18:02:21 #startmeeting Infrastructure (2015-04-09) 18:02:22 Meeting started Thu Apr 9 18:02:21 2015 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:02:22 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:02:22 #meetingname infrastructure 18:02:22 #topic aloha 18:02:22 #chair smooge relrod nirik abadger1999 lmacken dgilmore mdomsch threebean pingou puiterwijk 18:02:22 The meeting name has been set to 'infrastructure' 18:02:22 Current chairs: abadger1999 dgilmore lmacken mdomsch nirik pingou puiterwijk relrod smooge threebean 18:02:22 #topic New folks introductions / Apprentice feedback 18:02:30 sorry about the delayed start. 18:02:39 I am in the go/no-go meeting and a bit distracted by that 18:02:52 * relrod here 18:03:02 but might duck out early 18:03:40 also many folks are at pycon. ;) 18:03:41 * adrianr here 18:03:43 Hello everyone, I'm an apprentice. 18:03:45 here 18:04:03 I will take over 18:04:25 #topic New folks introductions / Apprentice feedback 18:04:25 = Status / information / Trivia / Announcements = 18:04:38 hello d3prof3t 18:04:49 anything we can do to help you? 18:05:19 Yea, actually I was going through the easyfixes and I thought to start off with #50 ticket 18:05:27 And it was under you. 18:05:55 .... uhm do you have a link :) 18:06:12 Although I don't know exactly where to start working exactly, but I just chose it so that I atleast start getting to know about all of this stuff 18:06:14 wait 18:06:56 smooge: https://fedorahosted.org/fas/ticket/50 18:06:58 hola all 18:07:06 thanks 18:07:20 hmm 18:07:47 smooge: although It would be really helpful if I could start working on python stuff 18:08:09 smooge: as I am a bit fluent with it 18:08:27 * smooge looks at the ticket.... 18:08:42 ok everyone that would be helpful on that are at PyCon today 18:09:20 Ohh okay 18:09:38 where is today's Pycon happening? 18:09:45 *PyCon 18:09:57 we used to have a fakefas system which could be used for seeing how fas runs. I will see if I can find the steps for that and if you can email me (smooge@gmail.com) I will send them to you 18:10:05 I believe it is in Montreal Canada 18:10:48 smooge: Okkay I will mail u a Hello World :p 18:11:03 thanks any other apprentices with questions? 18:11:12 Hopefully the fas docs explain how to make a local one. ;) 18:11:52 nirik, they always seem to require a lot of 'well why the f doesn't that work anymore' 18:12:17 a great chance to help improve the docs. ;) 18:12:41 anyhow, you want to keep going running things, or want me to? 18:12:42 yeah.. the changes usually were 'hand problem to abadger1999' 18:12:54 nirik, is go/no-go over? 18:13:18 not officially, but since it's no go, we just are waiting for the singing at the end. ;) 18:13:19 if it is over I will be happy if you want to take over. if not I will keep running also 18:13:41 #topic announcements and information 18:13:41 #info New release of Fedora Badges deployed with updated leaderboard, profiles, and frontpage - ralph 18:13:41 #link https://badges.fedoraproject.org/ 18:13:41 #info two new pkgdb release (bug fixes on pkgdb-sync-bugzilla) - pingou 18:13:41 #info new anitya release (mostly bug fix) - pingou 18:13:42 #info 18:13:59 #info I need to learn cut and paste better.... 18:14:58 so ralph (at pycon) worked a lot on the badges.. and pingou (at pycon) did a lot of work with anitya, pkdb and I think new-hotness 18:15:52 #info we are in a freeze so changes to running systems require an on-list 2 +1 from sysadmin-main to occur 18:15:59 I find this badge concept really good :). Nice job 18:16:39 It is fun for people in the project 18:16:57 Any other announcements or move onto "Things we should discuss?" 18:17:05 move on. 18:17:14 I wanted to give a quick update about MM2 18:17:33 #topic Quick Update about MM2 18:17:37 adrianr, all yours 18:17:55 I have been bisecting the curl segfault for over a week and it seems the segfault is in every version of curl 18:18:08 :( 18:18:10 the Fedora 21 libcurl also crashes, but not so often 18:18:30 adrianr: so whats the next step? newer? f22? 18:18:32 I will now try to recreate the segfault outside of MM2 for easier debugging 18:18:45 I tried the newest upstream git 18:18:52 ah. 18:19:08 I hope this is possible without too much effort 18:19:33 Sir 18:19:46 but if pingou can make a release I could start moving the installation from staging to production 18:20:01 and if the segfault is fixed it would be ready 18:20:10 Can you give me a link to this project so that I can read what its all about 18:20:18 adrianr: well, yeah... but we need to get out of freeze before making the prod instances... 18:20:32 another solution would be to go back to a forking approach and ignore the threading problem 18:20:42 sure, after the freeze 18:20:54 d3prof3t: https://bugzilla.redhat.com/show_bug.cgi?id=1204825 18:21:13 that's it 18:21:18 adrianr: thanks 18:21:20 thanks for working on this adrianr 18:21:27 adrianr, thanks for all the work 18:21:35 adrianr++ 18:22:04 #topic do we need warm spare hosted/collab instances anymore? - kevin 18:22:14 nirik, you ok to run on this? 18:22:19 so, I thought I would bring this up. 18:22:39 right now we have hosted and collab (where mailing lists are), with 'warm spares' at another datacenter. 18:22:48 This is from back in the days when things were unstable with them. 18:23:01 We have never used the warm spare instances 18:23:12 so, I am thinking we might just retire them. 18:23:30 It would mean if osusol went off the net or our machines there died, those things could be down a bit... 18:23:36 we do have backups of them both 18:23:39 sounds good enough to me, as long as we have backups 18:23:59 so worst case we spin up new ones and restore, which could take a while, but should work fine. 18:24:56 anyhow, thought I would bring it up in case anyone felt strongly about it. 18:25:00 thats all. ;) 18:25:04 I am for it for no other reason than our use of gluster to clone the disks has always gotten a "really we said you could do that?" from the gluster filesystem gusy :) 18:25:19 smooge: we aren't using gluster tho. ;) 18:25:20 smooge: well, I think we no longer use gluster for this 18:25:21 it's all rsync 18:25:24 we're just rsyncing 18:25:31 every hour they rsync all the data from one to another 18:25:34 ah well there goes my being on top of things 18:26:03 gluster was the time where I had a cronjob that would recreate all trac databases because gluster would corrupt them 18:26:23 I think we did this setup after the great gluster blow up 18:26:30 yup 18:26:52 anyhow, thats all I had on this... 18:26:55 I am getting old.. I thought I ran into gluster on our boxes last week and it was those system 18:26:58 ok 18:27:00 I'm personally in favor 18:27:23 smooge: we do use gluster on some other boxes, e.g. askbot and nuancier, for uploaded files 18:28:09 nirik: perhaps we should consider renaming the boxes when they're migrated to RHEL7. I always think it's odd the primary node is hosted03 :) 18:28:30 puiterwijk: sure. I was going to make the new one hosted01 18:28:32 ah it was askbot 18:29:00 nirik++ 18:29:00 puiterwijk: You have already given 1 karma to kevin 18:29:05 ok so we are at the learning about section 18:29:23 and since no one signed up. we skip 18:29:25 well foo. I can't give adrianr karm but you can 18:29:27 which reminds me I should put SSO/SLO for next week 18:29:28 so on to open floor. ;) 18:29:41 #topic open floor 18:29:44 adrianr++ 18:29:47 What is SS0/SLO? 18:29:55 d3prof3t: Single SignOn and Single Logout. 18:30:06 okkay 18:30:07 a major project I'm currently working on to simplify authentication to our webapps 18:30:20 okkay 18:30:54 puiterwijk: I was having one thing on my mind 18:31:11 d3prof3t: what is it? 18:31:26 puiterwijk: I was looking at the apps and I noticed most of them had TG frameworks running 18:31:31 right? 18:31:36 d3prof3t: no, not anymore 18:31:42 we still have one TG1 app remaining, bodhi1 18:31:47 the rest is all rewritten in Pyramid or Flask 18:31:49 well, a few more 18:31:54 mirrormanager1 18:31:57 okkay 18:31:59 fas2 18:32:07 nirik: I think fas2 is tg2? 18:32:20 Actually nowadays I am doing some personal stuff on bottle frameowork 18:32:28 Its really just for begginers 18:32:41 puiterwijk: don't think so. 18:32:51 nirik: ah, okay. I might be mistaken. so we have a few 18:32:57 but most of them are flask/pyramid nowadays 18:32:58 I mean I havent seen it working anywhere 18:33:05 yea exactly 18:33:46 d3prof3t: I guess you were going to ask about the tg1 "single login" stuff where all apps just use the same cookie? or not? 18:34:00 So, what I was saying was if I could just be an assistance sort in any project of a web app, so that i can see how it works and then I can also give a lille contribution to web apps 18:34:07 yea 18:34:41 yea so I needed some assistance so that I can learn them 18:34:50 we had that before, but that depended on a shared key and central connections to FAS, but we got rid of that 18:35:10 hmm 18:35:23 nowadays, we're switching to OpenID for our apps 18:35:51 Okkay, so is there any resource to read for it.. or just google that thing 18:35:55 and single login/logout for OpenID is slightly more tricky. Thouigh I have a setup for it, and we're now adding that to all of our apps 18:36:14 if you ping me in #fedora-apps after the meeting, I'd be glad to explain everything :) 18:36:28 puiterwijk: Okkay I will 18:36:36 puiterwijk: Thanks 18:36:41 anything else for open floor anyone? 18:38:22 #endmeeting