15:02:08 #startmeeting RELENG (2020-10-13) 15:02:08 Meeting started Tue Oct 13 15:02:08 2020 UTC. 15:02:08 This meeting is logged and archived in a public location. 15:02:08 The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:02:08 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:02:08 The meeting name has been set to 'releng_(2020-10-13)' 15:02:08 #meetingname releng 15:02:08 The meeting name has been set to 'releng' 15:02:08 #chair nirik sharkcz pbrobinson pingou mboddu dustymabe ksinny jednorozec 15:02:08 Current chairs: dustymabe jednorozec ksinny mboddu nirik pbrobinson pingou sharkcz 15:02:08 #topic init process 15:02:14 Lets see who is around 15:02:23 hello 15:02:23 hi, I am here 15:02:25 morning 15:02:30 Hi! 15:03:27 Hello guys 15:04:06 I am in multiple meetings, so this meeting might go slow 15:04:10 To start with 15:04:56 gchang is an intern with Red Hat and she is going to help us out with some of the releng/infra work 15:05:13 #topic New Introductions 15:05:22 welcome gchang! 15:05:25 gchang: Do you wanna add something here? 15:05:26 Yeah! I'll be here until the end of April 15:05:42 welcome gchang! 15:05:46 Hm, I'm just excited to be here haha 15:06:26 Thanks gchang :) 15:07:05 #topic Automate Unretirement 15:07:22 This is something we wanted to do for a long time and we would like gchang to pick it up for us 15:07:34 Last week, jednorozec and I discussed about it 15:08:08 For now, gchang will help us to automate the unretirement process detailed in https://docs.pagure.org/releng/sop_unretire.html 15:08:27 Sounds like a plan 15:08:46 But once its done, we will look at incorporating that work into fedpkg and fedscm-admin 15:09:17 The final plan is that, people can request unretirement through fedpkg and fedscm-admin will process the unretirement 15:09:43 what do you guys think of the idea adding it to fedpkg and fedscm-admin? 15:11:07 sure I guess. 15:11:23 +1 15:11:33 good plan 15:14:15 Okay, moving on 15:14:50 #topic #9804 Orphan kyotocabinet package only on Rawhide 15:14:55 #link https://pagure.io/releng/issue/9804 15:15:09 This is a weird request and one of them makes me miss pkgdb 15:15:26 ha. yeah. 15:15:34 Anyway, other than https://pagure.io/releng/issue/9804#comment-695475, the only other way that I can think of is 15:16:21 well, they can change the bugzilla assignee for fedora bugs to orphan. 15:16:32 Orphan the package and make the reporter "Collaborator" with access to epel branches, but if orphaned, it might get retired eventually 15:16:36 but they still would be maintaing it 15:16:44 yeah 15:17:32 From what I understand, its more like he dont want to maintain them rather than just bz 15:18:24 so, give them the options we have... can't do much more than that 15:22:28 * nirik has an item for later after other topics. 15:22:55 Okay, I will update the ticket 15:24:21 #topic #9801 Allow build_from_srpm for fNN-coreos-continuous Koji target 15:24:26 #link https://pagure.io/releng/issue/9801 15:24:41 I think we should enable this just like the infra tags 15:24:45 nirik: Wdyt? 15:26:04 as long as we are 100% sure that it never will get pushed into the real releases... 15:28:36 nirik: Is there anyway that we can block it? 15:28:38 nirik: Is there anyway that we can block it? 15:28:51 can a policy be set up to ensure that? 15:28:58 Sorry for the double ping, my irc client is acting up 15:29:27 I don't think so... we could make policy refuse to tag things from one to the other, but I don't think it knows 'built from src rpm' 15:31:17 at least I don't think so. I can look more if you want... 15:32:25 We can ask koji folks for help on creating a policy like that 15:32:56 let me look at policy again and I can update the ticket. 15:34:35 Okay, thanks nirik 15:34:44 I will also take a look if I can find anything 15:40:22 #topic #9746 Please provide the src repo of eln-build 15:40:45 #link https://pagure.io/releng/issue/9746 15:41:04 nirik: We have to make a decision on storage 15:41:12 storage isn't the problem really. 15:41:31 it's more cpu every single time a newrepo is done... 15:42:42 anyhow, I guess we can enable it... 15:45:03 Okay, I will enable it 15:46:12 it should be a much smaller set than rawhide say 15:46:51 Definitely 15:46:56 #topic Open Floor 15:47:00 nirik: You are up 15:47:05 I have 2 things actually... 15:47:38 first: we have a lot of side tags... shall we set some policy around them? like: 15:48:09 sidetags older than 60 days will be removed. sidetags with 0 builds in them will be deleted after two weeks. 15:48:29 +1 15:48:37 +1 15:48:51 I am not sure about 60days older one, since KDE and gnome side tags definitely take more time 15:49:02 We would need a freeze break or waiting until after to implement it of course. 15:49:09 I would say if no new build in for 60days delete. 15:49:13 no, named sidetags wouldn't be affected. 15:49:24 here's the current list: 15:49:38 We asked them to start using bodhi side tags as it doesn't require releng help 15:50:12 https://paste.centos.org/view/3938b34c 15:50:37 those are the ones that would be deleted currently using the above policy 15:51:55 I would say last build date + N days as condition for deletion. 15:51:56 60 days is likely too long... perhaps 30 would be better, or less. 15:52:20 well, I am using the already implemented and existing script... :) 15:52:26 start with 60 and see 15:52:28 as the releng provides side tags won't be affected, then I would say go with the plan 15:52:38 makes a lot of sense 15:53:12 nirik: I guess last build + 60 days is fine 15:53:15 I'll file a ticket on it and we can track announcing/implementing. 15:53:21 Ack 15:53:35 well, I am not sure the script does that... 15:54:11 just the date of creation of the side tag is not really useful. 15:54:16 empty ones are likely made and never used. 15:54:31 because koji should be deleting them when there are no more builds in them 15:54:50 jednorozec: yeah, so perhaps we should get koji folks to adjust the script? 15:55:01 Yeah, lets do that 15:55:32 I might be misreading it too... 15:55:56 Lets create a ticket and track it 15:56:01 +1 15:56:11 nirik: You got a 2nd thing to discuss? 15:56:15 oh yeah. 15:56:39 So, we talked a while back about renaming the master branch in places... we created a doc, and then we got busy and didn't do anything more. ;( 15:56:49 Yup 15:57:00 I was wondering if folks are ok with me making a f34 change on it... 15:57:11 and updating the docs more 15:57:33 Sure, if you want, I can help you with the change and you can add me to that change as well 15:57:35 I can update things and ask for review? but I'd like to start moving it forward again 15:57:43 +1 15:57:44 mboddu: that would be great! 15:58:31 ok, thats it... ;) 15:58:40 I have one as well, do we have a plan to replace pdc? 15:58:54 * jednorozec is digging into pdc and import armhfp 15:59:07 jednorozec: We do, but I dont think we will do it anytime soon 15:59:25 mboddu, ok I will try to update PDC 16:00:22 but it is complex 16:00:34 yeah, we have had several plans... 16:00:41 but none of them went anywhere. :( 16:00:53 The last one I think was to try and use... 16:00:59 Can I drive this than? 16:01:13 Its a complex system which we got tied into and we dont have a replacement plan 16:01:22 jednorozec: Sure 16:01:31 mboddu, do we have requirements? 16:01:34 ack. I can't recall the thing... will have to look it up 16:02:12 jednorozec: Not exactly :( 16:02:39 ok, I will create one 16:02:47 anyhow, sure! if you want to look at this, please do 16:03:05 right now pdc is... old and not good. 16:03:15 Yup 16:03:23 yeah, from what I have seen. We dont need this kind of complex system 16:03:25 Anyway, thats all I got 16:03:28 the first try at replacing it was fpdc 16:04:08 * jednorozec magically resurrect fpdc 16:04:17 Haha 16:04:29 Okay, its over time 16:04:30 but I think fpdc was just a reimplementation in flask? not sure... 16:04:55 thanks! 16:05:12 We thought about postgrest or something like that which is a rest implementation over postgres db 16:06:05 Then there was kinto 16:06:09 Anyway 16:06:12 We are over time 16:06:34 Thanks everyone for joining 16:06:37 #endmeeting