16:01:00 <mboddu> #startmeeting RELENG (2020-02-12)
16:01:00 <zodbot> Meeting started Wed Feb 12 16:01:00 2020 UTC.
16:01:00 <zodbot> This meeting is logged and archived in a public location.
16:01:00 <zodbot> The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:00 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:01:00 <zodbot> The meeting name has been set to 'releng_(2020-02-12)'
16:01:01 <mboddu> #meetingname releng
16:01:01 <mboddu> #chair nirik sharkcz pbrobinson pingou mboddu dustymabe ksinny jednorozec
16:01:01 <mboddu> #topic init process
16:01:01 <zodbot> The meeting name has been set to 'releng'
16:01:01 <zodbot> Current chairs: dustymabe jednorozec ksinny mboddu nirik pbrobinson pingou sharkcz
16:01:05 <nirik> morning
16:01:10 <mboddu> Morning nirik
16:01:56 * sharkcz is here today
16:02:01 <nirik> whats on the docket for today?
16:06:20 <mboddu> #topic #9131 Delete the rpms/python-notebook momento branch
16:06:25 <mboddu> #link https://pagure.io/releng/issue/9131
16:06:34 <mboddu> This is related to removing branches
16:06:46 <mboddu> I know there is a policy not to remove them
16:06:54 <mboddu> But lately we are seeing many requests
16:07:16 <nirik> yeah...
16:07:28 <mboddu> Another example: https://pagure.io/releng/issue/9068
16:08:08 <nirik> well, we can take it to fesco again... but not sure anything will change.
16:08:14 <smooge> here (sorry for delay)
16:08:24 * mboddu waves at smooge and sharkcz
16:08:30 <nirik> basically I am fine with us deleting things, but only if we have a script or other testable way to know that they are ok to delete.
16:08:39 <nirik> last time I asked for that, no one would write it
16:09:11 <mboddu> IIUC, it has to check if there is a build submitted from any of the commits from that branch, is that all they need?
16:09:21 <mboddu> the script has to check*
16:09:25 <nirik> or perhaps they should go to fesco... ie, we should just close->can'tfix and tell them if they want it to happen, get fesco to approve a process.
16:09:48 <nirik> right. and perhaps the 'archive' thing would work also
16:10:13 <mboddu> May be
16:10:17 <nirik> https://pagure.io/fedora-infrastructure/issue/8390
16:10:25 <nirik> .ticket 8390
16:10:26 <zodbot> nirik: Issue #8390: Create custom git refs to archive deprecated branches in dist-git - fedora-infrastructure - Pagure.io - https://pagure.io/fedora-infrastructure/issue/8390
16:11:39 <mboddu> So, for now, we will close the ticket and ask the maintainers to take it to fesco to approve on a process
16:11:42 <nirik> anyhow, IMHO we should close those wontfix and ask the reporters to ask fesco for a process... we have limited time/manpower and it's not something we want to keep looking at each case
16:11:53 <mboddu> ack
16:11:58 <sharkcz> +1
16:11:59 <kalev> the script to check should be easy, something like this to see if the branch is reachable from master:
16:12:04 <nirik> and that process should be easy on releng.
16:12:07 <kalev> if git rev-list origin/master | grep `git rev-parse remotes/origin/momento` ; then echo "OK to remove" ; else echo "NOT ok to remove" ; fi
16:12:17 <kalev> where "momento" is the branch to check
16:12:47 <mboddu> #info We will close the tickets and ask people to take it to fesco for a process
16:12:49 <nirik> is reachable by master the right critera? I guess it's one...
16:13:01 <kalev> I think so
16:13:24 <kalev> I guess it's a start at least :) any other fXX branches should be valid to check as well, I guess
16:13:44 <nirik> or epel* except when they want to remove those.
16:14:25 * kalev nods.
16:14:25 <mboddu> kalev: But it doesn't tell if there is a build from momento branch, right?
16:14:34 <mboddu> build submitted from*
16:15:00 <kalev> no; but it makes sure that all the contents from the momento branch is merged into a branch we never remove (master)
16:15:08 * dustymabe lurks
16:15:27 <mboddu> Right
16:15:32 <kalev> so even if there is a build from that branch, then it's ok to remove it as all of the branch contents is merged to master
16:16:48 * kalev is just trying to help out, doesn't have anything to do with python-notebook.
16:17:18 <mboddu> Haha :)
16:17:25 <nirik> sure, that might work for a policy... not sure. I'd want to get it approved by fesco before we go deleting anything tho
16:17:45 <kalev> good plan I think
16:18:57 * Son_Goku waves
16:19:59 <Son_Goku> ugh
16:20:09 <Son_Goku> I just saw the status update for my ticket
16:20:56 <nirik> so if we get a changed policy we can act on it...
16:21:58 <Son_Goku> man, this is getting brutally annoying :(
16:22:11 <Son_Goku> it's not your fault, I know...
16:22:23 * Son_Goku throws it on his internal bucket o' tasks
16:22:42 <nirik> shall we move on for now?
16:22:45 <nirik> sorry Son_Goku
16:23:00 <Son_Goku> eh, it's fine
16:23:03 <mboddu> Sorry Son_Goku
16:23:10 <mboddu> #topic #9169 Block unretirement commits
16:23:17 <mboddu> #link https://pagure.io/releng/issue/9169
16:23:34 <mboddu> So, we decided to use the today-1 to fix this issue
16:23:41 <mboddu> nirik: Is this part of pdc_updater?
16:24:13 <nirik> it's that fedmsg listening thing... not sure what it's called.
16:25:46 * Son_Goku waves away
16:26:25 * mboddu checks
16:26:36 <nirik> yeah, pdcupdater.py in fedmsg hub on pdc-backend01
16:27:57 <mboddu> nirik: Yup, we need to update https://github.com/fedora-infra/pdc-updater/blob/9287defd89085105c7f125ee03fbdd6ee4bb33a4/pdcupdater/handlers/retirement.py#L192
16:28:10 <mboddu> nirik: Oh, you wanna hotfix it?
16:28:22 <mboddu> I am not sure how its deployed? Did we package it?
16:28:31 <nirik> good question, let me see
16:28:45 <nirik> no, it's packaged. we should update the package
16:28:45 <mboddu> Yup, we did - https://koji.fedoraproject.org/koji/packageinfo?packageID=23122
16:29:04 <mboddu> But its failing to build :(
16:29:17 * mboddu bangs his head to the desk
16:29:25 <mboddu> And for reason it feels better
16:30:40 <nirik> ha, well, we can continue on that out of meeting?
16:31:05 <mboddu> nirik: Which one banging my head or the issue? ;)
16:31:12 <mboddu> Anyway, moving on
16:31:27 <nirik> both! :)
16:31:46 <mboddu> Ah, we dont have Son_Goku anymore to discuss https://pagure.io/releng/issue/9246
16:31:49 <mboddu> Well, lets see
16:31:57 <mboddu> #topic #9246 Please block `fedora-cisco-openh264` from pungi composes
16:32:03 <mboddu> #link https://pagure.io/releng/issue/9246
16:32:23 <mboddu> Any idea on what he means by blocking in pungi composes?
16:32:24 <nirik> I don;t think it's enabled anywhere... is it?
16:33:00 <nirik> we don't compose it with the normal composes. it's seperate.
16:33:06 <nirik> so it should never get into them...
16:33:40 <mboddu> Thats what I thought, I checked pungi configs, kickstarts (now checking comps) just to make sure
16:33:46 <nirik> it's never in a tag that pungi-fedora refers to.
16:33:58 <nirik> so, I would say to this ticket: done! :)
16:34:13 <mboddu> Yes, just have to merge the fedora-repos PR's
16:34:24 <kalev> oh, I know what this is about
16:34:43 * mboddu is waiting eagerly :)
16:35:20 <kalev> he just wanted to make sure that if the openh264 repos are enabled by default, they won't get used by any of the media creation tools (pungi)
16:35:32 <kalev> if pungi uses a custom set of repos then this of course isn't an issue
16:35:47 <nirik> yeah, pungi makes all the repos it uses...
16:35:51 <kalev> great :)
16:36:11 <nirik> so I think it will be ok... of course I could be wrong, but hopefully not. ;)
16:36:22 * kalev nods.
16:36:34 <mboddu> Nope, its all good, we should just need to merge and build the fedora-repos PR's thats all
16:36:49 <kalev> relatedly, I build f33 openh264 package that needs sending to Cisco
16:37:08 <mboddu> kalev: And you did create one today, I will get to it in sometime
16:37:15 <mboddu> Maybe not today
16:37:25 <kalev> yeah, no rush from my side :)
16:37:57 <mboddu> #info Nothing is needed on this ticket other than reviewing the PR's on fedora-repos package and building it.
16:38:11 <mboddu> #topic #9229 Improving the state of rust packaging experience
16:38:17 <mboddu> #link https://pagure.io/releng/issue/9229
16:38:24 <mboddu> dustymabe: Are you around?
16:39:24 * nirik thinks this is a discussion ticket.
16:39:32 * dustymabe waves at mboddu
16:39:41 <dustymabe> i asked mboddu to bring this up
16:40:08 <dustymabe> basically I mostly want to bring attention to the ticket and also maybe talk about how we can organize and help make some progress on it
16:40:27 <dustymabe> would it be worth trying to set up some meeting with the rust sig + releng + koji devs to discuss ?
16:40:50 <nirik> well, from the releng side... I think we are happy to implement things that will make things easier for people...
16:41:30 <mboddu> dustymabe: +1 for the meeting
16:41:33 <nirik> perhaps a start would be a thread on devel or releng ? it looks like some koji stuff is pending...
16:41:52 <dustymabe> nirik: thread on releng? isn't that the same thing as the ticket pretty much ?
16:42:01 <dustymabe> i could send an email to point ot the ticket, I guess
16:42:02 <mboddu> nirik: Thread on devel would be nice
16:42:10 <mboddu> dustymabe: ^
16:42:49 <nirik> well, personally I find tickets are not great for 'what to do'... they are much better suited for 'we need to do X'
16:43:19 <nirik> mailing lists are better for discussion... but of course high bw realtime is better yet, but it's hard to get everyone who might care involved at the same time
16:43:49 <dustymabe> right, I find mailing lists the discussion gets lost and it's like it never happened unless it gets captured in a ticket :)
16:44:04 <dustymabe> i honestly think a lot of the "what needs to be done" is mostly there already
16:44:45 <dustymabe> it's mostly getting people together to discuss an execution plan
16:45:07 * nirik isn't sure what releng needs to do here... there's 2 koji pr's and some fedpkg work?
16:45:08 <dustymabe> so I can try to send an email to the group and organize a bluejeans meeting or something
16:46:48 <mboddu> releng partly owns fedpkg, but if someone is going to pick the work that would be great
16:46:56 <mboddu> But I am interested in knowing the process
16:47:02 * mboddu is trying to learn some rust
16:47:03 <dustymabe> nirik: you're right, but you guys also have some say about side tags, etc
16:47:13 <dustymabe> and also your input on how dist-git is used and koji is being used in important
16:47:40 <nirik> sure
16:48:19 <dustymabe> #action dustymabe will try to set up some time to meeting with rust-sig, fedpkg devs, koji devs, and releng
16:48:30 <nirik> might need to open a ticket for the fedpkg script thing?
16:48:36 <nirik> to get it on their radar
16:50:27 <dustymabe> nirik: +1 assuming the meeting goes well
16:50:33 <nirik> not sure if the koji request is on their radar either actually
16:50:47 <dustymabe> that's exactly what the meeting is for :)
16:51:06 <dustymabe> getting everyone some context and on the same page
16:51:09 <nirik> ah, it is.
16:51:20 <nirik> https://pagure.io/koji/issue/1998
16:52:08 <mboddu> All the tickets are filed (except for fedpkg), so lets get everyone together and figure it out
16:52:13 <nirik> FYI, the sidetag-plugin was merged into koji... but without the blocking/tagging stuff
16:53:17 <nirik> anyhow, sure, we can meet sometime...
16:53:38 <mboddu> Okay, moving on
16:53:53 <mboddu> Oh btw, thanks dustymabe
16:53:58 <mboddu> #topic #9222 Accidental upload of "sources" file to look-aside cache
16:54:04 <mboddu> #link https://pagure.io/releng/issue/9222
16:54:25 <mboddu> nirik: Again historical reasons/knowledge, can we just do this?
16:54:31 * mboddu never done it before
16:54:50 <nirik> sure, we can just nuke it...
16:54:57 <nirik> you want to, or want me to?
16:55:01 <mboddu> Okay, I can do it
16:55:14 <mboddu> I can do it, I just never done it, thats all :)
16:55:22 <mboddu> #topic Open Floor
16:55:34 <mboddu> So, I got couple of things
16:55:39 <nirik> sure, fire away
16:56:02 <mboddu> #info Mass branching is done and branched compose failed due to a dep issue which I untagged and reran the compose
16:56:43 <mboddu> #info Mass branching of modules is still on going
16:56:50 <mboddu> Thats it
16:56:53 <mboddu> nirik: Do you have any?
16:57:18 <dustymabe> mboddu: nirik: FYI i'm starting to work again on the automated ostree import work. mostly just wiring things up now and getting a cert so the importer can publish messages https://pagure.io/fedora-infrastructure/issue/8641
16:57:40 <nirik> dustymabe: cool.
16:57:49 <dustymabe> so I might bug you with a few questions this week as we get closer to migrating the ostree repos from /mnt/koji into the netapp volumes we created
16:58:02 <nirik> mboddu: well, we didn't get to much backlog today... I guess next week we are gonna start some more sessions on that?
16:58:15 <nirik> dustymabe: ok.
16:59:13 <nirik> otherwise I got nothing. ;)
16:59:18 <mboddu> nirik: Okay
16:59:28 <mboddu> dustymabe: ack
16:59:38 <mboddu> Thats all for today
16:59:44 <mboddu> Thank you all for joining
16:59:56 <mboddu> #endmeeting