17:00:24 #startmeeting RELENG (2018-07-19) 17:00:24 Meeting started Thu Jul 19 17:00:24 2018 UTC. 17:00:24 This meeting is logged and archived in a public location. 17:00:24 The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:24 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:24 The meeting name has been set to 'releng_(2018-07-19)' 17:00:24 #meetingname releng 17:00:24 The meeting name has been set to 'releng' 17:00:24 #chair nirik tyll sharkcz masta pbrobinson pingou puiterwijk maxamillion mboddu Kellin dustymabe 17:00:24 Current chairs: Kellin dustymabe masta maxamillion mboddu nirik pbrobinson pingou puiterwijk sharkcz tyll 17:00:25 #topic init process 17:00:39 morning 17:00:47 nirik: Morning, how goes? 17:01:01 just cleaning up some fallout from mass rebuild. ;) 17:01:17 nirik: Like? 17:01:44 Your personal maintained packages or something related to releng/infra? 17:02:55 yeah, things I noticed on my laptop so far: Thunar was updated and xfdesktop needed rebuilding again (well, it needs an upgrade to use the new Thunar, but a rebuild fixes the broken dep) 17:03:16 and gnome-shell failed to build due to sassc being broken. fixing sassc now... looks like it had a missing source. 17:03:31 and shotwell needed rebuilding for new LibRaw 17:03:35 is lurking 17:04:02 nirik: Okay 17:04:10 Okay, lets get started 17:04:18 I have only couple of topics to discuss 17:04:45 #topic #7645 module metadata in releases/28/Modular/$basearch/os/repodata/ has no context or arch 17:04:51 #link https://pagure.io/releng/issue/7645 17:05:32 yeah, I saw that 17:05:44 * nirik has questions for them. was going to add to the ticket, but haven't yet 17:06:07 nirik: Okay, but at the same time, we cannot run another F28 compose right? 17:06:14 I am not sure we want to rebuild something that was made at GA tho 17:06:18 right. we cannot 17:06:53 My big question was: does the old dnf understand the new metadata... 17:07:25 nirik: You mean, F29 metadata? 17:08:06 sure, if we don't change the f28 as they want... I mean what happens if someone installs a fresh f28 and then uses that dnf to try and work with modules... 17:08:39 but I guess we should just say we cant do that and leave it at that. 17:09:00 (although we could do some trickery like redirecting it to another place or something perhaps) 17:09:01 I guess so 17:09:27 "Trickery" sounds scary and I am not sure what that means 17:09:38 You mean in MM? 17:09:52 or baseurl redirect itself? 17:09:57 yeah, or on the master mirrors... 17:10:21 but we would also still need a way to make the new repodata without doing a compose, which I am not at all sure is possible 17:10:54 True, we might have to check with pungi people and probably make a one time compose and run it.... 17:11:03 Well, that seems like a lot of work 17:11:33 And seems like they are okay with "* don't use the new depsolving code on F28 and support it from F29" 17:12:04 yeah, that seems vastly easier 17:13:00 nirik: So, can you update the ticket and close it? 17:13:39 Or do you want me to? 17:13:39 sure. 17:13:43 either way 17:13:53 Since you said you were already planning on updating the ticket, so... 17:14:50 nirik: Better you go ahead, probably you will word it better and I might learn something out of it 17:14:52 :) 17:15:22 ha. sure, I can try 17:15:31 nirik++ 17:15:59 * cverna is around 17:17:22 #info nirik will update the ticket but we cannot rerun the compose and rebuild something that is already GA. We thought of other options like redirecting it to some other place in mirror manager or master mirror, but still we need a way to generate the metadata which we are not sure of. 17:17:27 hi cverna 17:17:35 mboddu: o/ 17:18:13 #info Since in the ticket they mentioned they can skip it for F28 and include it in F29, we are thinking that might be the best way to go. 17:18:24 Next 17:18:38 #topic FTBFS Updates 17:20:23 nirik: Till added https://pagure.io/releng/c/ccceff2232d8802dc680ecfab558c9a5649ce3f6?branch=master, but now its only attaching 1024 bytes of logs 17:20:33 Do you know whats the limit? 17:21:28 And what should we do with the already filed bugs? 17:21:40 hum... yeah, 2^10 is 1024. ;) 17:22:03 >>> print 2 ** 10 17:22:03 1024 17:22:35 And, currently the script failed while processing https://koji.fedoraproject.org/koji/taskinfo?taskID=28171629 with missing logs, so I fixed the script to exception handle them, so waiting on how to proceed further 17:23:00 I shouldn't say missing logs, for some reason the build itself didn't put any logs 17:23:18 Probably a bug in koji, I am not sure 17:23:24 well, I would up the limit. whats an example of one with 1024? 17:24:11 nirik: https://bugzilla.redhat.com/show_bug.cgi?id=1603242 17:24:15 could just resubmit that one? 17:25:11 I guess I would say up the limit and continue from where it left off. 17:25:34 people can always look at koji or do another build. 17:25:36 Okay 17:25:52 And, do you want me to resubmit that task? 17:25:57 with missing logs 17:26:03 and 1024 there was enough to tell it was the BR: gcc issue. ;) 17:26:21 Yeah, I was worried about that too 17:26:25 yeah, it should work now, it looks like it hit some time when the s390x cache was not working 17:26:26 But it was part of it 17:26:31 Okay 17:27:33 And, how much should I increase the limit to? 17:28:53 dunno.. how about 16k... 2 ** 14 17:28:55 ? 17:29:01 or is that still too small... 17:29:17 Probably 4096 - 2**12? 17:29:37 Okay, lets go with 2 ** 15 17:29:45 thats 32k 17:29:50 ok 17:31:55 Okay 17:32:04 Thats all I have, going to open floor 17:32:15 #topic Open Floor 17:32:20 Anybody got anything? 17:32:53 I will be on PTO for the first week of Aug 17:33:06 before FLOCK 17:33:22 cverna: ^ lets go together to FLOCK from Wroclaw 17:33:43 cverna: I still haven't got that train ticket yet :( 17:33:47 * nirik will be at flock 17:33:56 mboddu: yeah I got my ticket to Wroclaw I ll arrive the the 7th in the morning 17:34:23 pdx->ams->dresden on the way out, then train from dresden -> ams then plane ams->pdx. 17:34:30 cverna: Okay, lets catch the train together, I will ping you my contact number later 17:34:48 nirik: sounds like a long journey 17:34:54 mboddu: did the f27 container release completed ? 17:35:01 yep... but thats the way it is. :) 17:35:15 cverna: Yup, and I asked you about running F28 or not, but I guess it was too late in your time 17:35:39 mboddu: oh great yes I missed that, we can run f28 :) thanks 17:35:46 nirik: How long is the train from Dresden to Amsterdam? 8 hrs? 17:35:56 cverna: Cool, I will run it now 17:35:59 yeah, something like that... 17:36:14 you all enjoy FLOCK, wish I could join you 17:36:30 we will miss you Kellin. :( 17:36:33 nirik: Wow, at least its train, so you can relax 17:36:38 lots of folks not making it this time.... 17:36:41 I'll be at the US one next year 17:36:43 Kellin: You will be missed :( 17:36:51 Kellin: Yay 17:37:23 mboddu: we have started to discuss container updates in bodhi with Randy 17:37:44 mboddu: but we want you to be involved so you can give some feedback 17:37:56 I opened https://pagure.io/releng/issue/7641 17:38:14 cverna: Sure, I think I already talked to Randy about it, but its been a while 17:38:15 But I think we need to clarify which koji tags we will want 17:38:25 Cool 17:39:04 We can talk 17:40:13 both randy and I are on PTO tomorrow, so we could schedule something next week :) 17:40:23 mboddu: I ll send an email about it 17:40:35 cverna: Sure 17:40:37 cverna: +1 17:40:51 it would be nice to have container updates for f29 :) 17:42:19 cverna: Yup and I dont have to run any releases, thats another +1 for me :D 17:42:37 mboddu++ :) 17:43:12 Okay, anything else guys? 17:43:55 that was all from me 17:44:03 oh, one short note: 17:44:39 I was going to post on devel about the rawhide vs 29 naming stuff... but then I re-read all the tickets and wanted to ask some more questions, so continuing in that ticket. 17:45:16 nirik: Okay 17:45:25 hopefully we can reach some consensus and then take that to devel 17:46:24 #info nirik was going through rawhide vs 29 naming and wanted to ask some more questions in that ticket. 17:46:30 I have one more note 17:47:01 nirik: We lost kirin on #fedora-releng, instead we got nirik-, not sure who is better? :P 17:47:29 ha. will check, probibly just needs restarting (kirin is my phone BTW) 17:48:00 nirik: Okay, I know kirin is also belongs to you, not sure that its on your mobile 17:48:18 Anyway, thats all I got 17:48:25 Thanks guys for joining 17:48:37 * mboddu will get back to FTBFS 17:48:45 #endmeeting