
15:33:03 <davdunc> I am not excited about making that change and then changing it back all within the course of ~6 months.
15:33:15 <davdunc> that's a recipe for disaster.
15:33:34 <davdunc> better to just commit to earning our stripes.
15:33:57 <davdunc> so we agreed. We are doing this!
15:33:59 <davdunc> :D
15:34:15 <Eighth_Doctor> also, technically being an edition doesn't require us to fiddle with the actual releng stuff
15:34:41 <Eighth_Doctor> so unlike most changes, it actually doesn't matter whether it's F36 or F37 since the changes are mostly procedural and marketing
15:35:34 <davdunc> and we are beholden for some other discussion points here. Things we can get done during the f36 timeframe.
15:35:43 <davdunc> especially for marketing.
15:36:19 <davdunc> We need to have a good blog put together on Cloud and make sure we have structured release information.
15:37:00 <davdunc> And I think we need to put together a transfer of any reasonable documentation to the docs.fedoraproject.org
15:38:05 <davdunc> okay.
15:38:19 <davdunc> I am moving on to the next topic unless someone has something else to add?
15:38:32 <mhayden> 👍🏻
15:38:39 <davdunc> #topic #353 Public cloud tooling
15:38:50 <davdunc> #link https://pagure.io/cloud-sig/issue/353
15:39:29 <davdunc> mikelo2 seemed to believe that we might need a cloud tooling sig (sub-sig)
15:40:11 <davdunc> I don't see it as a sub-sig. I see it as a direct duty of the cloud working group to maintain the packaging or at least follow the efforts where the primary maintainer doesn't participate.
15:40:49 <davdunc> does anyone see a reason that a sub-sig would be useful or necessary?
15:41:02 <Eighth_Doctor> not me
15:41:11 <Eighth_Doctor> I think it's an aspect of what we do in the first place
15:41:41 <davdunc> exactly and I think it points out something that should be effectively in our mission.
15:42:17 <davdunc> do we need to take a vote or just identify that at this time we don't consider this a necessary division of duties?
15:42:36 <davdunc> mhayden: do you have an opinion on this you want to share?
15:43:24 <mhayden> yeah, i think this is something we can do as part of the cloud sig
15:43:26 <davdunc> I don't want to vote it withotu mikelo2. If he feels strongly about it, we can reopen the request and have a proper vote.
15:43:28 * dustymabe only notes that he wishes we had more people in our army
15:43:40 <davdunc> dustymabe: no doubt!
15:43:44 <mhayden> it's kinda core to what i do at $dayjob lately
15:44:13 <davdunc> mhayden: I think your $dayjob is overlapping with our goals a lot at least right now.
15:44:27 <mhayden> haha, luckily it is!
15:44:40 <davdunc> it is great fortune for us.
15:45:17 <davdunc> okay - let's table this for now and call it a closed issue, but let's be willing to address it if the reporter  wants to address it again.
15:46:26 <davdunc> okay. awesome.
15:46:35 <mhayden> woot
15:47:09 <davdunc> #topic #356 Add AWS Systems Manager parameters for Cloud Images
15:47:19 <davdunc> #link https://pagure.io/cloud-sig/issue/356
15:47:43 <davdunc> This one is going to need another initiative in front of it.
15:48:06 <davdunc> I want to hammer out the details on Fedimg
15:48:53 <davdunc> there are public parameters for identifying the latest of an AMI in Amazon EC2.
15:49:16 <davdunc> we don't have a setting for those (or a naming convention published in the TRD)
15:49:31 <davdunc> moreover, we don't have a publishing tool that can set them.
15:49:49 <davdunc> we have been talking about moving to kiwi for a while.
15:50:03 <davdunc> so...
15:50:56 <davdunc> How could we establish an automated way to set the public parameters? We can write some cloudformation templates and add them to the infrastructure deployments
15:51:29 <davdunc> ansible can write the substitutions in the cloudforamtion and then build the stack.
15:51:36 * Eighth_Doctor doesn't know this stuff at all...
15:51:42 <davdunc> yea.
15:52:25 <davdunc> Eighth_Doctor: this creates a generic name for all regions to get amis for specific parameters.
15:52:35 <davdunc> it reads like a key with separators
15:52:46 <Eighth_Doctor> oh nice
15:53:27 <davdunc> `/etc/service/fedora-cloud-base-x86_64-latest`
15:53:38 <davdunc> something like that.
15:54:36 <Eighth_Doctor> it's a static file, right?
15:54:37 <davdunc> not exactly like that, but still creates consistency across all of the regions for a specifically registered machine image base.
15:54:43 <Eighth_Doctor> so we should put it in `/usr` rather than `/etc`
15:54:49 <davdunc> it's an ssm parameter.
15:54:58 <davdunc> it's metadata for instance launch.
15:55:10 <davdunc> I should use something better.
15:55:39 <davdunc> `/aws/service/fedora/cloud-base-x86_64-latest`
15:56:55 <davdunc> and then in a yaml template for cloudformation, we could add `{ Image_Id: /aws/service/fedora/....}` to get the same image in every region.
15:57:48 <davdunc> so instead of haivng to write a complicated describe-images call to get the same content, you would just use "latest"
15:58:04 <Eighth_Doctor> nice
15:58:17 <davdunc> or you could be specific.
15:58:41 <davdunc> `/aws/service/fedora/cloud-base-x86_64-20220112`
15:58:47 <davdunc> or something like that.
15:59:51 <Eighth_Doctor> hmm
15:59:51 <davdunc> #proposed Add AMI public parameters to our publication process
16:00:08 <Eighth_Doctor> davdunc: +1
16:00:20 <davdunc> mhayden?
16:00:28 <mhayden> +1
16:00:32 <mhayden> sorry, in two meetings
16:00:36 <davdunc> dustymabe?
16:00:42 <davdunc> understood mhayden
16:01:07 <davdunc> I guess dustymabe is busy. he is handling a lot today.
16:01:14 <dustymabe> ehh
16:01:48 <davdunc> +1 from me.
16:01:55 <dustymabe> I mean I'm not opposed, but I hardly think it's a high priority thing for the group in general. We have a lot of other foundational work we could be doing that would be far more important IMO.
16:02:10 <dustymabe> then again I don't think anyone is ever opposed to someone stepping up and offering to implement something
16:02:12 <davdunc> it's not a priority. I just want it on the list for specifications.
16:02:19 <davdunc> right.
16:02:21 <dustymabe> +1
16:02:42 <davdunc> #agreed Add the public parameters to publication process
16:03:02 <davdunc> #action davdunc to investigate the best path forward for implementation
16:03:28 <davdunc> let's try that again.
16:03:50 <davdunc> #action davdunc to investigate and report on the best path forward to use ami public parameters
16:04:00 <davdunc> okay.
16:04:08 <davdunc> #topic open floor
16:04:41 <davdunc> #link https://devconfcz2022.sched.com/event/siJu/fedora-and-centos-hyperscale-and-cloud-meetup
16:04:50 <davdunc> don't forget devconf.cz next week.
16:05:07 <Eighth_Doctor> do we have a calendar invite for that?
16:05:11 <Eighth_Doctor> otherwise I have a nasty feeling I'll forget about being there
16:05:27 <davdunc> Eighth_Doctor: I'll add it to the calendar.
16:05:29 <davdunc> https://sched.co/siJu
16:05:44 <davdunc> okay.
16:05:50 <Eighth_Doctor> cool
16:05:58 <dustymabe> wow devconf.cz already
16:06:05 <davdunc> I know!
16:06:08 <dustymabe> i barely have energy for it :(
16:06:20 <davdunc> every day has been full since the holiday!
16:06:35 <davdunc> okay.
16:06:40 <davdunc> we are over time.
16:07:07 <davdunc> thanks for being here.
16:07:08 <davdunc> #endmeeting