17:06:48 <mboddu> #startmeeting RELENG (2017-09-28)
17:06:48 <zodbot> Meeting started Thu Sep 28 17:06:48 2017 UTC.  The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:06:48 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:06:48 <zodbot> The meeting name has been set to 'releng_(2017-09-28)'
17:06:48 <mboddu> #meetingname releng
17:06:48 <zodbot> The meeting name has been set to 'releng'
17:06:48 <mboddu> #chair dgilmore nirik tyll sharkcz masta pbrobinson pingou puiterwijk maxamillion mboddu Kellin
17:06:48 <mboddu> #topic init process
17:06:48 <zodbot> Current chairs: Kellin dgilmore masta maxamillion mboddu nirik pbrobinson pingou puiterwijk sharkcz tyll
17:07:04 * nirik is sort of here, but in go/no-go meeting also
17:07:14 <mboddu> nirik: me too
17:07:19 <mboddu> :)
17:07:27 * masta is lurking about
17:07:31 <Kellin> .hello kellin
17:07:33 <zodbot> Kellin: kellin 'None' <kellin@retromud.org>
17:09:45 <puiterwijk> .hello2
17:09:46 <zodbot> puiterwijk: puiterwijk 'Patrick "マルタインアンドレアス" Uiterwijk' <puiterwijk@redhat.com>
17:11:03 <mboddu> puiterwijk: I was about to cancel if no one shows up since Kevin and me are doing go/no-go and Jon is just lurking and Rob is busy with his work, but you joined ;)
17:11:21 <puiterwijk> mboddu: oh, don't keep the meeting on for just me
17:11:28 <puiterwijk> I can unjoin if you like?
17:11:56 <mboddu> puiterwijk: no no, I was just thinking about it, since everyone is busy I thought people might not join or just lurk here
17:12:14 * puiterwijk can pretend to be busy if you like
17:12:28 <mboddu> puiterwijk: But I need your or Kevin help on something so I might do a quick one
17:12:32 <puiterwijk> Sure
17:12:34 <puiterwijk> What's that?
17:13:28 <mboddu> Okay, lets do it as per meeting etiquette
17:13:32 <mboddu> #topic #7061 dist-git-pagure does not allow till to orphan pkgs or give a package to someone else
17:13:40 <mboddu> #link https://pagure.io/releng/issue/7061
17:14:14 <nirik> I need input from pingou on this one.
17:14:19 <mboddu> puiterwijk, nirik : How can we sync the cvsadmin group?
17:14:45 <mboddu> nirik: Oh, I thought all that it needs to sync the fas groups on to that machine?
17:14:52 <nirik> no
17:14:59 <puiterwijk> mboddu, nirik: no, it should autosync next time you log in
17:15:15 <nirik> puiterwijk: it doesn't know to autosync that group. it knows nothing about it.
17:15:16 <puiterwijk> If it does not do that, give me a ping and I will fix that
17:15:23 <puiterwijk> Oh, then just create the group
17:15:29 <nirik> where/how?
17:15:33 <puiterwijk> Give me 20 seconds
17:15:42 <nirik> I could not find a place.
17:15:55 <puiterwijk> Okay. I'll figure that out and get it fixed
17:15:57 <nirik> although someone in #devel said that they could create groups
17:15:58 <puiterwijk> That's simple enough
17:16:18 <mboddu> puiterwijk: isn't there a ansible playbook to sync them?
17:16:20 <nirik> [22:30:16]  <QuLogic> so apparently, I can create any group I want on src.fp.o; is that supposed to be allowed? How do I edit members then?
17:16:24 <puiterwijk> mboddu: no, that's something else
17:16:24 * mboddu looked at them but didn't find it
17:16:41 <nirik> I would think it would be restricted...
17:17:09 <puiterwijk> nirik: pretty sure they can't.
17:17:17 <puiterwijk> I am going to guess they were talking about something else
17:17:41 <nirik> yeah, was puzzling.
17:18:50 <mboddu> puiterwijk, nirik : there were at least 10 tickets which needs this issue to be fixed, and on top of that we cant orphan or unorphan a package
17:20:39 <puiterwijk> nirik: can you please log out and log back in to src.fp.o?
17:22:12 <nirik> sure
17:22:58 <nirik> there now
17:23:02 <puiterwijk> https://src.fedoraproject.org/group/cvsadmin
17:23:10 <puiterwijk> Done.
17:23:13 <puiterwijk> `ls
17:23:16 <puiterwijk> Almost
17:24:16 <puiterwijk> mboddu: that ticket should now be fixed
17:24:54 <nirik> so, what did you do? poke the db?
17:25:01 <puiterwijk> Yes. Just an insert on the database
17:25:09 <puiterwijk> Since the group creating code is disabled.
17:25:24 <puiterwijk> (which should be fixed, but meh)
17:25:44 <mboddu> puiterwijk: Oh nice, let me take a look
17:26:25 <mboddu> puiterwijk: yay, awesome
17:26:42 <puiterwijk> mboddu: hint for next time: just ping me :)
17:26:47 <mboddu> puiterwijk++
17:26:56 <puiterwijk> I'm unable to help with things that I don't know about
17:27:04 <puiterwijk> (I have yet to figure out how to read minds)
17:27:23 <mboddu> puiterwijk: sure :)
17:27:36 <mboddu> zodbot: are you sleeping?
17:27:54 <puiterwijk> mboddu: no, since it set the topic. You probably already gave me a cookie this FEdora release cycle
17:28:07 <puiterwijk> (we have a unique constraint on sender, receiver, release cycle.)
17:28:08 <mboddu> #info puiterwijk fixed the issue by inserting on the database
17:28:19 <mboddu> puiterwijk: oh probably
17:28:37 <mboddu> #undo
17:28:37 <zodbot> Removing item from minutes: INFO by mboddu at 17:28:08 : puiterwijk fixed the issue by inserting on the database
17:28:48 <mboddu> #info puiterwijk fixed the issue by inserting on the database since the group creation code is disabled
17:29:59 <puiterwijk> What else is on the roll for today?
17:30:29 <mboddu> puiterwijk: multilib
17:30:33 <mboddu> #topic #7073 python2-openmpi.i686 and python3-openmpi.i686 missing from f26 x86_64 repo
17:30:37 * puiterwijk hides and runs
17:30:40 <mboddu> #link https://pagure.io/releng/issue/7073
17:31:00 <puiterwijk> mboddu: ah. That one is known and will be fixed as soon as that package gets into testing
17:31:01 <mboddu> there are similar issues with multilib
17:31:04 <mboddu> another one - https://pagure.io/releng/issue/7071
17:31:09 <puiterwijk> Yes
17:31:19 <puiterwijk> This is known and gets fixed when the package is pushed to testing/stable
17:31:44 <mboddu> puiterwijk: yes, and there is a long lasting issue regarding them - https://pagure.io/releng/issue/4084
17:31:51 <puiterwijk> Yep
17:31:52 <mboddu> puiterwijk: is there a way that we can fix this?
17:33:48 <puiterwijk> mboddu: see the summaries in the ticket you pointed to. For the taskotron entries: that's known and harmless, as the multilib things will get into the repos when the updates are pushed
17:34:34 <puiterwijk> Yes, ideally this would get fixed. Practically, getting the taskotron check correct there is very hard
17:34:56 <kparal> I wonder if we're doing something wrong there
17:35:10 <puiterwijk> kparal: nah. Not if you want taskotron to run immediately
17:35:37 <puiterwijk> kparal: it would work if you run the depcheck test on the repo where the update is pushed (since then the multilib deps are also in the repo)
17:36:01 <puiterwijk> But that would mean that the taskotron check needs to wait until the next push kicks off *and* finishes after the update is submitted before it can run that check.
17:36:24 <puiterwijk> The alternative is to teach taskotron about the multilib packages, but that will quickly go into a huge mess
17:36:32 <kparal> I heard koji could get a feature to create a repo for each build? that would help us fix this, I think
17:36:46 <mboddu> kparal: oh is it?
17:37:20 <puiterwijk> I guess they're adding that for the CI purposes? If that exists, that would fix the problem if you run the taskotron check on that repo, yes
17:37:31 <kparal> I might be mistaken. but I think I heard someone talking about it.
17:37:49 <kparal> anyway, just wanted to point out an easy solution, if koji ever gets this
17:38:12 <puiterwijk> Right. That would be a great solution, yes.
17:38:24 <mboddu> puiterwijk: so, if I understand correctly, taskotron sends the dep issues emails based on the updates repo and hence the multilibs builds gets those dep issue emails until they end up in updates repo?
17:39:04 <puiterwijk> mboddu: I think it runs it on updates and updates-testing (maybe kparal knows that part?), but it runs the tests *before* the next repo creation, from the downloaded RPMs
17:39:23 <puiterwijk> So yes, it will continue until that update gets pushed into updates-testing. After that, if it uses updates-testing, it would succeed
17:39:24 <kparal> I believe it's fedora+updates repo
17:39:42 <kparal> we use rpmdeplint for that, I can dig out source code if you're *really* interested :)
17:39:45 <puiterwijk> Okay. So then it would indeed continue until that update dragging in the new multilib packages goes to stable
17:39:56 <puiterwijk> kparal: I don't want to know :)
17:40:00 <nirik> kojis repo thing would be pretty expensive for each update... but I guess it could work
17:40:10 <mboddu> puiterwijk: ^ so, it should get into updates repo then
17:40:11 <puiterwijk> nirik: yeah, that was my thought too, indeed
17:40:34 <puiterwijk> mboddu: sorry, I did not understand your statement
17:41:03 <puiterwijk> But I think we can close the releng ticket 7073 with "Sorry, this is a limitation of the current design. Please bug the QA people."
17:41:04 <kparal> I was hoping we're not producing invalid results for every multilib update out there. I thought it happens in just some corner cases. if we do, I guess I should look into it more
17:41:06 <puiterwijk> (sorry kparal :) )
17:41:12 <kparal> but that won't be very soon I'm afraid
17:41:22 <mboddu> puiterwijk: as far as kparal knows, its fedora updates repo not updates-testing repo, so that means the multilib has to land in updates repo not just updates-testing
17:41:43 <puiterwijk> kparal: well, it only happens if they add new multilib requirements in an update that have never been a multilib dep for anything else
17:42:04 <puiterwijk> mboddu: right, yes. And the multilib lands there the same time the update that drags it in lands there
17:42:05 <kparal> feel free to redirect people to https://pagure.io/taskotron/task-rpmdeplint , we can have a look at the errors
17:42:14 <puiterwijk> kparal++ thanks!
17:42:22 <puiterwijk> So let's add that link to the releng ticket and close it?
17:42:40 <mboddu> kparal++
17:42:40 <zodbot> mboddu: Karma for kparal changed to 7 (for the f26 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
17:42:44 <mboddu> kparal: thanks
17:43:03 <mboddu> puiterwijk: I can do it, now that I understood whats going on
17:44:18 <puiterwijk> Good
17:45:10 <mboddu> #info Multilib issue emails are sent on whats available in fedora+updates repo, so the multilib lands in the repos the same time the update that drags it in lands in those repos, until then people will get these emails. And this issue can be better tracked at https://pagure.io/taskotron/task-rpmdeplint
17:45:49 <mboddu> puiterwijk: ready for next topic?
17:45:57 <puiterwijk> mboddu: sure..
17:46:10 <mboddu> #topic Alternative Architectures updates
17:46:13 <mboddu> puiterwijk: :)
17:46:21 <mboddu> puiterwijk: Any update on s390x?
17:46:44 <puiterwijk> S390x /mnt/koji not yet fixed, but I will get that done tonight since I'm not going to get any happier with it
17:47:12 <mboddu> puiterwijk: tonight?
17:47:17 <puiterwijk> Sure.
17:47:31 <mboddu> puiterwijk: Okay, thats awesome
17:47:49 <mboddu> puiterwijk: Can you help what you are trying to do in simple terms :D
17:47:51 <mboddu> ?
17:48:05 <puiterwijk> #info puiterwijk wishes we would have enough money for an s390x mainframe in PHX2
17:48:23 * mboddu wishes for the same
17:48:46 <puiterwijk> mboddu: basically, adding an ssh agent on the buildvm
17:48:52 * mboddu wonders how much does it cost, I know its a lot but...
17:49:01 <mboddu> puiterwijk: ahh okay
17:49:20 <mboddu> #undo
17:49:20 <zodbot> Removing item from minutes: INFO by puiterwijk at 17:48:05 : puiterwijk wishes we would have enough money for an s390x mainframe in PHX2
17:49:31 <mboddu> puiterwijk: sorry, I have to undo it :(
17:49:33 <puiterwijk> :( that's not any less true
17:50:13 <puiterwijk> mboddu: the system Z pricing is not even available online. I think that says enough :)
17:50:25 <mboddu> puiterwijk: sure, but the minutes go out to the mailing list :(
17:50:30 <puiterwijk> Ah, right. Okay
17:50:34 <mboddu> puiterwijk: haha
17:50:51 <nirik> if I had enough money for 2 of them I would be sure and get one for phx2. ;)
17:51:09 <mboddu> #info puiterwijk is planning to fix the mounting s390x issue tonight by adding an ssh agent on the buildvm
17:51:20 <puiterwijk> #undo
17:51:20 <zodbot> Removing item from minutes: INFO by mboddu at 17:51:09 : puiterwijk is planning to fix the mounting s390x issue tonight by adding an ssh agent on the buildvm
17:51:28 <puiterwijk> #info puiterwijk is planning to fix the s390x issue tonight
17:51:39 <puiterwijk> mboddu: I'd rather not attract attention to the agent stuff
17:51:45 <puiterwijk> I really, really hate it
17:51:48 <mboddu> puiterwijk: oh no particulars, got it
17:51:54 <mboddu> puiterwijk: understood
17:52:29 <mboddu> nirik: I hope you will a lottery :D
17:52:37 <mboddu> nirik: enough to buy 2 of them
17:52:52 <mboddu> Okay, last topic
17:52:56 <mboddu> #topic Open Floor
17:53:34 <mboddu> #info We just got a GO for the f27 beta for RC 5
17:53:39 <puiterwijk> \o/
17:53:49 <puiterwijk> Congrats everyone on that one :)
17:53:50 <mboddu> anyone want to share anything else?
17:54:46 <mboddu> puiterwijk: what is your secret on working 24 hours a day? I need some help, cant keep my eyes open today(had like 5 hrs of sleep)
17:55:03 <puiterwijk> mboddu: hehe. It's a secret :).
17:55:12 <mboddu> puiterwijk: PM me then :D
17:55:20 <puiterwijk> Just open a PR against the earth for longer days?
17:55:53 <puiterwijk> I'm sure the length of a day is just a configurable somewhere
17:55:54 <mboddu> puiterwijk: where is the code. "Dont talk, show me the code" - Linus
17:56:14 <puiterwijk> mboddu: it's proprietary. Just request :)
17:56:23 <mboddu> puiterwijk: heheheheh
17:56:38 <mboddu> puiterwijk: sorry, I want to stick with open source :D
17:56:45 <puiterwijk> :-)
17:56:48 * puiterwijk agrees
17:57:33 <mboddu> Anyway, we no one has anything else, then I will end the meeting and give 2 mins back to you guys :)
17:57:33 <nirik> no no, it's 100% open source...
17:57:47 <nirik> just that the issue tracker is impossible to find. ;)
17:57:53 <puiterwijk> Right :)
17:58:00 <mboddu> nirik: what are they using? Bugzilla?
17:58:16 <mboddu> :)
17:58:29 <puiterwijk> mboddu: nah, bugzilla is findable. I'm sure it's a quantum issue tracker
17:58:55 <mboddu> puiterwijk: ahhh, thats right
17:59:10 <mboddu> Anyway, thanks for joining guys, at least able to give 1 min back
17:59:14 <mboddu> #endmeeting