16:00:45 <mboddu> #startmeeting RELENG (2019-08-29)
16:00:45 <zodbot> Meeting started Wed Aug 28 16:00:45 2019 UTC.
16:00:45 <zodbot> This meeting is logged and archived in a public location.
16:00:45 <zodbot> The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:45 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:45 <zodbot> The meeting name has been set to 'releng_(2019-08-29)'
16:00:46 <mboddu> #meetingname releng
16:00:46 <zodbot> The meeting name has been set to 'releng'
16:00:46 <mboddu> #chair nirik tyll sharkcz masta pbrobinson pingou puiterwijk maxamillion mboddu dustymabe ksinny jednorozec
16:00:46 <zodbot> Current chairs: dustymabe jednorozec ksinny masta maxamillion mboddu nirik pbrobinson pingou puiterwijk sharkcz tyll
16:00:46 <mboddu> #topic init process
16:00:51 <nirik> morning
16:01:01 <mboddu> morning nirik, how goes?
16:01:09 <nirik> mboddu: perhaps we should adjust the chairs. ;)
16:01:18 <nirik> some of those folks aren't around much anymore.
16:01:34 <mboddu> nirik: Yeah, I removed few of them and need to adjust more
16:01:43 * mboddu goes and does that
16:02:55 <mboddu> Done
16:03:09 <mboddu> So, lets get started
16:03:32 <mboddu> I have 3 tickets to discuss today
16:03:51 <mboddu> #topic #8616 Block perl-bootstrap modules from entering composes
16:03:53 <dustymabe> hola
16:04:03 <mboddu> #link https://pagure.io/releng/issue/8616
16:04:11 <mboddu> dustymabe: hola, buenos dias :)
16:04:59 <mboddu> So, I am not sure how we can achieve this, blocking from composes but not from buildroot
16:05:54 <nirik> hum, looking
16:06:18 <nirik> we have a thing like this for packages.
16:06:33 <nirik> but not sure on modules, might need pungi work?
16:07:10 <mboddu> nirik: Yes, not for modules,
16:07:15 <nirik> yeah, I'd say we need a pungi RFE
16:07:50 <mboddu> I am thinking like a variants.xml (where we added modules initially for the composes) but instead block them from entering into composes
16:08:26 <nirik> well, it could just be like filter_packages...
16:08:30 <nirik> filter_modules=
16:08:50 <mboddu> #action mboddu to file a pungi RFE to filter out modules from composes
16:09:35 <mboddu> #topic #8581 installations from Live images use excessive client CPU on /dev/loop1
16:09:43 <mboddu> #link https://pagure.io/releng/issue/8581
16:10:15 <nirik> I'm fine using it once we have a way to in koji. ;)
16:12:35 <mboddu> nirik: Sorry, I didn't look at the linked ticket, so, once koji starts supporting it, we will try it and see
16:13:09 <mboddu> #info waiting on koji support - https://pagure.io/koji/issue/1622
16:13:31 <mboddu> #topic #8513 F32 system-wide change: x86-64 micro-architecture update
16:13:39 <mboddu> #link https://pagure.io/releng/issue/8513
16:14:14 <nirik> this didn't get accepted/happen
16:15:12 <mboddu> nirik: Not accepted yet
16:16:15 <nirik> I don't think it's likely to ever be
16:17:15 <mboddu> Okay, I am not very happy to accept the change, but lets punt it for now
16:17:33 <mboddu> #topic Open Floor
16:17:40 <mboddu> dustymabe: I think you have something?
16:17:46 <dustymabe> mboddu: maybe: https://pagure.io/releng/issue/8421#comment-593032
16:18:56 <nirik> dustymabe: I wonder... is your tagger doing a loop?
16:20:23 * nirik isn't sure there's a seperate dist-repo perm... if so, I'm fine adding it to that group
16:20:54 <mboddu> nirik: Me neither, probably a koji hub policy?
16:21:34 <dustymabe> nirik: did I answer your question here: https://pagure.io/koji/issue/1630#comment-593345 ?
16:21:38 <nirik> would have to dig...
16:21:59 <nirik> ah ha. ok, that makes sense.
16:22:34 <nirik> not sure what the solution is here...
16:23:14 <dustymabe> nirik: could we just serialize distrepo calls for a given tag ?
16:23:24 <dustymabe> in the koji code itself probably
16:23:51 <dustymabe> we can discuss in the ticket.. my question here was mostly can we get permission to regen the repo if it's in a bad state
16:23:59 <dustymabe> it sounds like you guys aren't opposed
16:24:25 <mboddu> I am not opposed to giving permission, but I am not sure if there is a specific permission like that
16:24:51 <mboddu> Also, to fix it, we either need fixes to koji distrepo or tag2distrepo
16:25:06 <nirik> either in koji or tag2distrepo...
16:25:19 <dustymabe> mboddu: correct. but we're separating the discussion here into two topics
16:25:33 <dustymabe> 1. we can get distrepo permissions
16:25:47 <dustymabe> from what I can tell the answer is yes if the permission exists
16:25:54 <dustymabe> 2. can we fix the race condition
16:26:02 <dustymabe> we don't need to answer #2 in this meeting
16:26:11 <mboddu> dustymabe: Yes, you are right
16:26:16 <dustymabe> $ koji dist-repo --non-latest coreos-pool 429476b4 cfc659b9 3c3359c4 12c944d0
16:26:19 <dustymabe> 2019-08-27 16:45:40,207 [ERROR] koji: ActionNotAllowed: dist-repo permission required
16:26:29 <dustymabe> at least koji names the permission as `dist-repo`
16:26:33 * mboddu checks
16:26:36 <dustymabe> so sound like the perm exists
16:26:51 <nirik> yeah.
16:26:56 <nirik> just was digging on it.
16:27:37 <nirik> dustymabe: who should have it? can we do this out of meeting?
16:27:50 <dustymabe> nirik: yes we can do it out of this meeting
16:28:07 * nirik can add folks and we can work on 2...
16:28:10 <dustymabe> the answer to your question is `coreos-continuous` group
16:28:26 <dustymabe> +1 thanks for the discussion folks
16:28:44 <nirik> I'm not sure that works at all.
16:28:48 <mboddu> nirik: We dont have it enabled(if there is a permission like that) - https://paste.fedoraproject.org/paste/zTZVvv1HwWJA9vi2xNRluw
16:29:03 <nirik> thats a set of permissions for tagging.
16:29:50 <dustymabe> maybe we can move this discussion to the releng channel after the meeting?
16:29:59 <dustymabe> I just don't want to consume the meeting with it :)
16:30:10 <nirik> sure.
16:31:19 <mboddu> Anything else?
16:33:24 <dustymabe> not from me
16:33:39 <nirik> nothing I can think of off hand.
16:34:18 <mboddu> Okay, thanks for attending everyone
16:34:29 <mboddu> #endmeeting