/srv/irclogs.ubuntu.com/2016/07/19/#juju.txt

lazyPowerx58 - yeah you're finding some rough edges with charmhelpers vs reactive01:02
lazyPowerreally sorry about that too :( We've been moving kind of fast, and its left a trail in the tooling that we haven't been able to circle back and get either updated, or deprecated, just yet.01:03
=== frankban|afk is now known as frankban
kjackalhello juju world07:51
gnuoyhey stub, I see that https://code.launchpad.net/~stub/charms/trusty/nrpe/py3/+merge/300153 has a prereq branch which isn't proposed yet. Is that an oversight? I assume that the normal workflow here is to merge the prereq branch then this one.08:47
neiljerramMorning10:35
neiljerramDoes anyone happen to know if I can control the names of the machines that Juju creates when I ask it to deploy a bundle?10:36
neiljerramIn other words, I'd prefer something more meaningful than 'juju-466b6192-d900-4e7a-8fa4-e3935e5c25f5-machine-0' or 'juju-26a5f7-0'10:37
=== rogpeppe1 is now known as rogpeppe
bryan_attgnuoy: ping13:39
bryan_attany charm experts here that can help with a layer question?13:40
marcoceppi bryan_att ask away13:41
bryan_attmarcoceppi: I have an issue where a value is not getting set in one of the config templates. It seems that the "layer"  is not passing the right value, or the value name changed. How do I debug this?13:42
marcoceppibryan_att: is this a layer.yaml option or a config.yaml option?13:43
bryan_atthttps://github.com/blsaws/charm-congress - see https://github.com/blsaws/charm-congress/blob/master/src/templates/liberty/congress.conf13:43
bryan_attthe value "connection = {{ shared_db.uri }}" is not getting set13:43
bryan_attthis comes from I think includes: ['layer:openstack-api'] in https://github.com/blsaws/charm-congress/blob/master/src/layer.yaml13:44
bryan_attBut I don't know how to check this "layer" to see if the variable has changed somehow.13:44
bryan_attI further need to document this dependency (if the root of my current build/deploy issue) if it can result in CI/CD failure - so I need to know what is the root issue here and how to debug such (potential) issues with layers13:46
marcoceppibryan_att: ah, so shared_db.uri comes from the mysql-shared interface layer13:46
marcoceppibryan_att: the openstack-api layer pulls in interface:mysql-shared: https://github.com/openstack/charm-layer-openstack-api/blob/master/layer.yaml13:48
marcoceppibryan_att: I don't see a .uri key for that library though13:49
bryan_attmarcoceppi: sorry, I don't know what a .uri key is, and can't see from the file what is being referenced (or at least where to find the real code)13:50
marcoceppibryan_att: sorry, is shared_db.uri a key in your template or a key in the openstack-api template?13:51
bryan_attI don13:51
bryan_attI don't know13:51
jamespagebryan_att, hey - can you join #openstack-charms please13:51
bryan_attsure13:52
=== Guest73831 is now known as med_
=== scuttle|afk is now known as scuttlemonkey
acovrigI get this error running conjure-up openstack: cannot find network interface "lxdbr0": route ip+net: no such network interfaceERROR invalid config: route ip+net: no such network interface15:29
stokachuacovrig: try running lxd init15:29
acovrigstokachu: still get the error...15:31
acovrigand I don’t see lxdbr0 in ifconfig15:31
stokachuacovrig: ugh, such a pain point for users15:31
acovrigstokachu: what if I install docker first?15:32
stokachuacovrig: that has nothing to do with lxd15:32
stokachuacovrig: sec15:33
acovrigyea, but wasn’t sure if it would configure the if, but I guess it uses docker0 so I guess not15:33
stokachuacovrig: what does, sudo ifup lxdbr0 do?15:33
acovrigstokachu: Unknown interface lxdbr015:33
stokachuacovrig: that bridge is socket activated so something has to touch it to turn on15:33
stokachustub: one sec15:34
stokachuacovrig: one sec15:34
stokachustub: disregard15:34
stokachuacovrig: try running 'lxc list' as a normal user15:35
stokachusee if that activates the bridge15:35
acovrigstokachu: it doesn’t show anyting in the list, would a reboot help?15:36
lazyPowerstokachu - wasn't there a bit where if you had lxc prior, you had to dpkg-reconfigure to get the proper bridge?15:36
stokachuacovrig: now what does `ifconfig` show15:36
stokachulazyPower: yea we do that for them but sometimes it doesn't active the new bridge :(15:36
lazyPower:|15:36
acovrigstokachu: 2 physical, lo, and openstack015:36
stokachuacovrig: :((((15:37
* stokachu sad15:37
stokachuacovrig: ok give me another second15:37
stokachuacovrig: for kicks try running `conjure-up openstack` again15:38
stokachuacovrig: what about `ifconfig -a`15:38
acovrigstokachu: same as ifconfig; should I pick OpenStack or OpenStack with Nova-LXD?15:39
stokachuacovrig: if youre doing it on a single machine go with novalxd15:39
=== scuttlemonkey is now known as scuttle|afk
acovrigstokachu: intend to run on a cluster of 3 machines (once I get an update to fix a bug in maas)15:39
=== scuttle|afk is now known as scuttlemonkey
stokachuacovrig: the other openstack spell requires i think 4-5 machines15:40
acovrigstokachu: same error… (cannot find network interface "lxdbr0")15:40
stokachuacovrig: bah15:41
* acovrig sudo reboot pending15:41
stokachuacovrig: yea try that :|15:41
acovrigstokachu: ifconfig shows lxdbr0, running conjure now15:47
stokachuacovrig: yay15:47
stokachuacovrig: it shouldn't be this difficult to start that bridge15:47
stokachuwe may try to handle it better in conjure-up15:48
stokachuacovrig: are you running the pre-release?15:48
acovriglol, yea; also, when installing it added me to the lxd group, but I had to logout and back in to get it to accept that, I wonder if I should have rebooted then15:48
acovrigstokachu: I’m running ubuntu 16.0415:48
stokachuacovrig: yea or log completely out15:48
stokachuacovrig: what does dpkg -l conjure-up show15:49
=== natefinch is now known as nfinch-baby-chic
=== nfinch-baby-chic is now known as nfinch-babychick
acovrigstokachu: 0.1.2 and juju 2.0~beta7-0ubuntu1.16.04.115:49
=== nfinch-babychick is now known as nfinchbabychicks
acovrigstill bootstrapping15:49
stokachuacovrig: ok, do you want to test the latest stuff?15:49
stokachuacovrig: if you're curious https://www.irccloud.com/pastebin/91OmlH1S/15:50
acovrigstokachu: I presume that last line should be “conjure-up openstack” instead of “conjure-up openstack|”15:53
stokachuyea sorry15:53
acovrigI may look at it myself, I’m currenlty setting this up at work to compare an openstack cluster w/an windows hyper-v cluster so I may set it up at the servers at my house to tinker w/it.15:55
acovrigstokachu: this looks promising, it’s deploying 17 services and is ~1/2 done15:56
stokachuacovrig: cool, should be good to go15:56
acovrigstokachu: yup, now if I can get a bug fixed in maas, I’d scale this out across 3 machines lol15:56
stokachu:D15:57
stokachuwhat bug in maas?15:57
acovrigstokachu: https://bugs.launchpad.net/maas/+bug/160412815:57
mupBug #1604128: [2.0RC2] Unable to add a public SSH Key due to lp1604147 <cdoqa-blocker> <verification-done> <MAAS:Fix Committed by allenap> <MAAS 2.0:Fix Committed> <MAAS trunk:Fix15:57
mupCommitted by allenap> <maas (Ubuntu):Fix Released> <maas (Ubuntu Xenial):Fix Released> <maas (Ubuntu Yakkety):Fix Released> <https://launchpad.net/bugs/1604128>15:57
stokachuah ok rc215:58
stokachushouldnt be long for that15:58
acovrigstokachu: yea, I’m talking to someone on #maas and they said it’s fixed in -updates, I apt upgrade’d and got 2.0.0~rc2+bzr5156-0ubuntu2~16.04.2, but it still isn’t letting me put a key in...15:58
stokachuhmm15:59
acovrigyea, and I think they’re AFK (~1hr)16:00
=== rcj` is now known as rcj
stokachuacovrig: is this just putting the key in the UI16:00
stokachucopy and pasting?16:00
=== rcj is now known as Guest10260
=== Guest10260 is now known as rcj
acovrigstokachu: yup16:03
stokachuacovrig: that is weird16:03
acovriga differnet install (VM) (I think it was the same version) accepted this key16:03
acovriglol, was about to ask how long I should expect this to take (conjure-up openstack), but then I looked at top…44% iowait, 12% idle… *sigh* slow hardware...16:55
geethaHi, My IBM WAS Base charm needs 2 fixpack packages, when fixpack already installed and user wants to upgarde using latest fixpack packages, I have written code like this : http://paste.ubuntu.com/20056926/. But when I run juju attach command for latest fixpacks, upgrade_charm hook will be triggered and comparing checksum values for existing fixpack packages and new fixpack packages. Here 2nd packges is not getting updated with the lates16:58
geethawhen I try to verify checksum , for second package it's showing same checksum value for old and new fixpack package17:00
stokachuacovrig: yea it takes a toll on the hardware for a single install17:01
stokachuacovrig: we usually recommend at least 16G ram for that spell17:01
geethacould any one please suggest me how to check whether the packages are old or new in multiple package scenario17:01
acovrigstokachu: lol, yea I have exactly 1617:01
stokachuacovrig: ok yea, should take about 45 minutes17:01
acovrigstokachu: yea, it’s been going for ~1.25hr, it’s done with 9/18, 8 installing apt packages, 1 incomplete relations: identity17:04
stokachuacovrig: ok17:06
stokachuacovrig: if it takes that long i would kill it and try the latest from our ppa17:06
stokachuupdate juju to beta12 as well17:06
stokachutons of fixes17:07
acovrigstokachu: “our ppa” being “ppa:juju/devel” and “ppa:conjure-up/next”?17:07
stokachuacovrig: yea17:07
=== frankban is now known as frankban|afk
acovrigstokachu: apt purge juju conjure-up && apt install juju conjure-up or just upgrade? (I.E. will there be any broken leftovers if I ^c the current conjure process)17:08
stokachushouldn't be any leftovers if you ^c17:09
stokachuacovrig: your first one would work17:09
stokachuor apt dist-upgrade17:09
acovrigstokachu: apt upgrade? wouldn’t dist-upgrade upgrade my distribution of ubuntu?17:10
x58Any guides on publishing stuff to the charm store?17:10
magicaltrouthttps://insights.ubuntu.com/2016/04/16/charm-charm-tools-2-0/17:11
magicaltroutthe docks currently are a bit lousy because it changes17:11
magicaltroutd17:12
magicaltroutyou need to charm push17:12
magicaltroutcharm publish17:12
magicaltroutcharm grant17:12
magicaltroutthe help on the cli is pretty useful17:12
x58I'm logged into the charmstore and there is a button that says "create new" but clicking it does all of nothing.17:15
magicaltroutdunno what that is17:16
magicaltrouton the command line17:16
magicaltroutcharm login17:16
magicaltroutthen you can interact with the CS from the cli17:16
magicaltroutand push your stuff17:16
x58Found some docs, will follow, will report back if I fail.17:16
magicaltroutyou don't need to setup anything on the web17:16
x58https://jujucharms.com/docs/devel/authors-charm-store17:19
x58This tells me I need to login to the charm store in my browser.17:19
magicaltroutyeah17:19
magicaltroutbut then login on the CLI17:19
magicaltroutoh yeah they were the docs I was grepping for17:20
x58Published my first charm =)17:29
x58https://jujucharms.com/u/bertjwregeer/staticroutes17:29
x58I read the docs, and the README could be a rst file, yet it is not being rendered correctly :-(17:30
magicaltroutwoop17:30
magicaltroutchange rst to md17:30
magicaltroutignore the lies!17:30
x58I'm a python dev... rst is my life.17:31
magicaltroutsad times!17:31
magicaltroutbetter than this crazy go stuff :)17:32
x58configuration also loses my formatting (specifically newlines that make it readeable and understandable :-()17:33
magicaltroutokay 2 things a) not sure https://jujucharms.com/u/spicule/pentahodataintegration/trusty/8 thats one of mine, the formatting is acceptable17:34
magicaltroutb) you've not granted public read writes17:34
magicaltroutso I can't see it and have no clue :)17:34
x58How do i grant it?17:35
x58I can see it, clearly I am the most important person ;-)17:35
magicaltroutcharm grant cs:~spicule/wily/dcos-master --acl=read everyone17:36
x58magicaltrout: done.17:36
x58charm grant cs:~bertjwregeer/staticroutes everyone (from the docs)17:37
magicaltroutyeah17:38
magicaltroutnot sure how rst works17:38
magicaltroutbut for blocks17:38
magicaltroutjust indent 417:38
magicaltroutand headings use #, ##17:38
magicaltroutusual markdown stuff17:38
x58Yeah, that's markdown... rst has headings, so ------- underneath a line is a top-level, heading <h1>, ~~~~~~ underneath a ------- heading is a <h2>17:39
x58.. code:: is valid rst.17:39
magicaltroutyeah, but rst has never rendered for me on the charms store17:39
x58See rendering here: https://gitlab.com/bertjwregeer/juju_staticroutes17:39
x58Ah17:39
magicaltroutso unless you're magic, it'll be broke :)17:39
x58That's a different story. I followed docs. Docs said md or rst.17:39
magicaltroutyeah i have no clue17:39
magicaltroutthe samepl charm is .rst file17:40
magicaltroutbut it never formats when its deployed17:40
=== nfinchbabychicks is now known as natefinch
x58Where can I file bugs against the charmstore?17:40
magicaltrouthttps://github.com/juju/charmstore17:40
magicaltroutprobably17:40
lazyPowerhttps://github.com/CanonicalLtd/jujucharms.com/issues17:40
magicaltrouti lied17:41
lazyPowerx58 - link is in the footer of the page :)17:41
magicaltroutthere be no bugs17:41
magicaltroutits written by professionals17:41
jrwrenwhere do docs say RST? that is incorrect and has never been supported AFAIK.17:41
lazyPowerjrwren - RST was a supported format back in the juju 1.x days, in the old store. Kapil made heavy use of RST17:42
magicaltroutjrwren: the charm template creates you a RST17:42
magicaltroutor it did last time i tried17:42
* jrwren is disappointed.17:42
x58Well I just filed two bugs in the wrong place :P17:46
magicaltroutlol17:47
x58https://jujucharms.com/docs/devel/authors-charm-policy#readme.md17:47
x58that is where .rst is mentioned.17:47
magicaltroutlies all round17:49
magicaltroutchatting with Bill on Thursday lazyPower, got most of my PDI charm done.. and then I get sidetracked writing a maven plugign for snappy *boom*17:50
magicaltroutI really need to stop doing new stuff17:50
lazyPowerWhats the fun in that?17:50
magicaltrouthopefully still get DC/OS polished off this week17:50
x58People tend to complain when people DON'T read docs, and now I read the docs and it's wrong...17:51
x58That just hurts.17:51
lazyPowerx58 we apologize, and will take steps to address the bug17:51
=== cmagina is now known as cmagina_away
magicaltroutblame it on evilnickveitch17:51
* magicaltrout runs for the hills17:51
magicaltroutlazyPower: well there is no fun in that, but apparantly no one had thought of adding snapcraft to existing build chains17:52
magicaltroutwhich seems rather weird17:52
magicaltroutso i'm taking some steps to address that17:52
lazyPower:D17:52
lazyPowerthats phenomenal17:52
magicaltroutwell if you have a build setup like a big maven multi module project or something, why would you replace that with a snapcraft maven weird thing?17:53
magicaltroutespecially if you've got CI, reporting etc17:53
magicaltroutand I'm sure its not just me in java world17:53
magicaltroutbut it seems reasonable to build artifacts then programatically create a snap definition to create your image post build17:53
magicaltrouta bit like with juju where you might want to replace specific bits of infrastructure but not replace all your nicely crafted puppet scripts or  cookbooks17:54
x58lazyPower: I18:02
x58lazyPower: lol. I've filed bugs :-)18:02
lazyPowerx58 thank you, sincerely :) i appreciate the effort18:02
x58lazyPower: I maintain WebOb (python request/response library) and work heavily with the Pylons project on Pyramid. I understand that docs can be out of date or wrong. No-one likes writing docs.18:02
magicaltroutsee the funny thing is if i said a sentence like that i'd be taking the piss18:02
magicaltroutlazyPower is being honest though I think18:03
x58magicaltrout: haha18:03
magicaltroutx58: where you based?18:04
lazyPoweri am :) I have no idea how many users come by and dont file bugs when we're just plain wrong on something18:05
lazyPowerto see someone take a couple minutes and file a bug, which makes us better makes me happy18:05
magicaltroutsee there's the american sincerity shining through18:06
magicaltroutyou don't get that in the uk ;)18:06
magicaltroutwe might think it, we just don't articulate it in that manner ;)18:06
lazyPowermagicaltrout - i appreciate you too sir18:07
lazyPower:D18:07
magicaltroutlol18:07
magicaltrout*cringe*18:07
x58magicaltrout: Denver, CO. Currently got a mattrae sitting next to me working on our production deployment.18:23
x58magicaltrout: I am from The Netherlands though. So we like to tell people how it is, Dutch Directness.18:23
magicaltrouthehe18:24
magicaltroutmaybe but the dutch isn't as scary as german18:24
magicaltrouts/isn't/aren't18:24
x58Germans are scary because their language always sounds angry. Even when expressing love.18:24
x58We Dutch have toned down that "anger" in our language.18:24
magicaltroutyeah i'm sat in a bar now, and every 10 minutes the woman walks past and shouts "1 more?" at me18:25
magicaltrouti feel like i should say yes everytime18:25
x58magicaltrout: Where do you work? Found an email address on one of your charms and the bare naked domain takes me to an ESXi welcome screen :P18:26
magicaltroutdoes that still happen?18:27
x58analytical-labs.com18:27
magicaltroutI don't even know who owns that box :)18:27
magicaltrouti'm part self employed data guy based in London, part NASA JPL data scientist based in Pasadena and part time Juju/Snappy hacker18:28
magicaltroutdid apachecon in denver a few years ago, jesus christ, i'm never drinking in denver again18:31
x58magicaltrout: Welcome to alttitude.18:35
x58Learn to drink up here, and you can drink like a fish at sea level.18:35
magicaltroutlol18:35
magicaltrouti'm sure18:35
magicaltroutthe only place i've switched on a satnav in a hotel basement parking lot and it warns me I should get lower18:36
x58I don't drink anymore, but in Denver I could put away a fair bit. Ended up in Vegas for DEF CON... let's just say I spent a lot of money.18:36
magicaltrouthehe18:36
petevgcory_fu: I'm chewing on a thing with the interaction between puppet and juju in the Bigtop charms.19:14
petevgThe idea behind puppet is that it should be safe to run arbitrarily on a running system. It will reset config files that you've manually modified, but that's expected -- you're supposed to change the puppet config if you want stuff to change.19:15
petevgSo the question is: I have a config value that I may want to set at deploy, or may want to change while my server is running.19:16
petevgWhat's the best practice for handling that situation?19:17
petevgBasically, I want to wind up in a situation where I can do juju upgrade, and have the value stay set.19:18
petevgcory_fu: you mentioned having a juju action manually update the juju config on the host machine. Do I have to worry about updating the config locally, too? Or will juju respect the config on the host if it's re-run?19:19
cory_fupetevg: I did not mention having an action update config.  I mentioned having a config option update the config.  ;)  But yes, the puppet scripts ought not wipe out any data but c0s suggested that they might (for HDFS, at least).  If they do, the most correct (if perhaps not the most expedient) solution would be to fix the Bigtop puppet scripts19:21
cory_fuFor the case of Kakfa, it seems unlikely that the scripts will wipe out any data, but it's worth verifying19:21
petevgcory_fu: Got it. So the charming way to do things would be to update your config and re-run juju deploy, but there may be broken behavior in Bigtop. I'll try to test for it. Thx.19:23
cory_fupetevg: I assume you mean re-run puppet apply?19:23
petevgcory_fu: no, I didn't. But that's why I typed it -- I wanted to make sure that I wasn't walking away with wrong things to do :-)19:24
petevgWhat I'm intending to do is to have a layer option that sets the "bind" address for kafka.19:25
petevgIf I want to change that, I can update the layer option, but then what command do I want to run to tell juju to update the config on the host?19:25
petevg*sets should be "specifies"19:26
cory_fupetevg: No, no.  Not a layer option, a config option.  Those can be changed after deployment (and on a per-deployment basis), while layer options cannot19:26
petevgcory_fu: got it. So I do juju deploy --config <some config>.yaml kafka19:27
petevg... and then later, I change my mind about the config. What do I run to update it?19:27
stokachucory_fu: just making sure you saw some PR's i created https://github.com/juju-solutions/bundle-apache-processing-mapreduce/pulls19:28
cory_fupetevg: juju set-config listener=0.0.0.019:29
cory_fupetevg: juju set-config kafka listener=0.0.0.019:29
petevgcory_fu: Ah, cool. And then we call the config-changed handler, which I believe runs puppet, and all is good.19:29
petevgCool. Thank you :-)19:29
cory_fustokachu: Yeah, sorry, I saw it.  It seems fine to me, but I'm still a bit unclear about the drive behind conjure.  Also, we're deprecating that bundle in favor of hadoop-processing, which lives in Apache Bigtop19:30
stokachucory_fu: if you run the instructions in the PR it'll give you a better idea19:31
stokachuis apache bigtop a different repo?19:32
stokachucory_fu: im guessing this is it https://github.com/juju-solutions/bigtop/tree/master/bigtop-packages/src/charm19:33
stokachucory_fu: and what hoops do i have to jump through to make contributions?19:33
petevgstokachu: juju-solutions/bigtop our fork apache/bigtop. You can make and push branches off of it, but you want to make your pull requests against apache/bigtop. (github will do that by default).19:35
petevgAs for hoops, there are a few more :-)19:35
magicaltroutwill you get sucked up into apache/bigtop without an ICLA etc signed?19:35
petevgNo.19:36
stokachui just care about the charms, is opening a PR against that repo enough?19:36
petevgHang on. Typing up an explanation :-)19:36
petevgBasically, the Bigtop folks will be happiest if you first file a ticket in their Jira: https://issues.apache.org/jira/browse/BIGTOP/19:37
petevgThen when you create your PR, you want it to have a title exactly in the following format:19:37
petevgBIGTOP-<ticket number> <Ticket Title>19:38
petevg... where <ticket number> is the four digit (so far) ticket number, and the title is the same string as you put in the title of your ticket.19:38
petevgThat way, the PR will get slurped up automatically into the Jira issue, and people will be able to find it easily.19:39
stokachuok19:39
magicaltroutjust caring about charms makes the apache software foundation sad.....19:39
magicaltrout;)19:39
petevgThey actually seem to be pretty happy about charms ... so long as we play nicely with their Jira automagic github integration thing :-)19:40
cory_fustokachu: Also, even though new dev is focused on Bigtop, the vanilla Apache bundle isn't going to go away, so we can accept the PR into that in the meantime19:41
magicaltroutpetevg: i can speak as an insider its pretty important else you just end up with a bunch of unclosable pull requests19:41
magicaltroutand its a right pain in the balls19:41
stokachuso just to clarify, a new contributor will need to first learn charm development with the reactive layer, then try to figure out where the charms live under different projects and learn their submission process?19:42
stokachuobviously the charmstore doesn't show this information19:42
magicaltroutwell there is talk of making charms live with their parent projects19:42
magicaltroutwith the new push mech19:42
petevgstokachu: when you put it that way ... :-/19:42
magicaltroutdoubt that flies with cory_fu19:42
stokachuright, where do you go to find all this out?19:42
magicaltroutbut you know, technically the charms could live in the BigTop19:43
stokachuright now im going to 3 different places to figure out how to make a contribution to a charm19:43
magicaltroutwell if you found the source, and they live with their projects, you're in the right place?19:43
stokachuhttps://jujucharms.com/apache-processing-mapreduce/19:43
stokachuthats not the source of bigtop which seems to be where development is heading, right?19:43
magicaltroutyeah, it is true, but its also legacy19:43
magicaltroutthis stuff will iron itself out over time19:44
stokachuhow do i know that19:44
magicaltroutwell you don't19:44
stokachu:\19:44
magicaltroutbut19:44
magicaltroutsince the new charm push/charm publish stuff is available now19:44
magicaltroutit makes charms in their respective projects much easier19:44
stokachujust b/c you dont have to push to launchpad first before it gets in the charmstore doesn't make discovery easier19:45
stokachufor people who want to contribute19:45
petevgstokachu: I'm definitely paying attention and thinking about it ...19:45
magicaltroutno, but if you go to a charm, it will give you a like to the source (doesn't it?)19:45
stokachuhttps://jujucharms.com/apache-processing-mapreduce/19:45
stokachuis that the same source as bigtop??19:45
petevgPutting stuff in the bigtop repo is meant to centralize the big data charms and make them easier to find.19:45
magicaltroutoooh19:45
petevgBut you are having to deal with two separate communities, which is a little tricky.19:45
magicaltroutjujucharms directs me to LP19:46
stokachuso you have 2 big barries, getting people to learn charm layers, then having them go through a potential new process everytime they want to submit something19:46
stokachubarriers*19:46
magicaltroutokay thats a bit shit19:46
magicaltroutthe charm layer contains a repo path19:46
magicaltrout(i may just have not filled mine in)19:46
magicaltroutit should direct you to the source19:47
magicaltroutand if you manage the source of the charm with the source of the project then thats fine19:47
stokachupetevg: a start would be to update the README in the toplevel dir about how to contribute a charm19:50
petevgstokachu: That's an excellent idea.19:50
petevgI'm also going to badger the big data team about updating our metadata so that there's always a link to click to view source from the charm store.19:51
stokachupetevg: ok thanks19:51
petevgNo problem. Thank you for the feedback!19:51
stokachupetevg: the readme could be as simple as what you told me here on irc19:52
stokachucory_fu: ok, give that PR a test run and let me know if you have questions19:52
stokachucory_fu: once you run it once it'll give you a good idea what conjure-up does19:52
magicaltrouton a more useful note... i ask my colleagues if my diet of beer and hotel peanuts will kill me19:55
magicaltrouthttp://www.ncbi.nlm.nih.gov/pubmed/1211958019:55
magicaltrouti got sent that as a response19:55
petevgit sounds like the paper says eat as many peanuts as you want: you won't even get tired of them. I approve of these findings :-)20:00
magicaltrouthehe20:01
magicaltrouti'm working on an authentic conclusion20:01
lazyPower"The effects of hotel peanuts and beer on magicaltrout over a decade, as computed by apache-realtime-streaming"20:04
lazyPower"and visualized with zepplin"20:04
magicaltrouti think it would likely show a sharp decline in something20:05
magicaltroutproductivity probably20:05
magicaltroutthat said, i've been sat here doing nothing whilst waiting on infra for about 3 hours20:06
magicaltroutluckily20:06
magicaltrouti'm paid by the hour20:06
lazyPoweror, you consistently reach the ballmer peak and you gain superhuman coding powers for 4 hours a day, and manage to charm all the things, thus rendering the rest of us puny mortals insignificant in your charming wake20:06
lazyPowerclearly i've thought about this...20:06
magicaltroutwell... yesterday i tested my pdi <-> hdfs tweaks to the charm20:06
magicaltroutand they worked first time20:07
lazyPowerbanking on unnatural phenomenon20:07
magicaltroutso maybe you have a point20:07
lazyPowersee?20:07
lazyPowermay the lesson be:  feed magicaltrout more beer20:07
lazyPowerand hotel peanuts*20:07
magicaltroutoh i can vouch for the beer part20:07
magicaltroutI did all my coding at uni whilst drunk and that was one of the few modules i passed with a 1st20:07
magicaltroutand lets face it, i work a lot for NASA in the evenings.....20:08
lazyPoweranecdotal evidence is still evidence right?20:08
magicaltroutindeed20:08
magicaltroutof the highest order20:08
=== cmagina_away is now known as cmagina

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!