13:59:55 #startmeeting Infrastructure (2018-05-03) 13:59:55 Meeting started Thu May 3 13:59:55 2018 UTC. 13:59:55 This meeting is logged and archived in a public location. 13:59:55 The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:59:55 Useful Commands: #action #agreed #halp #info #idea #link #topic. 13:59:55 The meeting name has been set to 'infrastructure_(2018-05-03)' 13:59:55 #meetingname infrastructure 13:59:55 The meeting name has been set to 'infrastructure' 13:59:55 #topic aloha 13:59:55 #chair smooge relrod nirik pingou puiterwijk tflink 13:59:55 Current chairs: nirik pingou puiterwijk relrod smooge tflink 14:00:06 .hello2 14:00:07 x3mboy: x3mboy 'Eduard Lucena' 14:00:10 morning 14:00:26 hello everyone. 14:00:29 * relrod_cld waves 14:00:29 hello 14:00:33 hello 14:00:36 * sayan waves 14:00:37 Hello people! 14:01:00 hello. In fixing stuff, so not realyl here. 14:01:15 .hello2 14:01:16 bowlofeggs: bowlofeggs 'Randy Barlow' 14:01:17 .hello2 14:01:19 tflink: tflink 'Tim Flink' 14:01:24 i'm breaking stuff, so i am really here 14:01:44 nirik, can you move the fedocal time and channel to this channel and time? 14:01:58 .hello2 14:02:00 yep. 14:02:01 dustymabe: dustymabe 'Dusty Mabe' 14:02:16 .hello sinnykumari 14:02:17 .hello sayanchowdhury 14:02:18 ksinny: sinnykumari 'Sinny Kumari' 14:02:21 sayan: sayanchowdhury 'Sayan Chowdhury' 14:04:24 #topic New folks introductions 14:04:25 #info This is a place where people who are interested in Fedora Infrastructure can introduce themselves 14:04:52 helo 14:05:27 hello sapo 14:05:39 how are things for you? 14:06:11 Good, thanks for asking. 14:06:31 * aikidouke_ is here 14:06:33 I'm working on the network map ticket. 14:06:35 welcome sapo 14:06:58 I may have some questions when we reach the topic 14:07:13 sapo, ok thanks 14:08:35 Do I still count as new? 14:08:38 I would like to thank Anatoli for offering pull requests 14:08:47 THanks Anatoli! 14:09:08 OK next up as I focus on one ping at at a time 14:09:11 #topic announcements and information 14:09:11 #info Fedora 28 released! Kudos to everyone! 14:09:12 #info Freeze is over 14:09:12 #info Pagure 4.0 being deployed 14:09:12 #info New openshift being deployed 14:09:13 #info A Bodhi 3.7.0 release is being worked on. It mostly addresses some usability issues with test gating, and also adds support for links to privacy/legal. currently targeting a release early next week. 14:09:36 #info mass update/reboot cycle next tuesday 14:10:33 #info Ibiblio data move 2018-05-09. Fedora people and other systems will be down for N hours 14:10:59 #info short switch outage later today (21UTC) 14:11:17 #info who gets the last info 14:11:36 any other announcements? 14:12:13 i announce that i enjoy working with all of you. 14:12:22 awww 14:12:33 bowloftapioca - thanks! 14:12:40 :) 14:12:54 ok I think we are now into mushy stuff so next topic 14:12:58 #topic Tickets discussion 14:12:58 #info https://pagure.io/fedora-infrastructure/issues?status=Open&priority=2 14:13:08 bowloftapioca++ 14:13:11 There's only a few left from last week... 14:13:24 oh, one actually. 14:13:34 so it looks like we have 1 to discuss this week 14:13:52 we ran out of time last week and didn't really get to this one much 14:14:03 https://pagure.io/fedora-infrastructure/issue/6241 14:14:44 My first question comes from trying to figure out where this sits. To me this looks like releng, but we write the apps 14:15:07 yeah, the state is not super clear. ;) 14:15:32 so the orig case was say you maintain package foo and don't want to maintain it's epel branches. 14:15:33 wihtout per-branch ACLs i don't see how we could do this 14:15:51 you can make bugs get assigned to orphan 14:15:53 well i guess i don't mean ACL exactly, but we don't have per branch "point of contacts" 14:15:58 we do tho 14:16:05 oh, well, sorta only 14:16:15 it's epel vs fedora only 14:16:23 yeah 14:16:29 but epel is still a branch 14:16:38 I expect this is going to also happen for other types of branches 14:16:48 i ugess there's that weirdo yaml file in some releng repo somewhere that defines some of this stuff 14:17:10 so you can set orphan to be assigned all epel bugs... but then the only way we know to retire it or have someone take it is that assignment 14:17:21 i am not a fan of the way that works btw - i wanted to adjust that once and finding the docs was hard, and then it took over a week to get my PR merged (all so i could have the-new-hotness file bugs for me) 14:17:48 another problem is that we don't have anything or anyone looking for orphaned things anymore 14:18:01 there was a fesco discussion about this sometime recently 14:18:08 they are supposed to be retired after X time... 14:18:17 so if they get orphaned they will stay 4eva as it is 14:18:18 and I think some folks were looking into reviving that 14:18:21 unless someone fixes that 14:18:22 yeah 14:19:07 So, I guess I am kinda at: lets document this and move on... 14:19:08 should we mark this ticket as waiting on someone to fix that? 14:19:23 "that" being the orphan reaper script? 14:19:36 or yeah, we could just document how to make it an orphan 14:19:44 i guess we need to document it anyway 14:19:53 ok who is documenting it? 14:19:56 yeah, at least in the 'new' docs 14:20:08 is it already documented in those docs i said were hard to find? 14:20:12 * bowlofeggs isn't sure he could find it again 14:20:23 ha 14:20:30 I can document it somewhere 14:20:55 it's this i think? https://pagure.io/releng/fedora-scm-requests 14:21:22 though i don't know if that talks about orphaning 14:21:40 ok I hae assigned it to nirik who has no tickets 14:21:44 yep.. thats where to do it... docs might be wiki or the new packager area 14:21:48 so are we suggesting that they just change the BZ assignee for epel to orphan? 14:22:03 yeah, and then whoever is retiring packages also look at that 14:22:08 cool 14:22:50 ok, shall we move along? 14:22:53 Next up 14:22:58 #info https://pagure.io/fedora-infrastructure/issues?status=Open&priority=3 14:23:10 These are tickets which are looking for people to fix them 14:23:52 i would appreciate that waiverdb token for staging bodhi in the next couple days if it's a possibility - it would help me test the 3.7 beta i'm working on. it's not a huge deal if that's not possible 14:24:14 i.e., it'd be real nice, but i've been living without it and the world hasn't ended :) 14:24:17 or better: https://pagure.io/fedora-infrastructure/issues?status=Open&priority=3&assignee=0 14:24:29 (ones with no assignee) 14:24:35 thanks 14:24:49 bowlofeggs: I am not 100% sure where to get that... so I have left it alone. 14:25:09 i fear that only patrick knows how to do it and i don't want to pile too much on him :/ 14:25:15 #info https://pagure.io/fedora-infrastructure/issues?status=Open&priority=3&assignee=0 [Tickets that need assignment] 14:25:20 as i said, it's not critical 14:25:28 and i don't want to pull him off more important things 14:25:31 What do I know? 14:25:32 Please look at these and see if you can adopt one. We will start reviewing them next week 14:25:44 patrick you know way too much :) 14:25:44 bowlofeggs: what ticket do I know? 14:25:54 puiterwijk: https://pagure.io/fedora-infrastructure/issue/6891 14:26:00 checking 14:26:16 Okay. Right. Ping me post-meeting, and I can take care of that 14:26:26 i can take #6872 14:26:34 Takes about 5 minutes, but I'm currently trying to fix repo stuff so I need to focus 14:26:35 now that we are unfrozen 14:26:54 puiterwijk: yeah you are working on More Important Thingsā„¢ 14:27:17 OK next topic? 14:27:38 #topic Apprentice Open office minutes 14:27:38 #info A time where apprentices may ask for help or look at problems. 14:27:52 Are there any questions from apprentices on items? 14:28:11 o/ 14:28:21 hello sapo. 14:28:26 Any tasks/projects coming up that need assistance? 14:28:31 So about that network map ticket. 14:29:03 I planned to make actually two different ones: 14:29:19 1: Virthosts / VMs 14:29:38 2: Actual network map with all machines all IPs 14:29:40 creaked: see the lists above. ;) Some likely require access or the like, but some might be possible to do a lot on without any access. 14:30:22 I already wrote a script that collects all the VM Host / VM relations 14:30:26 * nirik realizes he owes a review to an apprentice too. 14:30:56 I used the virthost-list.out file from Nagios 14:31:04 Is that OK in your opinion? 14:31:30 https://pagure.io/fedora-infrastructure/issue/6505 <- nagios plugin for review... I just keep not getting to it if anyone else wants to help test. 14:31:34 sapo: sounds fine to me. 14:31:56 OK. The second part. 14:32:09 @nirik hehe I've been checking that list daily/ github for things to fix. I just did not know if there was anything outside of pague, etc. 14:32:11 Where could I collect all the IP addresses? 14:32:17 ..from? 14:32:31 Hello 14:32:34 sapo: ansible has them in inventory/host_vars/ 14:32:37 sapo, on batcave there is also /var/log/virthost-lists.out 14:32:46 I was looking for something to do 14:32:55 OK Thanks I'll look into it. 14:33:47 sapo for ip addresses. I think you can read-only clone the dns db. on batcave: git clone /git/dns 14:34:01 https://pagure.io/fedora-infrastructure/issues?status=Open&tags=easyfix 14:34:29 * nirik is happy to expand on whats needed for any of those easyfix tickets if any sound interesting 14:35:14 OK thanks smooge 14:35:58 thanks nirik and sapo 14:36:27 also apprentices can look at the ones in https://pagure.io/fedora-infrastructure/issues?status=Open&priority=3&assignee=0 and see if any of them might be things they could comment on 14:37:54 #topic Open Floor 14:38:05 OK I have one announcement from a different meeting I was just in 14:38:14 I had a quick thing to bring up.. can after smooge 14:38:20 bowlofeggs: other idea about the tokens: how about we have a talk about how to issue the tokens later, and you try to write a small py script to generate them? 14:38:26 Sorry. 14:38:41 #info Outage 2018-05-15 for systems on virthost-cc-rdu01 14:39:08 will put in announcement for that after this meeting. Should be a 'short' outage. I will be onsite to fix problems 14:39:23 that is all. over to you nirik 14:39:28 networking stuff? 14:39:35 yep. 14:39:38 puiterwijk: sure we can do that - it'd be good for someone else to know how to do it :) 14:39:49 bowlofeggs: absolutely. 14:40:08 puiterwijk: can i schedule a meeting with you to transfer that knowledge? maybe tomorrow since today sounds slammed for you? 14:40:16 I just read the other night https://blogs.akamai.com/2013/10/you-must-try-and-then-you-must-ask.html (pointed at from https://medium.com/@michaeldehaan/achieving-agility-in-engineering-processes-d7e929e65b59 ) and thought it was a good think we should all try and do... 14:40:33 basically the idea is that you work on something, if you get stuck... 14:40:35 bowlofeggs: I'll talk to you in a bit 14:40:43 cool thanks for the offer! 14:40:45 try 15min more and then ask others to help. 14:41:16 and during those 15min you document all that you tried or did or what the problem is. so you can show it to whoever you ask 14:41:23 nirik: +1 14:41:24 anyhow, I thought it was a nice rule of thum 14:41:32 yeah very much agreed 14:41:36 +1 14:41:40 i have some thoughts to piggy back on that too 14:41:43 two thoughts 14:41:43 Rules of the Ancient Pharoah Thum 14:41:58 0) oncall seems to be awesome - we should keep doing it and get people used to it more 14:42:03 nirik: that reminds me: I think most of openshift should be good. We might want to rerun to be sure, but there's a bunch of manual things to do in the middle... 14:42:19 1) i think it'd be good for more of our requests to be tickets and fewer of them to be "irc support" (if that makes sense) 14:42:20 puiterwijk: I was going to ask after the meeting. 14:42:23 ack 14:42:29 i.e., the oncall person could very frequently direct people to file a ticket 14:42:37 bowlofeggs: agreed on both counts 14:42:40 that way we have a good way to prioritize work, and keep track of things 14:42:51 it also gives more visibility into the work being done 14:43:02 which is nice when you want to tell your manager what you've been working on :) 14:43:12 we need to protect Brent^Wpuiterwijk and the rest of us. ;) 14:43:17 hahah 14:43:19 yes 14:43:29 i'm guilty of bothering the sage myself 14:43:38 i've been trying to remember to file tickets more and ask for IRC support less 14:43:53 yeah, a big problem with irc is focus 14:43:58 yeah 14:44:07 even just asking someone to look at something later means they stopped doing what they were to reply 14:44:10 sometimes i close my client when i need to focus on something 14:44:20 OK next up 14:44:20 and just check it like once and hour or so 14:45:09 who is oncall starting Monday? 14:45:19 puiterwijk has done it for a week 14:45:33 does oncall person need to be sysadmin-main? 14:45:34 I can take it, or relrod if he's ready for it 14:45:34 Well... I've done 3.5 days and I'm already going crazy 14:45:53 puiterwijk will have done it for a week 14:46:00 bowlofeggs: I'd say it doesn't need to be, but it should be someone who knows enough to prioritize stuff 14:46:11 (and then probably know who to ask) 14:46:14 i can't volunteer to do it next week because i'm kinda overloaded right now, but i could do it in the future. maybe the week after? 14:46:23 and someone who is on irc a lot to direct people 14:46:33 * puiterwijk thinks we can write a doc for that, and get random people to do that. Maybe call it a service desk 14:46:39 i wouldn't mind getting in the rotation to help out with that 14:46:45 puiterwijk++ 14:46:45 nirik: Karma for puiterwijk changed to 15 (for the f28 release cycle): https://badges.fedoraproject.org/tags/cookie/any 14:46:46 and i can direct people to file tickets too :) 14:47:03 * ksinny is curious to know what works an oncall person needs to do 14:47:06 I will be writing our helldesk sop after this meeting 14:47:12 it's basically level 1 support :) 14:47:19 it is a bit more than that 14:47:24 "can i speak to your manager?" 14:47:31 haha 14:47:38 haha helldesk 14:47:52 smooge: was that a hilarious typo or a hilarious on-purpose? 14:47:58 answer questions on irc. try and fix things if possible. direct people to file tickets. talk to others to find out how to fix the things you didn't know how to fix 14:48:10 yeah true 14:48:15 it was a typo that I kept so it became on-purpose 14:48:44 check nagios regularly, ack issues and answer pages 14:48:57 Thanks for the info 14:48:57 actually i can't do the week after next, or at least not one afternoon, because i am paying a man to put a hole in my tooth one day that week 14:49:06 so i can do all but one afternoon 14:49:28 ugh, and i need to pick up someone from the airport another day 14:49:34 yeah so 14:49:35 we can figure it out. 14:49:39 I don't think we need you to always be a full week 14:49:39 that 14:49:42 bowlofeggs: I am postponing my dentist appointment from past 1 month :/ 14:49:45 but! in 3 weeks my calendar looks much more clear 14:49:56 bowlofeggs: that's why we tell people to try .oncall rather than trying to remember who's on call 14:49:58 ksinny: :) 14:50:04 We can switch that like every minute 14:50:10 puiterwijk: true 14:50:18 well i'm willing to do a lot of the week in two weeks :) 14:50:25 thanks 14:50:36 we are at 10 minutes until the end of the meeting. 14:50:41 nirik, did we get your item? 14:50:49 wait we did 14:50:54 yeah, just wanted to note that post for folks. 14:51:04 were there any other open floor items? 14:51:16 Just one 14:51:30 your turn 14:51:35 There havce been some issues with Fedora 28 repos today and yesterday, those should be fixed. There's been metalink issues today, those should be fixed now too 14:51:42 If you still have trouble, please call the oncall 14:52:07 ok thanks puiterwijk 14:52:33 and with that.. bowlofeggs say goodnight 14:52:34 puiterwijk: I can confirm it's fixed. 14:53:19 ah sorry I messed up the George Burns straight line 14:53:32 ok I will call it an end of meeting 14:53:36 thank you all for coming 14:53:40 #endmeeting