16:03:34 #startmeeting RELENG (2019-05-16) 16:03:34 Meeting started Wed May 15 16:03:34 2019 UTC. 16:03:34 This meeting is logged and archived in a public location. 16:03:34 The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:03:34 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:03:34 The meeting name has been set to 'releng_(2019-05-16)' 16:03:34 #meetingname releng 16:03:34 The meeting name has been set to 'releng' 16:03:34 #chair nirik tyll sharkcz masta pbrobinson pingou puiterwijk maxamillion mboddu dustymabe ksinny jednorozec 16:03:34 Current chairs: dustymabe jednorozec ksinny masta maxamillion mboddu nirik pbrobinson pingou puiterwijk sharkcz tyll 16:03:34 #topic init process 16:04:46 morning 16:04:58 nirik: Morning, how goes? 16:06:10 it goes. 16:06:40 .hello2 16:06:43 bcotton: bcotton 'Ben Cotton' 16:07:14 Okay, lets get started 16:07:30 I want to bring up couple of quick things 16:07:37 #topic compose-tracker 16:09:18 nirik: In our taiga grooming today, we talked about this and dustymabe, jednorozec and I are interested in going over the deployment process, so when do you think you have cycles to do this and can we have a screen share session for our learning experience? https://pagure.io/fedora-infrastructure/issue/7770 16:09:18 which is this? the failed composes thing? 16:09:30 nirik: ^^ yes 16:09:38 cool. sure. 16:09:46 perhaps friday after fesco meeting? 16:09:51 i can help with compose-tracker if needed; i need some more practice with openshift 16:10:37 mizdebsk: Awesome 16:10:58 hopefully it shouldn't be too complex an app :) 16:11:13 nirik: So, can I create a meeting for this along with all the folks? 16:11:23 After your FESCo meeting 16:11:38 sure I guess. 16:11:53 I just got in so I haven't seen todays fires, but hopefully friday will be quiet. 16:12:23 nirik: Okay, we can reschedule it if anything comes up 16:12:31 sure, sounds good. 16:13:07 nirik: Also, we are working on porting it to fedora-messaging, but for now it works on fedmsg, so we planned to deploy it with fedmsg for now and change it to fedora-messaging later on 16:13:34 I have a PR waiting to get merged, and couple of questions to get answered 16:13:39 ok. 16:14:30 #info We are planning to deploy compose-tracker on Friday and nirik will have a screen share session to help us understand how the deployment works in fedora infra's openshift 16:14:58 #topic #7718 The FTBFS cleanup policy is not happening 16:15:03 #link https://pagure.io/releng/issue/7718 16:15:14 nirik: I think you completed this request, can I close this? 16:16:23 I still need to put them in as cron jobs... ;( 16:16:44 so, leave it for now, I will try and get that this week. 16:17:08 nirik: Okay, let me know if you need any help 16:17:38 #info nirik is planning to close this sometime this week, the only thing remaining is put it in a cron job 16:17:41 happy to have help if you want to work on it. ;) Just adding those to ansible and cron jobs 16:18:48 nirik: Okay, that should be easy 16:19:05 #topic #8176 Asking for agreement for a new micro-service Message-Tagging-Service 16:19:49 nirik: I am thinking of closing this, have you heard anything lately about this? 16:20:00 I haven't... there's an infra ticket too. 16:20:25 * mboddu bbiab, hiccups, need some water 16:20:33 https://pagure.io/fedora-infrastructure/issue/7563 16:21:10 we should probibly ping there and ask what the status is 16:23:02 * mboddu is back 16:23:17 nirik: You ping on infra ticket and I will ping in releng ticket ;) 16:23:38 ha. sure, if you like 16:23:55 nirik: Also, I am not sure how much of a help MTS provides us 16:24:11 But, better to check 16:24:46 hard to say... it could... 16:24:52 Yup 16:25:37 #info There is no status update on this recently, so we decided to ping in the ticket to get an update. 16:25:52 #topic #7288 Request permissions to untag packages 16:25:58 #link https://pagure.io/releng/issue/7288 16:26:14 mizdebsk: Do you think is there an easy way to achieve this? 16:26:32 * mizdebsk reads the ticket 16:26:55 yes, this is configurable with hub policy 16:26:55 mizdebsk is my go-to guy for koji-hub policy changes :) 16:26:58 * nirik also reads since I don't recall what this was about 16:27:08 so should be easily doable 16:28:41 I'm fine with a new group or whatever, but we need to avoid people in general untagging things that have already gone out in a rawhide, etc. 16:28:42 mizdebsk: Just so you know, he/she should be only allowed to untag the builds in rawhide if they are not shipped yet and only the builds that are built by them 16:28:54 nirik: Exactly 16:29:02 actually there's a work around here I think... 16:29:55 say foo-1.2-1.fc31 is broken, koji tag-pkg f31-pending foo-1.1-1.fc31, it would be signed again (if not already) and then tagged into f31 and be used since it's last tagged right? 16:30:05 the "if not shipped" part will need to be evaluated by the person untagging builds 16:30:32 but note that ordinary users can already re-tag older builds without checking anything, which will have similar effects to untagging 16:31:19 I guess both have certain unnecessary effects 16:34:13 I think we need some sort of automagic - if a maintainer feels the build is broken, he will create a ticket or something which should be automatically processed and untag the build or retag the last build before it 16:35:26 There's one case where the retagging older build won't work... if it's the initial build of something... but I think thats pretty rare. 16:35:53 gating would solve much of the problem with untagging 16:36:04 yeah 16:36:09 mizdebsk: True 16:36:19 it would prevent some of broken builds in the first place; and bodhi enablement allows to submit buildroot overrides for rawhide 16:36:23 I'm afraid if we document/tell people about retagging old builds it would get misused. 16:36:44 so, perhaps we tell tibbs about this for now and wait for gating otherwise. 16:40:36 nirik: I think we should not tell them and say gating would solve most of these issues 16:41:39 sure in general. I'm fine mentioning to a few provenpackagers that they can retag. 16:42:36 nirik: Yeah, just being careful that it wont be spread around and get misused 16:42:50 sure 16:44:08 #info There isn't a good way to solve this one, but rawhide gating should solve help the maintainers a lot in these cases. 16:45:39 #topic Open Floor 16:45:55 Anyone has anything to share? 16:46:11 hi, have one little thing 16:46:32 I've updated puiterwijk's check_metalink, so it works again :-) 16:46:48 adds checks for modular repos and more 16:47:09 oh nice. 16:48:15 it should avoid the broken repos (wrong url in mirrormanager, etc) we see sometimes 16:48:48 sharkcz++ 16:48:48 mboddu: Karma for sharkcz changed to 1 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 16:48:54 Thanks sharkcz 16:49:02 mboddu: np :-) 16:49:12 sharkcz: merged that. Thanks! 16:49:15 sharkcz++ 16:49:15 puiterwijk: Karma for sharkcz changed to 2 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 16:49:18 puiterwijk: yup, thx 16:49:53 #info sharkcz updated the check_metalink and added checks for modular repos as well 16:50:33 sharkcz: So, how's alt arch doing? 16:50:55 mboddu: no problems, except the occasional issues with s390x 16:51:44 the srpm unpacking and sometimes images are broken, but today it (=cloud) went OK 16:52:14 sharkcz: I have been asking nirik for our own s390x machine, but he is neglecting my requests :P 16:52:49 #info Everything seems to be fine on alternate arches, except for some occasional s390x issues 16:52:50 hopefully some time we will :-) 16:53:02 ha. 16:53:19 I pinged on kvm s390x again the other day, no reply... 16:53:33 I can ping on tthe ticket too 16:53:56 nirik: Do you know the name of the person who is supposed to work on it? 16:54:26 yep. I don't know that I want to name people in public that may not want attention tho 16:54:28 If he is from Westford office, I can try to find him and ask about it 16:55:03 nirik: Sure, you can ping me on internal irc 16:56:29 hes remote I guess. 16:57:18 Then I am not of much help 16:57:52 Anything else before I close this meeting then? 16:58:00 nope 16:58:18 * mboddu waits a min before closing the meeting 16:59:35 Okay, thanks everyone for joining 16:59:39 See you all next week 16:59:41 #endmeeting