14:31:12 #startmeeting RELENG (2017-04-10) 14:31:12 Meeting started Mon Apr 10 14:31:12 2017 UTC. The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:31:12 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:31:12 The meeting name has been set to 'releng_(2017-04-10)' 14:31:12 #meetingname releng 14:31:12 The meeting name has been set to 'releng' 14:31:12 #chair dgilmore nirik tyll sharkcz masta pbrobinson pingou puiterwijk maxamillion mboddu Kellin 14:31:12 Current chairs: Kellin dgilmore masta maxamillion mboddu nirik pbrobinson pingou puiterwijk sharkcz tyll 14:31:12 #topic init process 14:31:23 morning 14:31:25 hi 14:31:41 * masta is here 14:31:49 .hello kellin 14:31:50 Kellin: kellin 'None' 14:36:07 Lets give couple more min for people to join 14:36:56 .hello maxamillion 14:36:58 maxamillion: maxamillion 'Adam Miller' 14:37:50 maxamillion is here, so we can start ;) 14:37:52 morning 14:38:20 #topic #6645 Make sure OSTree repos are sync'ed correctly after the masher completes 14:38:28 #link https://pagure.io/releng/issue/6645 14:38:42 Does anybody got a chance to look at it? 14:39:33 not I, I know that bowlofeggs had a meeting about the OSTree topic recently but I was on PTO and don't know if anything around this was addressed 14:39:56 Last time we decided to replace the cron job with fedmsg driven task. 14:41:39 no one has had time to work on this that I know of. 14:42:15 nirik: yeah, I can understand 14:42:31 maxamillion: okay, since bowlofeggs isn't here, I can talk to him later and update the ticket if anything related to this is discussed in his meeting with OSTree 14:42:33 * nirik suspects we should remove the meeting keyword, since there's nothing more to discuss on it 14:43:25 * mboddu nods 14:44:07 #info mboddu will update the ticket and will remove the meeting keyword, since there is nothing more to discuss about it. 14:44:45 mboddu: +1 14:44:59 #topic #6600 Mock and sigined rawhides 14:45:07 #link https://pagure.io/releng/issue/6600 14:45:43 nirik: any chance you created the RFE about koji support for multiple gpg keys? 14:45:58 yes, and it's a dead end currently. 14:46:25 nirik: dead end? may I know why? 14:46:32 I filed a koji RFE, they couldn't get rpm to work. I filed a rpm RFE and rpm devs say that it's supported by the format, but not any of the tools 14:47:26 so, I don't think we are likely to see this anytime soon 14:47:48 nirik: wow, we have to wait until the tools are supported then 14:48:38 #info nirik filed a koji RFE but it seems like its going to take a while to get it fixed. 14:49:46 There is some talk about detached signatures but that also will likely need a lot of work. 14:49:50 #info Reason: nirik filed a koji RFE, they couldn't get rpm to work, then he filed a rpm RFE and rpm devs said that it's supported by the format, but not any of the tools 14:50:39 nirik: oh well, we have to wait, thats it 14:50:56 yep. 14:51:55 #topic #5886 need method for distributing urgent fixes... urgently 14:52:01 #link https://pagure.io/releng/issue/5886 14:52:22 Its a very old ticket, does anybody have any idea where we stand on it? 14:52:45 The last update on it is even before I joined fedora :D 14:53:07 maxamillion, mboddu: re OSTree - we just talked about *how* to get OSTrees into the mirror network, not about any particular bugs/issues we're facing today 14:53:17 also, 14:53:20 .hello bowlofeggs 14:53:20 bowlofeggs: bowlofeggs 'Randy Barlow' 14:53:23 bowlofeggs: +! 14:53:25 bleh 14:53:29 +1 * 14:54:05 bowlofeggs: would that ticket need to be a function of bodhi? 14:54:09 so on this... I'd say we should just try and make composes faster/more parallel. I think the urgent path stuff is not worth the effort. Others may disagree. 14:54:24 bowlofeggs: thanks for the info, so we have to wait on it as well until we can get it to work 14:54:36 nirik: I dunno, it sounds like it's asking for a 'security updates' channel 14:54:38 maxamillion: i don't think the bodhi user has permission to do that - i think something else has to listen to a fedmsg from bodhi 14:54:38 it's a gigantic pile of work for... a few times a year at most 14:55:11 Kellin: yeah, but everyone misunderstands all the things we do to make a update get to a user. And we have to do most of them anyhow... 14:55:25 so the "urgent/faster" process ends up being... the same as the normal process. 14:55:36 just with extra work and more duplication of effect. 14:55:51 it's pretty fast to push just one update 14:56:06 maybe not minutes.... but not hours 14:56:24 I'm not sure how that could be the case. 14:56:29 masta: can we specifically push just one update? 14:56:33 no 14:56:59 if you push a update to say f24-updates... you have to mash all of f24-updates... even if you are just adding 1 single new update to it. 14:56:59 well... isn't that exactly what we do when we push FE's or blockers ? 14:57:03 maxamillion: unless you are in freeze and there is an FE or blocker fix 14:57:12 Those are not the same path. 14:57:19 those just tag it in koji and say done. 14:57:27 and the nightly compose picks them up and pushes them out. 14:57:58 nirik, oh, well yeah that is true 14:57:59 yeah 14:58:20 I didn't think you could just say "I want this one package, go forth and release it" 14:58:36 yeah, there is no way to do that. 15:00:02 just drop the build in the existing master repo and run createrepo_c on it, but that sounds like a terrible idea. 15:00:27 Then it bypasses everything else we normally do for updates. -100 15:00:56 no bug updating, no drpms, no karma, no checking for signatures, no deps check, no mutillib, etc etc 15:01:26 anyhow, I would say ping on the ticket and see if anyone still really wants this. 15:04:28 nirik: +1 15:05:10 nirik: okay, I can do that 15:05:56 * nirik fears someone will, but I really don't know that it's worth it. ;) 15:07:20 #info mboddu will ask in the ticket if anyone wants it anymore 15:07:52 #topic #6746 Produce a slimmed-down compose whenever certain packages appear in an update 15:07:59 #link https://pagure.io/releng/issue/6746 15:08:27 nirik: I can see the use of it, but are we equipped to do this? 15:09:50 seems like there are lots of things needs to be looked at before we can do this 15:09:57 dunno. I haven't really followed this one... I think it should be doable, but will need some work 15:10:32 so basically turn the compose into a CI system? 15:10:48 new anaconda, whoo... lets run a compose to test it out ? 15:11:13 yeah... if x, y, z update, build new images and say we have... 15:11:14 masta: yes, and probably we can add other pkgs based on QA requirement 15:12:23 nirik: it is doable but where should we need to put the configs and where should the compose should reside and all other questions needs to be answered 15:12:25 probibly should try and gather stakeholders for a meeting on this one and see how we want to approach it. 15:12:35 yep. lots of questions. 15:13:44 nirik: I can create a meeting with releng, infra and QA. Do you think of anyone else that should be added to the meeting? 15:13:48 seems doable 15:14:08 well, definitely threebean also from modularity... 15:15:12 masta: yes, if you look at it all that it needs a pungi config and a fedmsg driven task, but there will be lots of questions around it 15:15:31 nirik: okay, I will add threebean as well, thanks 15:16:41 #info mboddu will create a stake holders(releng, infra and QA) meeting to discuss more about it 15:17:12 #topic Alternative Architectures updates 15:17:20 nirik: any update on s390x? 15:17:55 yeah, waiting on firewall changes now. 15:18:04 the instances are up and I can ping them 15:18:07 but thats it. 15:18:36 will ask networking again today 15:18:37 well that's something :-) 15:19:21 #info nirik is waiting on firewall changes and the instances are up and running 15:19:32 progress at least 15:19:40 thanks nirik, thats a great news :) 15:19:54 #topic Open Floor 15:20:07 anybody got anything to share? 15:20:51 nirik: and we will want a reboot of the hypervisor too to get it more memory (my question in the ticket) 15:21:14 sharkcz: oh yeah, indeed. 15:23:34 Since bowlofeggs is busy, on behalf of him: Bodhi 2.5.0 is deployed in prod 15:23:45 \o/ 15:24:13 #info Bodhi 2.5.0 is deployed in prod 15:24:14 whoot 15:25:08 anything else? 15:25:15 5 15:25:17 4 15:25:18 yay, though this one is a bit bumpy 15:25:23 3 15:25:27 2 15:25:31 1 15:26:02 Okay, lets end it, its been a while we completed the meeting before the time is up 15:26:09 thanks for joining guys :) 15:26:13 #endmeeting