18:01:05 #startmeeting Infrastructure (2017-04-20) 18:01:05 Meeting started Thu Apr 20 18:01:05 2017 UTC. The chair is puiterwijk. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:05 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:01:05 The meeting name has been set to 'infrastructure_(2017-04-20)' 18:01:06 #meetingname infrastructure 18:01:06 The meeting name has been set to 'infrastructure' 18:01:08 #topic aloha 18:01:10 #chair smooge relrod nirik abadger1999 dgilmore threebean pingou puiterwijk pbrobinson 18:01:10 Current chairs: abadger1999 dgilmore nirik pbrobinson pingou puiterwijk relrod smooge threebean 18:01:12 morning 18:01:12 Let's try this again :) 18:01:14 hey 18:01:14 Hello everyone! 18:01:18 hello 18:01:19 hello 18:01:20 Sorry for the late start, I messed up my meetbot-fu :) 18:01:21 hello 18:01:22 Hello everyone! 18:01:23 Hey 18:01:31 hi, I'm Shailendra 18:01:34 #topic New folks introductions 18:01:34 * cverna is here 18:01:44 Hi, any new folks that want to give a one-line introduction? 18:02:14 Hi everyone, Michael O'Brien from the UK, I've been working as a contractor for past 15 years mostly as Unix Admin/Engineer and last 5 years as an infra architect. 18:02:14 hi all 18:02:14 I'm an android developer basically 18:02:42 Having the basic of linux, done RHCE participation certificate 18:02:47 michaelob, yadavShailendra: welcome! Are you more interested in system admin stuff or application development? 18:02:56 I'm looking to keep my hand in Unix admin which is my favourite thing, helping on the fedora project seem the prefect opportunity to do this. 18:02:58 I'm up 18:04:01 Happy to help out where need, I would like to used my OpenStack skills and learn some Ansible if possible, thanks 18:04:43 michaelob, yadavShailendra: great. Ping me after the meeting in #fedora-admin and I can get you started 18:04:48 hi everyone 18:05:01 okay thanks 18:05:03 welcome new folks. :) 18:05:09 cheers 18:05:12 If no other new people, let's move on to the announcements 18:05:18 #topic announcements and information 18:05:20 #info beta freeze will start 2017-05-17 18:05:22 #info infra hackfest in RDU 2017-05-08 to 2017-05-12 - everyone 18:05:24 #info bodhi-2.6.0-0.1.beta is deployed on stg. 2.6.0 will likely be released/deployed next week - bowlofeggs 18:05:27 Anyone else with more? 18:05:55 .hello bowlofeggs 18:05:56 bowlofeggs: bowlofeggs 'Randy Barlow' 18:06:03 .hello jcline 18:06:04 jcline: jcline 'Jeremy Cline' 18:06:40 Let's move on to discussion then 18:06:42 #topic jenkins cleanup (https://jenkins.fedorainfracloud.org/view/infrastructure/) - cverna 18:06:47 cverna: take it away 18:07:07 * nirik added a few status items too, but thats fine. See them in gobby. 18:07:22 just a quick one about jenkins 18:07:43 we seems to have quite a few project that are not building since a long time 18:07:45 nirik: oh, oops... my gobby crashed it seems :/ 18:08:15 or even never even built successfully once 18:08:48 I was wondering if some spring cleaning could be done 18:09:09 We could ask the project owners, sure. But "not building" does not mean that jenkins is useless :) 18:09:28 (for example, I know that Ipsilon fails four out of five times due to something in jenkins, but it's still useful to me) 18:10:00 we could disable builds for projects that are failing for too long, to conserve resources, and let owners fix them and re-enable 18:10:21 for example last successful build for fedocal was 9 month and 18 days 18:10:29 this is not very useful :s 18:10:34 Right. Some projects could definitely be cleaned 18:10:42 or well, maybe/probably :) 18:10:47 well, by cleaning... you mean help them get working? or ? 18:11:10 I got the feeling they mean to remove them 18:11:18 I think it might be 1 first pass to identify if the builds are still needed 18:11:33 for example do we need fedora-mobile ? 18:11:46 right, there may well be some projecs that no longer exist 18:12:07 or where maintainer doesn't want jenkins anymore 18:12:26 (or where jenkins is still following fedorahosted! :) ) 18:12:27 and a second activities would be to investigate the failures and try to get them working 18:12:28 i would be more motivated to fix the bodhi one if we had the github PR testing plugin 18:12:45 bowlofeggs: got a fedora infra ticket requesting it? 18:12:52 puiterwijk: i do 18:12:55 Cool 18:12:59 https://pagure.io/fedora-infrastructure/issue/5592 18:13:14 bowlofeggs++ 18:13:47 without that, i only know about failures after they are already merged 18:15:07 packaging up those would be a lovely thing for some one to do 18:15:08 so to wrap up, we need to identify which project could be removed, then try to get the long time failure up and running again 18:15:42 PR verification may be problematic security-wise: anyone can submit PR and execute arbitrary code on builder 18:16:17 mizdebsk: agreed. But we're already doing that for Pagure. 18:16:37 (not saying that that makes me happy, but well, it's cloud. And it's what we've been doing for a while now) 18:17:31 I think a number of projects handle that by only running on project members PRs or ones they ask for testing on 18:17:43 dunno if jenkins has any way to tell that 18:18:00 nirik: reasonably sure Pagure's jenkins plugin has no way to do that 18:18:07 i bet it does, jenkins has plugins for virtually anything :) 18:18:18 mizdebsk: true there. Just not our current setup :) 18:18:46 But I think that could be useful, sure 18:18:56 mizdebsk: we could have a list of pre-approved people to test PRs for, and if the PR is from someone else it could watch for a keyword from an approved person 18:19:40 the keyword would be nice, because i could kick off a test after reviewing the PR with my soft human brain 18:20:07 Anyway, I think the main question is if we wanted to clean out jenkins. We can totally continue this discussion in the ticket we have, or after we get through the cleanup 18:20:11 discussions 18:20:32 So, do we think we need to clean up jenkins, and if yes, is anyone willing to take a first hit at it? 18:20:56 pagure does already allow retesting on a keyword... 18:21:13 I'd like to take a shot.. if I can get to that host 18:21:27 or provided I have the perms 18:21:27 Skeer: well, you would just look at the web interface I guess :) 18:21:34 https://jenkins.fedorainfracloud.org/ 18:21:42 Ahh *facepalm* 18:21:44 Look for projects that haven't built for a long time, or that have only failed for a while 18:22:01 Roger that 18:22:30 We use Jenkins here at work so I can navigate well enough 18:22:39 but we also then need to ping maintainers/owners. 18:22:47 So maybe just create an infra ticket for it, and track the long failing projects there 18:22:47 is that info available? 18:23:13 nirik: I think most (all?) jenkisn projects have been requested via the trac/pagure, so the tickets should still be there, not? 18:23:29 yeah, but might be not easy to find. ;) 18:23:31 And if unsure, someone in #fedora-admin probably knows 18:23:36 no, i don't think owners are exposed in web for anonymous users 18:23:42 but i can make a list if needed 18:23:51 (it's all stored in xml files) 18:23:57 That sound sgood 18:24:19 I can raise an infra ticket to track the work 18:24:29 cverna: was just going to ask. Thanks :) 18:24:42 wouldnt the owners or at least the creator listed if project-based security is enabled? 18:24:44 more tickets. yea! 18:24:45 :) 18:24:46 ;) 18:25:03 Skeer: should be. but not clear to who that info is available 18:25:05 nirik: I can just hardcode Pagure to always show a count of "20 open tickets" for fedora-infra :-) 18:25:23 nirik: I can see that info for some projects, but not all 18:25:25 man, still 20 tickets. I am making no progress at all! 18:25:46 Skeer: huh, well, we can figure it out in ticket 18:25:50 mizdebsk: If you generate that list can you paste/attach it to the ticket? 18:26:19 Skeer: sure, will do 18:26:32 unless someone is against that 18:26:39 Nope, sounds good to me. 18:26:55 Anything else on this? 18:27:21 Are you going to assign it to me or would it be better open? 18:27:35 Skeer: just let us know and we can assign it to you 18:27:47 Go ahead then.. Ive got this. 18:28:00 Okay. As soon as cverna creates the ticket I'll assign it then 18:28:07 #topic announcements and information - continued 18:28:07 I am on it :) 18:28:13 Seems I missed two announcements: 18:28:15 #info koji ssl cert expired reminder cron removed - kevin 18:28:17 #info s390x builders are pretty close to ready to go - kevin 18:28:45 #topic Apprentice Open office hours 18:28:59 Any apprentices with questions or remarks? 18:29:33 What shall I do? Just solve tickets? 18:29:50 Meto1: did you see our list of easyfix tickets? 18:29:54 yes 18:30:04 I solved one 18:30:10 Great! 18:30:35 You are more than welcome to look through the non-easyfix tickets as well if you want. 18:30:47 oh hey puiterwijk can you assign that packagereview ticket to me as well? 18:30:49 Just let us know and we can help you get started with stuff. 18:31:00 Skeer: sure 18:31:17 I will, thanks. Going msg you if I get stuck on something 18:31:25 Meto1: sounds good 18:31:36 If/ when I decided to reactivate, just an email to the list? 18:31:49 puiterwijk: Skeer https://pagure.io/fedora-infrastructure/issue/6003 18:31:54 odin2016: yeah, or just say hi on #fedora-admin. We don't require another hello :) 18:32:01 cverna: thanks! 18:32:11 thanks cverna 18:32:12 odin2016: just let us know if you want to be re-added to the apprentice group 18:33:16 Cool, thanks. I likely will in the next little bit here... things are quieting down, the arch boxes I have are playing nice. ansible is making my life easy, so.. ;) 18:33:30 Cool. You're always welcome. 18:34:01 Oh hey question abotu git'ing 18:35:03 I was under the impression that we should be keeping commits and pull requests seperate, is this not the case? 18:36:07 Doesnt matter to me either way.. I just want to keep things organized how ya'll like 18:36:21 Well, it's useful to have PRs be centered around a single change 18:36:53 * nirik nods. 18:37:05 Note that we're not very strict with the git or PR use in Infra. We try to point out if things are a big mess, but as long as it's reasonably sane, you should be good 18:37:18 If you have any concrete problems where you're wondering how to split it up, just ask. 18:37:33 OK I guess his wording just lost me a bit in the comments 18:37:34 We have some people that are experienced with git, and some that are less. We're always willing to help guide 18:37:50 I def fall on the lesser side of that scale :) 18:38:29 So, a rule of thumb is that PRs can be the same split as a ticket. In many projects, smaller tickets equal one PR that closes that ticket 18:38:32 often people like 1 commit on a PR... instead of 1 that does part and a fix that does some more and another fix, etc... 18:38:44 but it's pretty easy to squash those down if desired. 18:39:30 my two tickets, adding teo different sops into infra-docs, I guess logically it seemed better to be to keep them apart 18:39:31 i'm a big believer in commits being atomic - meaning that they fix exactly one thing 18:39:41 ^ makes total sense 18:39:48 i don't mind multiple commits in a PR if they are related or if they build upon each other 18:39:51 Skeer: yeah, I would definitely say 2 there... 18:40:15 bowlofeggs: right, but if they are just changing the same thing 3 times because you were not sure what to change... it's kinda noisy 18:40:17 So I guess Im lost on the squashing of commmits 18:40:25 so, for example, if you want to fix a style related thign you notice in code while fixing a bug there, it's good to fix the style thing in a separate commit but it's fine to put it in the same PR 18:40:57 yeah i don't like multi-commit PRs when the commits are completely unrelated 18:41:12 LIke mine here: https://pagure.io/infra-docs/pull-request/25#commit_list 18:41:25 I readily admit.. this was my first pr and it's rather nasty 18:42:01 yeah, but multiple commits there may have been fine if they were doing distinct things... 18:43:13 we are all pretty easygoing I think, so in the end just ask or adjust if someone feels strongly, otherwise just getting things done is the important part. 18:43:16 I had issues early on with adding two files into the same commit and pr then technicial difficulties removing the correct file correctly and.. it was a mess 18:44:18 bowlofeggs: if you point me in the right direction to cleaning up those commits Ill take car eof them 18:45:28 Skeer: this article looks helpful: https://github.com/edx/edx-platform/wiki/How-to-Rebase-a-Pull-Request#squash-your-changes 18:45:52 basically, squashing lets you rewrite the history of your branch and turn those commits into one commit 18:46:21 btw, in the future you can use git commit --amend to amend the most recent commit, and then git push -f to force push the rewritten history to your fork 18:46:27 that's what i usually do 18:46:43 so that way you never end up with extra commits 18:46:43 Oh nice.. here I was thinking you used squash as just a slangy term for somehitng else. Had no idea it was a thing, lol 18:46:51 you always just have the one and you edit it until it's accepted 18:47:02 hahaha yeah it's a git thing 18:47:16 sweet Ill work on that and update the comments when done 18:47:18 also good to keep in mind that you should never do either of these thigns to a shared repo 18:47:32 like, i would not do that to the bodhi repo, only to my fork of it 18:47:39 Id expect the history is a good thing to have in that case 18:47:41 Rewriting the history of the main repo is a bad idea :) 18:47:43 because rewriting history breaks other clones 18:47:47 right. that messes up the history for others 18:47:48 right, right 18:47:57 * cverna uses --amend a lot when bowlofeggs reviews his PRs :D 18:48:01 understood 18:48:17 Does that make things more clear? 18:48:21 def 18:48:25 thanks fellas 18:48:35 Great. If you have any concrete questions, you can always ask and we'll try to answer 18:48:45 bowlofeggs++ 18:48:45 Skeer: Karma for bowlofeggs changed to 12 (for the f25 release cycle): https://badges.fedoraproject.org/tags/cookie/any 18:48:48 Any other apprentices with questions? 18:49:04 awww thanks 18:49:20 :) 18:50:25 puiterwijk just make sure they're marked as easyfix ;) 18:50:43 If nothing else, let's go on to open floor 18:51:03 ExoUNX: just let us know if you think tickets should be marked as such :) 18:51:24 #topic Open Floor 18:51:36 Any other questions, remarks, or favorite modes of transportation? 18:51:49 hover board.. 18:52:05 The mexican "Metro" 18:52:16 transporter bean? 18:52:20 beam 18:52:26 is that a special kind of coffee? 18:52:39 :) 18:52:49 what about nullcycle? it's like a unicycle but it has 0 wheels instead of 1 18:52:57 nirik ++ 18:53:02 sftp is mine. 18:53:05 bowlofeggs++ 18:53:05 smooge: Karma for bowlofeggs changed to 13 (for the f25 release cycle): https://badges.fedoraproject.org/tags/cookie/any 18:53:10 tcp ;D 18:53:26 everyone sees me moving, no clue that it's me. 18:53:38 bowlofeggs: think of how little friction it has? 18:53:42 I am quite fond of TLS :-) And in the physical world, trains. 18:53:45 cyberworm54_ SYN-ACK 18:54:05 Trains are nice, Especially steamy's. 18:54:08 * puiterwijk expects an RST now, because of missing SYN 18:54:36 Anyway, if nothing else, let's call it a meeting. Please continue all conversations in #fedora-admin and #fedora-apps. 18:54:38 A sql attack walks into a bar and orders a DROP tables; and all the tables in the bar fall over. 18:54:44 Thanks all for coming! 18:54:48 bowlofeggs++ 18:54:49 nb: Karma for bowlofeggs changed to 14 (for the f25 release cycle): https://badges.fedoraproject.org/tags/cookie/any 18:54:49 nirik: hah. 18:54:54 Hi team 18:54:58 nirik++ 18:55:01 nice, nirik 18:55:05 hahah thanks ☺ 18:55:08 #info nirik gets the joke of the meeting award 18:55:15 nirik++ 18:55:15 ExoUNX: Karma for kevin changed to 40 (for the f25 release cycle): https://badges.fedoraproject.org/tags/cookie/any 18:55:16 #info (decided by me) 18:55:22 #endmeeting