20:00:16 <mmcgrath> #startmeeting Infrastructure
20:00:16 <zodbot> Meeting started Thu Jul  8 20:00:16 2010 UTC.  The chair is mmcgrath. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:16 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:00:18 <mmcgrath> #topic Who's here?
20:00:20 * skvidal is here
20:00:23 * CodeBlock here
20:00:24 * mmcgrath here
20:00:28 * sijis is partially here
20:00:29 <LinuxCode> here
20:00:31 * nirik is lurking in the cheap seats.
20:00:36 * badone present
20:00:53 <mmcgrath> ianweller: you around?
20:01:05 <smooge> hwew
20:01:08 <mmcgrath> lets get started
20:01:12 <mmcgrath> #topic Meeting tickets
20:01:14 <mmcgrath> .ticket 2234
20:01:16 <zodbot> mmcgrath: #2234 (request for resources: datanommer) - Fedora Infrastructure - Trac - https://fedorahosted.org/fedora-infrastructure/ticket/2234
20:01:25 <mmcgrath> abadger1999: if ianweller and ricky are gone, you're the only other one listed here.
20:01:33 <dgilmore> hey
20:01:35 <mmcgrath> delivery / expiration date is set for the end of the month, what's the scoop?
20:01:39 <abadger1999> heh
20:01:50 <abadger1999> Dunno -- it's ianweller's baby.
20:02:10 <skvidal> where did ian go?
20:02:17 <mmcgrath> .any ianweller
20:02:17 <zodbot> mmcgrath: ianweller was last seen in #fedora-meeting 5 days, 21 hours, 20 minutes, and 45 seconds ago: *** ianweller has joined #fedora-meeting
20:02:18 <skvidal> I thought he was working from rdu
20:02:21 <LinuxCode> hes at ui now or ?
20:02:28 <mmcgrath> he might just be up and about at the moment.
20:02:28 <skvidal> uni shouldn't start until aug
20:02:29 <LinuxCode> uni
20:02:31 <mmcgrath> we can go back to it.
20:02:33 <LinuxCode> skvidal, true
20:02:53 <mmcgrath> #topic Infrastructure - Caching + expiration
20:03:02 <mmcgrath> I've started to setup more agressive caching and expiration schemes for the websites.
20:03:12 <mmcgrath> I tested against pkgdb today
20:03:14 <smooge> mmcgrath, how is that going
20:03:29 <mmcgrath> smooge: it's going ok, I'm trying to get the stuff done that I can control
20:03:37 <mmcgrath> there's other fixes we should probably do that are much trickier.
20:03:50 <mmcgrath> like removing the some 30 or so css downloads as part of pkgdb
20:04:09 <abadger1999> <nod>
20:04:30 <mmcgrath> I don't mean to pick on pkgdb, lots of our apps have this problem I just know it's under active development and toshio happens to be very easy to find :)
20:04:30 <abadger1999> At some point I need to look at compiling css and javascript into single files.
20:04:44 <mmcgrath> abadger1999: the other thing that's unclear to me is js downloads.
20:04:45 * mdomsch is here
20:04:48 <abadger1999> i've just haven't had time to look at optimizing :-(
20:04:56 <mmcgrath> something I *think* you're doing in pkgdb for icons, and something we're doing a lot of in community
20:05:06 <ianweller> whoops sorry guys
20:05:08 <ianweller> was distracted
20:05:14 <mmcgrath> ianweller: no worries, one sec
20:05:21 <mmcgrath> anywho, anyone have any questions or comments on it?
20:05:32 <smooge> what are the setting at?
20:05:36 <mmcgrath> I really only bring it up because we've had some issues in the past so keep your eyes open.
20:06:01 <mmcgrath> well for caching I'm now using the htcacheclean instead of tmpwatch, cache itself should be determined by the headers.
20:06:10 <mmcgrath> when no expiration headers are set, I'm setting them to 5 days.
20:06:42 <mmcgrath> some sitse, like yahoo have hilarious expiration headers
20:06:45 <mmcgrath> .headers http://l.yimg.com/a/i/mntl/ww/events/p.gif
20:06:46 <zodbot> mmcgrath: content-length: 43, accept-ranges: bytes, expires: Tue, 23 Jun 2020 19:54:01 GMT, server: YTS/1.17.23.1, last-modified: Wed, 13 Jan 2010 07:16:59 GMT, connection: close, cache-control: max-age=315360000, date: Sat, 26 Jun 2010 19:54:01 GMT, content-type: image/gif, age: 1037564
20:06:50 <mmcgrath> yeah, 2020.
20:07:29 <mmcgrath> anywho, moving on :)
20:07:33 <mmcgrath> #topic Meeting tickets
20:07:34 <smooge> wow, like they will be in business then
20:07:41 <mmcgrath> .ticket 2234
20:07:42 <zodbot> mmcgrath: #2234 (request for resources: datanommer) - Fedora Infrastructure - Trac - https://fedorahosted.org/fedora-infrastructure/ticket/2234
20:07:45 <mmcgrath> ianweller: ^^ what's the scoop?
20:07:46 <ianweller> yo.
20:07:52 <mmcgrath> delivery date on this is the end of the month
20:08:11 <ianweller> well, after talking with mmcgrath and spevack the scope/implementation of this project has changed entirely
20:08:22 <ianweller> so i'm thinking of closing out that RFR and just opening a new one, with a later date
20:08:35 <LinuxCode> ianweller, little mouse to huuge elephant ?
20:08:37 <ianweller> i do have (hacked-together) code for the new implementation up in the git repo
20:08:42 <ianweller> LinuxCode: more like the other way around
20:08:45 <LinuxCode> hehe k
20:09:05 <mmcgrath> ianweller: has max and company given you some pretty good requirements on what they're looking for?
20:09:17 <ianweller> mmcgrath: yeah.
20:09:44 <mmcgrath> can you give us a preview?
20:09:51 <ianweller> hang on a sec
20:09:57 * ianweller will delve through mailing list archives
20:10:32 <ianweller> ok well i guess i've never typed up the full implementation. so i'll talk about that
20:10:37 <ianweller> mmcgrath: or did you want to hear about what max is looking for
20:10:42 <mmcgrath> sure
20:11:20 <smooge> what max is looking for first, then your implementation
20:11:41 * ianweller is pulling up what max is looking for
20:11:45 <ianweller> it's... somewhere ;)
20:12:16 <ianweller> ok so here's what max wrote down for what he is looking for
20:12:27 <ianweller> 1) Place that enables people to articulate questions that could be answered.
20:12:40 <ianweller> 2) Fully automated answers to questions above:
20:12:43 <ianweller> - modular
20:12:49 <ianweller> - examples of it running against Fedora
20:12:58 <ianweller> - straightforward to run it against anything
20:13:09 <ianweller> 3) Docs/tutorials for anyone to work on #2
20:13:25 <ianweller> 4) Aggregate a collection of queries into a report (presentation layer).
20:13:55 <ianweller> (end of his wants)
20:14:08 <mmcgrath> 'mm kay
20:14:12 <ianweller> so 1) is currently happening now on [[Statistics 2.0]], although it's a bit stale
20:14:12 <mmcgrath> sounds massive.  Good luck with that :)
20:14:22 <ianweller> and 2) is what datanommer is going to attempt to do.
20:14:37 <ianweller> i need to do an actual writeup of how it is slated to work and post a blog post.
20:14:45 <ianweller> 2-4 is what datanommer will do, sorry
20:15:15 <ianweller> right now, the code will produce a graph that looks something like this: http://ianweller.fedorapeople.org/test.png
20:15:44 <ianweller> mmcgrath: the implementation is all in the code -- do you want me to talk about it now, or just blog it?
20:15:55 <ianweller> or put it in a new RFR? ;)
20:15:58 <mmcgrath> ianweller: ok, I think we have a general idea of what all you're working on.  I say blog about it.
20:16:05 <mmcgrath> we can discuss implementation on infra list as well if you like
20:16:06 <ianweller> ok
20:16:09 <ianweller> <nod>
20:16:12 <mmcgrath> anyone have any questions for the weller?
20:16:29 <ianweller> mmcgrath: should i just do a new RFR at this point, since it's *completely* changed?
20:16:34 <mmcgrath> yeah
20:16:38 <mmcgrath> probably best.
20:16:39 <ianweller> i think the only thing i kept from the old RFR is hte name
20:16:42 <ianweller> ok, will close
20:16:55 <smooge> no questions from me
20:17:03 <mmcgrath> alllrighty
20:17:13 <mmcgrath> #topic How to be a successful contributor
20:17:45 <mmcgrath> So I created this page recently in response to the high turnover rate we've been seeing - https://fedoraproject.org/wiki/How_to_be_a_successful_contributor
20:17:53 <mmcgrath> I sent it to FAB but I wanted to know what people here thought.
20:17:57 <mmcgrath> especially those of you that are newer.
20:18:10 <mmcgrath> so thoughts?
20:18:13 <mmcgrath> did anyone happen to see it yet?
20:18:22 <smooge> I am slowly making my way through it
20:18:35 <smooge> I am using it for a template for next weeks "How to Be a New Sysadmin"
20:18:59 <ianweller> mmcgrath: it's very good, and is fairly generic to the entire fedora project (and i think that was your goal?)
20:19:05 <mmcgrath> it was
20:19:15 <dgilmore> mmcgrath: i thought it was well written
20:19:25 <mmcgrath> dgilmore: that's because I had stickster proof it :-P
20:19:30 <ianweller> i'm going to try and find a place to link to it either from the main page or the join page
20:19:40 <sijis> i thought it was good.
20:19:42 <mmcgrath> I handed him a document that was full of numbers and a couple of shapes.  Lots of purple.  He converted it to english :)
20:20:03 <sijis> ianweller: i'm thinking the join page.
20:20:04 <stickster> mmcgrath: *snort
20:20:13 <LinuxCode> mmcgrath, great idea
20:20:19 <ianweller> sijis: i am too.
20:20:28 <LinuxCode> mmcgrath, the concerns I had I told you the other day
20:20:28 <mmcgrath> CodeBlock: so you're a relatively new voluteer, what are your thoughts?
20:20:29 <stickster> It was a great doc, something we've needed for a while.
20:20:36 <smooge> wait stickster can do that?
20:20:47 <CodeBlock> mmcgrath: I had not seen it yet, I'm reading it now :)
20:20:52 <smooge> do you work with pink too?
20:20:52 <ianweller> although i'm thinking we could stick it just before "follow these instructions" in the second paragraph of [[Fedora Project Wiki]]
20:20:53 <tobyh> I've had a brief read - thought that to some extent the bit about the commitment required may scare some people off, but then the project requires those who will stick around.
20:20:59 * badone is reading
20:21:12 <tobyh> I'd go with adding it the joining page also
20:21:17 <ianweller> although, maybe not the best idea to the main page.
20:21:22 <mmcgrath> tobyh: I'd honestly rather scare them off right now instead of in 6 weeks when they're in the middle of a project.
20:21:29 <stickster> smooge: All colors except puce, vermillion, and sea-foam green.
20:21:34 <mmcgrath> I'm sponsoring 6 people right now, it's been crazy hard.
20:21:39 <sijis> is ther an area about expectations?
20:21:51 <mmcgrath> sijis: that's pretty much what the whole thing is.
20:21:56 <tobyh> mmcgrath: in that case you've got your point across well. I'll get my coat!
20:22:00 <tobyh> just kidding
20:22:22 <abadger1999> mmcgrath: Something seems missing from the doc but I haven't put my finger on it yet.
20:22:31 <mmcgrath> a smiley face?
20:22:34 <abadger1999> :-)
20:22:48 <ianweller> that's an easy fix ;)
20:22:49 <mmcgrath> One of the thing that's bugged me lately is giving people access and having them up and vanish.
20:22:54 <mmcgrath> that has happened a _LOT_ lately.
20:23:00 <mmcgrath> I need to go through and prune the groups down.
20:23:19 <smooge> stickster, ah crap.. I did it in sea-foam. maybe I can make it 1975 avocado
20:23:20 <mmcgrath> it's almost as if people are showing up for the thrill of the hunt and then get bored when they realize it's really just work to be here and not some crazy sex orgy.
20:23:31 <mmcgrath> or technology orgy.  I don't know what they think we do :)
20:24:13 <CodeBlock> haha, great analogy
20:24:23 <abadger1999> I think there should be something about "Does it scratch an itch for you to work on this?"
20:24:25 <badone> it isn't?
20:24:32 <abadger1999> That's a good motivator for staying involved..
20:24:34 <rbergeron> i feel so deflated
20:25:01 * badone is sure it said that in the brochure
20:25:22 <mmcgrath> Do people generally think this is an area in Infrastructure we should be focusing a bit more on?  Or is it just a fact of life now?
20:25:23 <abadger1999> Maybe something about... find out if there's any low hanging fruit that you can work on -- things that are small but give immediate results.
20:25:30 <LinuxCode> lol @ crazy sex orgy
20:25:33 <stickster> abadger1999: Excellent point. I have an itch or two that will become my "20% time" in my next position. I'm working on them precisely because they matter to me.
20:25:46 <mmcgrath> abadger1999: it kind of mentions that people should expect to be getting some small tasks first.
20:26:01 <smooge> mmcgrath, I am hoping to have something written up better this week on how to sponsor/get through sysadmin adn how to leave
20:26:02 <badone> junior jobs?
20:26:15 <badone> like KDE?
20:26:18 <LinuxCode> mmcgrath, good idea
20:26:21 <mmcgrath> One problem in Infrastructure is we don't have too many low hanging fruit anymore.
20:26:24 <LinuxCode> just to work them in
20:26:24 <abadger1999> mmcgrath: <nod>  but I guess it's a question of who's finding the tasks.
20:26:33 <abadger1999> mmcgrath: I have plenty of coding low hanging fruit :-)
20:26:40 <LinuxCode> mmcgrath, maybe have some sort of advice centers ?
20:26:41 <mmcgrath> abadger1999: do we really though?
20:26:49 <LinuxCode> like a web person
20:26:53 <LinuxCode> a email person
20:26:58 <LinuxCode> a dns person
20:27:00 <LinuxCode> ?
20:27:02 <mmcgrath> because the last time I went through to find coding work of stuff that needs to be done I ran into our SA problem and some db / logic issues in smolt.
20:27:06 <smooge> mmcgrath, I don't know we seem to have lots of low-hanging fruit at times... but not sexy fruit. More like those under sized apples you get in an elementary cafeteria
20:27:23 <mmcgrath> smooge: like what?  what's a low hanging fruit?
20:27:28 <smooge> documentation
20:27:37 <jsmith> I wouldn't call that low hanging
20:27:38 <mmcgrath> that's not low hanging at all.
20:27:42 <smooge> oh.
20:27:46 <LinuxCode> smooge, that requires experience ?
20:27:47 <mmcgrath> because to do it you have to have intimate knowledge of the system.
20:27:53 <LinuxCode> yeh
20:28:01 <abadger1999> mmcgrath: So.. I guess we need to think of 1) What qualifies as low hanging fruit?  (For instance, our SA problems are to me... they need someone to pretty mechanicall port code).
20:28:12 <jsmith> It requires both intimate knowledge of the system, and the ability to describe in a clear and concise manner
20:28:15 <LinuxCode> couldnt larger task been broken up ?
20:28:16 <mmcgrath> the SA one's aren't to me because of how complex they are.
20:28:20 <LinuxCode> and then let teams form ?
20:28:23 <mmcgrath> it'd take several weeks just to understand the problem.
20:28:40 <smooge> I have always thought of it as the first thing you needed to do. Get to know a system well enough to write a book about it so you can take it over.
20:28:44 <mmcgrath> well to me anyway.  but I'm not a coder, perhaps I should have higher expectations of an actual coder.
20:29:22 <badone> smooge: then you won't have any contributors :)
20:29:33 <mmcgrath> abadger1999: to me, low hanging fruit is something that can be done with sysadmin-test or no access at all.
20:29:52 <smooge> ah
20:29:54 <abadger1999> mmcgrath: Ah... so you're thinking of access and I'm thinking of code knowledge.
20:30:01 <mmcgrath> smooge: the problem is no one wants to sit around and learn about a system with the promise of writing it all down.
20:30:08 <mmcgrath> abadger1999: well coding too.
20:30:08 <tobyh> low hanging fruit... awstats not working - run the update script to see if it fixes the prob, if not investigate the error.
20:30:15 <abadger1999> mmcgrath: Maybe we need to set up a "staging"like environment in publictest?
20:30:22 <tobyh> and report to some with access
20:30:34 <tobyh> as an example
20:30:34 <mmcgrath> abadger1999: the problem I've run into with recent code people is this.  "You got anything to work on?"  "Sure!  Get the fas code from here and get it running, then I'll assign some tickets to you"
20:30:42 <mmcgrath> that step alone, I have yet to have someone get back to me about.
20:30:44 <abadger1999> So that anyone can get access to hack on a packagedb instance/fas instance/etc without having to setup the db, etc.
20:30:54 <abadger1999> <nod>
20:31:02 <mmcgrath> If they can't even get it running.  I'm just not sure what to do from there.
20:31:44 <mdomsch> mmcgrath, yeah, i've burned some possible contributors because MM is somewhat difficult to get started with in an environment.  not as simple as 'yum install'
20:31:45 <abadger1999> mmcgrath: Well -- getting it running is a sysadmin task.
20:31:45 <mmcgrath> tobyh: yeah but that's sort of a one off thing, like.  Lets say we got that done.  What other low hanging fruit you got?
20:32:11 <mmcgrath> abadger1999: but I'd expect someone who's going to work on it could at least get it running from source.  Is that an unreasonable expectation?
20:32:18 <abadger1999> mmcgrath: So to be successful with developing our apps on your own machine, you have to grok both the sysadmin side and the coding side.
20:32:36 <LinuxCode> abadger1999, which makes it more demanding
20:32:45 <mmcgrath> but what good is a coder that can't get the system they're going to be coding on running?
20:32:58 <mmcgrath> I may have completely unreasonable expectations there.
20:33:16 <smooge> mmcgrath, well most coders are that stage.. thats why we have sysadmins :)
20:33:20 <abadger1999> mmcgrath: Depends -- I don't htink it's unreasonable but at the same time -- if we think it's a problem that a lot of people aren't successful at setting up a fas instance to hack on, (ie the people are smart enough to code on it but the seetup is a problem) then we should do something aboiut that.
20:33:46 <mmcgrath> the question is why didn't they get that far?
20:33:48 <mmcgrath> was it really that hard?
20:34:01 <smooge> I would say from my experience 80% of the coders aren't sysadmins and 80% of the sysadmins aren't coders
20:34:03 <mmcgrath> or was it they just didn't realize that volunteering means actual work and that means time
20:34:06 <abadger1999> mmcgrath: So those people who are devoting 10-30 hours a week on an app really should be able to set it up from a sysadmin side.
20:34:08 <sijis> is there any support docs that could help this person build fas locally?
20:34:21 <abadger1999> They're likely to be making decisions that affect the deisgn and how deployable the app is.
20:34:26 <mmcgrath> sijis: yeah the README that comes with FAS is a step by step setup process.
20:34:33 <mmcgrath> last I checked on it it worked just fine.
20:34:41 <abadger1999> So if it's painful for them, they have the ability to fix it so it's not so painful and then everyone benefits.
20:34:42 <mmcgrath> there might be an SA bug or two in the setup process now.
20:34:46 <abadger1999> But the 1-2 hour a weekers...
20:35:08 <abadger1999> They're just going to be fixing bugs in code, implementing peripheral features.
20:35:10 <mmcgrath> and that's something to bring up.  I don't want to downplay any volunteer status.  But can you be a volunteer in Fedora at 1 - 2 hours / week?
20:35:23 <mmcgrath> In Infrastructure if you volunteer for 1 hour a week, that's the meeting we're in right now.
20:35:29 <smooge> well before I got this lovely job.. my time I could spend on CentOS went something like this 20 hours a week for a couple of weeks and then 5 months of 1-2 hours a week
20:35:29 <mmcgrath> it's practically no time at all.
20:35:55 * mmcgrath wishes these were easier questions to answer :)
20:36:02 <abadger1999> For them spending five hours figuring out how to set up the code is several weeks investment which doesn't directly influence what they're primarily working on.
20:36:02 <smooge> or I could focus on Fedora for about the same.
20:36:26 <mmcgrath> abadger1999: well do we want to go forward offering to setup fas and company up for people on pt machines?
20:37:40 <abadger1999> mmcgrath: I could see the benefit to doing that, yes.
20:37:50 <smooge> mmcgrath, actually I would look at the following: Set it up once and train someone through it to get it done. so we could say to the people who just want to code... (or find that they can do 8 hours of coding but 2 hours of sysadmin kills them) "Hey tobyh can you make a pt fas box for sgallagh for us?"
20:37:53 <abadger1999> mmcgrath: But maybe we should do this frst --
20:38:01 <abadger1999> mmcgrath: Make a list of low hanging coding fruit.
20:38:37 <abadger1999> mmcgrath: If we have five bugs that people need fixed in pkgdb, say, that's five reasons that having that kind of setup in pt would be worthwhile.
20:38:42 <mmcgrath> we're in a tough spot right now because in about a 2 month span we lost Nigel, Jon and Ricky
20:38:50 <abadger1999> Yep.
20:38:59 <mmcgrath> we didn't really lose Jon but the hours are a bit off and we've certainly taken a hit from it.
20:39:39 <tobyh> mmcgrath: maybe as part of the joining process it should be more clear about what is deemed a suitable amount of time per week to be volunteering.  If 1-2 hours is not seen as worth while
20:39:48 <abadger1999> mmcgrath: Okay so here's another thought -- do we have anyone in sysadmin-test that we're ready to promote to something higher?
20:39:55 <mmcgrath> not really
20:39:59 <mmcgrath> at least not that's only in sysadmin-test
20:40:11 <abadger1999> mmcgrath: How do we change that?
20:40:27 <badone> maybe make setting up fas a "rite of passage"
20:40:30 <dgilmore> abadger1999: have more things in test that need to become production
20:40:35 <abadger1999> mmcgrath: I've got some coders in sysamdin-test, for instance, but when is it appropriate to move them up the ladder?
20:40:39 <tobyh> smooge: I agree, if you have a new starter who maybe lacks skills in some areas, train them up and then get them to train a few others afterwards once they no it inside out.
20:40:39 <mmcgrath> well, for one most people I sponsor go from sysadmin-test to production pretty quickly or they vanish.
20:40:42 <mdomsch> abadger1999, it's 2 parts: 1) someone able and with the right time; 2) someone we've seen be capable of that promotion.
20:40:44 <mmcgrath> like CodeBlock
20:40:56 <mmcgrath> he went from sysadmin-test to sysadmin-noc or web within a week.
20:40:57 <abadger1999> mmcgrath: So... mbacovsk?
20:41:02 <mmcgrath> he showed up every day, worked, etc.
20:41:04 <abadger1999> maploin?
20:41:15 <mmcgrath> what about them?
20:41:18 <LinuxCode> mmcgrath, I think confidence is an issue too
20:41:20 <CodeBlock> mmcgrath: :)
20:41:26 <abadger1999> Should they go to sysadmin-web now?
20:41:29 <LinuxCode> you dont want to do something wrong, as a professional
20:41:32 <mmcgrath> why would they?
20:41:45 <mmcgrath> correct me if I'm wrong but they're doing coding for pkgdb right?
20:41:51 <abadger1999> Correct.
20:42:02 <mmcgrath> AFAIK it's not implied that it's an upstream project means downstream sysadmin-web
20:42:15 <abadger1999> Also correct.
20:42:35 <mmcgrath> now, if they wanted to take a step into sysadmin web and had a specific project in mind.  Yeah, their work on pkgdb means they're great candidates for it.
20:42:36 <abadger1999> But... until someone else working on the pkgdb has syadmin-web, I'm a big bottleneck.
20:42:59 <mmcgrath> what's the bottleneck though?  the release process
20:43:09 <mmcgrath> because it still goes from source -> rpm -> deploy
20:43:55 <abadger1999> 1) release.  2) testing (having a separate group for sysadmin-staging migt work there) 3) making fixes at whatever time in the morning because something is horribly broken.
20:44:19 <mmcgrath> eh, I hesitate to make an outside group for staging because of security reasons.
20:44:29 <abadger1999> mmcgrath: Right.
20:44:31 <mmcgrath> but a third 'official' test environment might be feasible.
20:45:10 <abadger1999> mmcgrath: Which means A) We need to promote people into sysadmin-web or we need to make changes to how we improt data and passwords into staging.
20:45:31 <abadger1999> s/or/or b)/
20:45:54 <mmcgrath> is that a real problem?
20:45:56 <abadger1999> A third test env would help with 2)
20:46:01 <mmcgrath> it sounds like we just need a proper sample data set?
20:46:09 <abadger1999> wouldn't help with 3 though.
20:46:11 <smooge> mmcgrath, the big boy places I worked at in SNL had : Devel, Testing, Staging, Production environments
20:46:16 <abadger1999> proper sample set of what?
20:46:20 <abadger1999> ah.
20:46:26 <abadger1999> I understand now.
20:46:49 <mmcgrath> It just seems wrong that faster development means access to production data.
20:47:18 <abadger1999> If we're worried about the security of data in the staging env, then we have to munge the data that we put into it.  If we are concenred enough that we're  creating a third testing env, then we need to miunge the data going into that.
20:47:20 <mmcgrath> if the problem is the sysadmin side of things, a more official publictest environment might be in order.
20:48:30 <abadger1999> We don't necessarily need production data but the closer it is to production data, the more useful it becomes.
20:48:50 <mmcgrath> <nod>
20:49:04 <mmcgrath> well it sounds like this should come up for discussion at our next meeting for sure.
20:49:16 <mmcgrath> for now anyone have a problem with us moving on (we've only got 10 minutes or so left for the meeting.
20:49:29 <abadger1999> <nod> -- note I may not be here (going to be at FUDCon Santiago)
20:49:50 <mmcgrath> that's ok, this is more of a long term goal anyway, not something we need to fix tomorrow.
20:49:51 <abadger1999> (Week after, FISL in brasil)
20:50:10 <mmcgrath> #topic Open Floor.
20:50:16 <mmcgrath> anyone have anything they'd like to discuss?
20:50:27 <CodeBlock> Nagios 3 migration status
20:50:42 <mmcgrath> CodeBlock: what's blocking right now?
20:50:55 <CodeBlock> well
20:51:00 <CodeBlock> tobyh: heard anything back from peter yet?
20:51:48 <LinuxCode> mmcgrath, we will be switching to rhel6 right ?
20:51:59 <CodeBlock> or not. - well mmcgrath you mentioned rhel 6 on the list - is rhel 6's default nagios in the repos going to be 3?
20:51:59 <mmcgrath> we can
20:52:01 <LinuxCode> rather soonish after its out ?
20:52:12 <LinuxCode> CodeBlock, should be
20:52:17 <mmcgrath> CodeBlock: it should be but someone's got to do the work, that's where the maintainer comes in
20:52:22 <LinuxCode> epel is bumping everything as far as I understood
20:52:34 * mdomsch will be out all next week again
20:52:42 <dgilmore> LinuxCode: everything in epel-6 was branched from F-12 or rawhide
20:52:59 <LinuxCode> dgilmore, k we did have 3 in there didnt we ?
20:52:59 <dgilmore> mdomsch: toshio and i will be out also and the week after
20:53:06 <LinuxCode> 3.25 is in f13
20:53:08 <dgilmore> LinuxCode: i believe so
20:53:13 <CodeBlock> alright - suppose we just hold on it then. I'm still in the process of moving nagios to a proper puppet module, but that's...an ongoing/doing in my spare time thing
20:53:21 <LinuxCode> sorry 3.2.1
20:53:47 <mmcgrath> mdomsch: <nod> have fun
20:54:02 <CodeBlock> Hosting request automation - I've been working on the script with baddog, weaving out bugs, and talking with mmcgrath and -websites for ideas.... well
20:54:07 <LinuxCode> CodeBlock, I will be doing some testing too with 3 vs 2
20:54:09 * CodeBlock just pastes what he wrote up before the meeting
20:54:14 <CodeBlock> I've been working with baddog quite a bit lately on the hosted automation script, trying to get that going again. I've found a problem in that the 3-4 tickets that have been submitted since we changed template, 100% of the users altered the questions to shorten them (which makes the script not work for that question.)
20:54:15 <CodeBlock> After talking with some people, the proposed fix is to create an html form that generates the data in a form ready to be pasted into the ticket body. The user will be instructed to not change the ticket body once it has been generated. This is ongoing, though as I showed mmcgrath two days ago, I do have something worth having -websites test out and see what they think.
20:54:17 <LinuxCode> I can let ya know how that goes
20:54:37 <mmcgrath> CodeBlock: <nod> that's a problem.
20:54:47 <mmcgrath> CodeBlock: how far are you from deploying your form thing?
20:55:16 <CodeBlock> mmcgrath: I will talk with -websites about it tonight or tomorrow and see what they say, to my knowledge it's ready to go, as far as I'm concerned
20:55:38 <CodeBlock> I'm assuming we're just going to integrate it into /web/new
20:55:38 <mmcgrath> CodeBlock: you could probably add that as a replacement to the /new page (I assume that's what you had in mind)
20:55:42 <mmcgrath> yeah
20:55:50 <sijis> CodeBlock: i can help you test that form.
20:56:15 <CodeBlock> sijis: alright, it's just some html/JS stuff, nothing too complex ;)
20:56:23 <mdomsch> completely different topic - cleanup of crap left behind by puppet...
20:56:42 <mmcgrath> CodeBlock: yeah lets meet up after the meeting if you have other questions
20:56:45 <mmcgrath> mdomsch: such as?
20:57:05 <mdomsch> when deleting stanzas in puppet init.pp files, rather than just deleting the stanza, consider adding "ensure => 'absent'", and if it's a directory, 'force => true'
20:57:20 <mdomsch> such as deleting a cronjob from a module, or deleting a directory you had previously created in the module
20:57:48 <mmcgrath> <nod> that should be in the SOP
20:57:51 <mdomsch> if you simply delete a cron {} stanza, from init.pp, puppet won't delete it from the servers it's deployed to
20:57:58 <mmcgrath> is there some specific things you're worried about?
20:58:02 <mmcgrath> ah, so a training issue you mean.
20:58:12 * mmcgrath probably could put another puppet training session together
20:58:23 <mdomsch> nah, I just ran into it cleaning up mirrormanager stuff today, so thought I'd throw that out for FYI
20:58:24 <LinuxCode> mdomsch, thats a bane of puppet
20:58:32 * nirik points mmcgrath to #fedora-classroom. That would be a great class. ;)
20:58:57 <mmcgrath> <nod>
20:59:14 <mmcgrath> Ok, we've got one minute left.  anyone have anything earth shattering?
20:59:50 * CodeBlock thinks the puppet clsas is a good idea
20:59:53 <CodeBlock> *class
21:00:02 <mmcgrath> allllrighty.
21:00:02 * badone too
21:00:07 <mmcgrath> #endmeeting