19:01:16 #startmeeting Infrastructure (2013-10-10) 19:01:16 Meeting started Thu Oct 10 19:01:16 2013 UTC. The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:01:16 Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:01:24 #meetingname infrastructure 19:01:24 The meeting name has been set to 'infrastructure' 19:01:31 #topic welcome to all and to all a welcome 19:01:41 #chair smooge relrod nirik abadger1999 lmacken dgilmore mdomsch threebean pingou puiterwijk 19:01:41 Current chairs: abadger1999 dgilmore lmacken mdomsch nirik pingou puiterwijk relrod smooge threebean 19:01:48 * relrod here 19:01:49 * anshprat here 19:01:52 What is kickin infra team 19:02:06 * lmacken 19:02:08 hello 19:02:20 hi lmacken how is the weather in Co? 19:02:24 * ianweller here 19:02:50 We are having a front go through so I may get knocked off from a thunderstorm 19:02:57 smooge: pretty decent today, a little overcast 19:03:03 * robyduck here 19:03:15 * abompard here too 19:03:19 #topic New folks introductions and Apprentice tasks 19:03:28 Hello any new folks here? 19:03:48 new here 19:03:50 hi im new 19:04:00 Hello 19:04:01 sorry bad joke 19:04:06 hi akshayb 19:04:18 Hi 19:04:21 welcome 19:04:27 hello akshayb 19:04:30 hi frankieonuonga 19:04:36 #chair frankieonuonga 19:04:37 Current chairs: abadger1999 dgilmore frankieonuonga lmacken mdomsch nirik pingou puiterwijk relrod smooge threebean 19:04:40 Hi 19:04:44 hi smooge 19:04:46 hola 19:04:59 hello akshayb what are you looking at? 19:05:02 hi guys 19:05:16 abadger1999: ¿como estas? 19:05:38 I was looking at doing some sysadmin stuff 19:05:38 dgilmore: ola amigo 19:05:47 is dgilmore still in cusco? 19:05:55 :) 19:06:03 robyduck: no en Illinois 19:06:09 @smooge Have recently started with python 19:06:18 dgilmore: un poco enfermo. 19:06:48 abadger1999: :( 19:06:58 akshayb, cool. I am too (sort of.. ) 19:06:59 frankieonuonga: óla 19:07:21 akshayb: any bit in particular interest you? 19:08:19 I am wanting to go over the Apprentice program and flesh it out in the next 2 weeks. It really needs some docs and work to make it better (like all things). People who have been or are apprentices should email me at smooge@fedoraproject.org with suggestions of what would help 19:08:25 dgilmore: No preferences as such 19:09:00 * croberts steps in late 19:09:03 hi everyone 19:09:24 #info Apprentices please email smooge@fedoraproject with ideas on what apprenticeship means to them and what they want to do. 19:09:34 welcome croberts.we have just started.not gone too far 19:09:45 smooge: I ve a few, will send them to you.. 19:09:58 ok next item on the agenda will be 19:10:06 * frankieonuonga offering to help smooge to do the write up if it is after wed next week 19:10:28 thanks frankieonuonga it will be a couple of days 19:10:40 s/days/weeks/ to get the information so yes 19:10:45 I would welcome the help 19:10:54 #topic Applications status / discussion 19:11:06 abadger1999, lmacken etc.. up to you 19:11:23 any app work you want to cover? 19:11:32 slow week I think -- many people are out this week. 19:11:50 I've been sick and busy with fesco/fpc things so not much from me. 19:11:55 bodhi1 and 2 work underway, probably going to do a bodhi update on releng4 today 19:12:04 the wallpaper election in nuancier-lite worked. 19:12:13 there were a few things that I had to hotfix for it. 19:12:38 but nothing that blocked valid votes. 19:13:00 abadger1999: ah, I missed that.. to which list was the info for the election sent? 19:13:05 lmacken: bodhi1.x update? Or all the way to bodhi2? 19:13:07 sorry to hear you have been sick. 19:13:17 abadger1999: just a bodhi1x bugfix update. not there yet :) 19:13:17 bodhi3 with python3 :) 19:13:20 19:13:34 I'm going to try and make some headway this week on our app logging setup in stg 19:13:37 .ticket 3627 19:13:39 lmacken: #3627 (Configure centralized logging for all of our webapps.) – Fedora Infrastructure - https://fedorahosted.org/fedora-infrastructure/ticket/3627 19:13:44 frankieonuonga: what is the status on your search app 19:13:56 I did some Fedora Mobile and HRF week last weekend - HRF (Human Readable Fedmsg: https://github.com/fedora-infra/hrf/) can now query datagrepper directly, which saves a slow http request from Fedora Mobile, which makes it show the fedmsg newsfeed about twice as fast :) 19:14:03 anshprat: I think the design list but I'm not sure -- pingou did hte heavy lifting. I just took on the role of bug fixer since he's on vacation and the election was ongoing. 19:14:18 abadger1999: get well soon 19:14:23 thanks 19:15:00 the whole datanommer/datagrepper stack is up to date in staging, and working 19:15:03 abadger1999: alright , guess will sign up for those as well.. what are the other elections we are planning to use nuancier for? 19:15:09 Ah -- one thing I noticed when working on nuncier hotfixes, we need to work on our app deployments a bit. 19:15:13 croberts: still working on something..nothing that I have pushed out.Java was not accepted so need to understand what was written and optimize those algorithms . will take at least 2 weeks before i can comment 19:15:17 before going into production with it i would like to have a plan of what to update on what machines in what order 19:15:38 ianweller: excellent. 19:15:48 i am fairly certain i can go like this: 19:16:13 busgateway01 update everything -> datagrepper01 update everything -> everything else update fedmsg 19:16:15 * abadger1999 looking for tickets 19:16:43 i just am not sure what order the update on busgateway01 should go in, in order to minimize downtime 19:16:44 ianweller: didn't we hit a bus -> irc issue this week or something? 19:16:59 lmacken: that was related to the fedmsg plugin not being loaded in supybot 19:17:16 i should probably get someone to give me admin access to supybot just in case that happens again 19:17:36 er, no, that was irc -> bus. 19:17:38 In particular, this ticket is probably an easyfix if an apprentice wants to take it on: https://fedorahosted.org/fedora-infrastructure/ticket/4046 -- HTTPS services need haproxy change 19:17:39 ah okay, so nothing mysterious 19:17:47 lmacken: we hit a bus -> irc issue that was quickly fixed by threebean yes 19:18:41 i intend to come up with a plan for datanommer/datagrepper upgrade in production, email the list today, and make it happen tomorrow. 19:18:50 (EOF) 19:19:19 ianweller: was that a packages-out-of-sync issue? 19:19:37 abadger1999: that was a legitimate bug that we found, that i'm looking for the diff for 19:19:38 or "new package, needed to resync the configs from ansible?" 19:19:40 I'd like to figure out how to do zero-downtime bus reboots. I'll bet we could do it with this feature: https://lwn.net/Articles/542629/ 19:19:41 okay. 19:19:56 https://github.com/fedora-infra/fedmsg/commit/c857d15f75736528c870a74a05e07181164f9012 found it 19:20:51 ah. 19:21:35 k 19:22:49 For ticket 4046 -- we just need our haproxy config modified so that it actualy checks all services to see if they're up or down. Right now, many aren't being checked so if one node goes down we get transient errors. 19:23:17 whenever haproxy sends a request to the down node. 19:24:01 and thus we spent our evenings and morning earlier this week :) 19:24:29 or something like that.. 19:24:37 anything else on applications? 19:24:53 I would need some help on askbot - rhbz 1010166 19:25:00 i hope i gave croberts a good answer 19:25:06 stuck on celery versions.. 19:25:29 frankieonuonga, it does make sense to me 19:25:42 anshprat, ok 19:25:46 .bug 1010166 19:25:50 smooge: Bug 1010166 askbot-0.7.49 is available - https://bugzilla.redhat.com/show_bug.cgi?id=1010166 19:26:32 I think I just fixed it for el6 - now need to check for rawhide.. 19:26:46 anshprat, very cool. 19:28:34 anshprat, kevin will be back on Monday I believe. He should be able to help on the items 19:28:50 I am going to move to the next topic. 19:28:53 smooge: alright 19:29:30 #topic Sysadmin status / discussion 19:29:41 Ok since Kevin is out.. it is mainly me. 19:29:51 We will be going into Beta freeze real soon 19:30:08 [Actually I need to make sure I didn't miss it already.] 19:30:16 how soon is real soon? :) 19:30:16 it's on tuesday 19:30:45 So next tuesday. Please make sure any changes to applications etc are done before Monday afternoon. 19:31:19 On tuesday 90% of changes will need to have two +1 from sysadmin-main people before being pushed to production. 19:31:57 I need to put out the beta freeze tickets tomorrow 19:32:27 lmacken, does that cause problems with your schedule? 19:32:41 just want to get any plans dealt with. 19:33:36 I may need some help getting a new blockerbugs version deployed (probably tomorrow) if the sudo issue I'm hitting is still around 19:33:52 tflink, I believe it is :/ 19:33:57 smooge: nope, should be fine 19:34:03 glibc is the problem with the sudo items 19:34:17 hmm need to report it agaisnt 6.5 and see if we can make it a blocker :) 19:34:18 frankieonuonga :) you did 19:34:19 ok, I thought that was just f19 nodes 19:34:32 oh sorry.. maybe different bug then 19:34:44 tflink, don't mind me. My coffee ran out 5 minutes ago 19:34:49 probably not, I don't pretend to know much about it 19:35:04 an offhand comment kevin made last week is about it :) 19:35:20 Oh then that might be F19 only then. 19:35:29 so I need to make it a different bug :) 19:35:33 Ok my items: 19:35:46 1) Cisco UCS. We have a blade server with 4 blades. 19:36:06 We should be able to make them buildvmhosts for dgilmore soon. 19:36:24 then next quarter I can order 4 more blades and work on another chassis 19:36:48 :) 19:36:50 Let me just say that Hardware as a Service is a completely weird concept if you don't know that is what you are working on 19:37:13 once I figured that out.. it became a lot clearer 19:37:17 I need to set aside one or two of the old blades to separate build channels 19:37:42 dgilmore, I don't know what is going to be the lifetime on the old blades. 19:37:51 will work with nirik on that after he gets back 19:38:07 smooge: well, we are going to need them for a bit 19:39:08 smooge: i feel you. i work for a cloud compute company building the platform and trust me..sometimes I dont even know what I am doing but it works 19:39:15 dgilmore, I will make sure that the warranty on it is extended one more year. 19:39:37 smooge: well if its not its not. 19:39:59 dgilmore, no it will be. IBM hardware senses when it has no warranty 19:40:06 down to the minute 19:40:18 :) 19:40:59 Third item for sysadmin items is it will be hardware inventory and maintenance time. We need to work out what hardware we own, what part of its lifetime it is in and what to replace it with and when 19:41:21 #topic Upcoming Tasks/Items 19:41:26 1) Beta time on Tuesday 19:41:33 2) Inventory and budget 19:41:42 just out of curiosity cant we spin of something where we are using commodity h/w 19:41:52 smooge: where is the inventory for h/w ? 19:42:09 (existing h/w) 19:42:14 anshprat: that is in the private repo 19:42:18 3) Re-enliving CSI documentation 19:42:22 puiterwijk: ok 19:43:09 smooge: CSI ? 19:43:27 anshprat: http://infrastructure.fedoraproject.org/csi/free-software-policy/en-US/html-single/ 19:43:36 "community services infrastructure standards" 19:43:37 Community Something Infrastructure. A bunch of documents that were written a while ago on how to run our infrastructure 19:43:43 (i'm not sure if that's the most recent copy) 19:43:57 ianweller, I expect it is 19:44:03 thanks! google gave me this.. https://fedorahosted.org/csi/ 19:44:18 yep that is upstream for them 19:44:23 ok.. 19:44:55 My plan is to have apprentices rewrite sections of them into asciidoc versus docbook and update to current practices. 19:45:50 While not glamorous it is one of the most important skills a sysadmin needs to have. Being able to document and explain what is going on. 19:45:58 Ok last itme 19:46:05 #topic Open Floor 19:46:21 any one have items for the open floor? 19:46:28 croberts shoule be able to guide other interns on csi I guess.. 19:47:21 that is a good idea. And something I need to write an email about. 19:47:23 Thank you 19:47:36 anshprat: csi 19:47:47 Ok I am going to close in 3 minutes if there isn't anything. 19:48:28 one unrelated thing - python-lamson needs to be dropped I guess due to license change.. 19:48:29 yes , we really need to encourage people to know more about other languages in genral. i realized when i tried to suggest something on java..it was evident most of the community was stuck to python. not a bad thing..but we really need to open minds up 19:48:34 where do I follow that up? 19:48:59 and I do not mean this in a negative way 19:49:01 ansphrat? I should guide interns? I am an intern to Fi-apprentice myself 19:49:15 anshprat, I would contact legal@fedoraproject.org 19:49:19 and the maintainer 19:49:41 croberts: you have experience with twiki and docs guidelines I guess, that's why I said.. 19:50:08 smooge: legal part is done, guess will check the maintainer, thanks 19:50:27 frankieonuonga, the main reason for our reluctance on other languages is that they all good and fine when there are people committed to staying at the project for long term (which usually means employed to work on it). 19:50:49 oh i see 19:50:57 When that doesn't happen we end up with zombie software which the next "I know X" person says is the worst code they have ever seen and want to replace with something new 19:51:11 * anshprat agrees with smooge 19:51:19 wow. ok then makes sense 19:51:29 incase i ever leave you are not left stranded 19:51:34 ok sounds good 19:51:48 I don't mind java if we have someone who is tasked with be the java person so that they can guide others in it. 19:52:08 i can take it off 19:52:12 not a problem 19:52:26 but as you said it might need support from someone inside 19:52:27 frankieonuonga, yeah.. we had a bunch of different little projects with special languages and people wanting us to deploy scala or ruby but not on rails 19:53:04 I don't see a problem in testing what we are working on if we can drive towards having some java people really help out. 19:53:37 But in any software there is always a social side that has to be solved as much as the technical side :) 19:53:37 I have one question 19:53:38 well i think this one will have to wait for kevin to get back cause he might think i am going behind his back 19:53:49 frankieonuonga, no problem 19:53:50 who is member of cloud team? 19:54:37 mirek_: dgilmore should be from what i recall . but i might be wrong . whats up ? 19:54:47 maybe i can help 19:54:53 mirek_: whats up? 19:55:01 I just have few question, some of them I sent to mailing list 19:55:17 dgilmore: is arm in fedora cloud? 19:55:20 mirek_: okay 19:55:23 mirek_: no 19:55:39 mirek_: not sure its really possible 19:55:39 dgilmore: and is possible to get arm as emulated on x86? 19:55:51 mirek_: need to look into it though 19:56:02 mirek_: not easily 19:56:11 dgilmore: and second question - how can i edit existing AMI or upload new one? 19:56:22 dgilmore: ok 19:56:22 mirek_: f20 is easier but still difficult 19:56:41 mirek_: an admin can upload new images 19:57:06 because I need to edit AMI used for builder for copr 19:57:33 ok I need to close out in 1 minute for the next meeting 19:57:43 mirek_: okay 19:57:44 mirek_, dgilmore could you work on #fedora-admin 19:57:49 ok thanks guys 19:57:57 yeah 19:57:57 Thanks for coming everyone. Do continue in #fedora-admin, #fedora-apps and #fedora-noc 19:57:57 thanks for coming by . 19:58:00 smooge: will do :) 19:58:01 went smooth 19:58:01 thanks all 19:58:08 #endmeeting