/srv/irclogs.ubuntu.com/2014/05/16/#juju.txt

tvansteenburghbac: you around?00:08
rick_h_tvansteenburgh: what's up? Anything I can help with?00:11
tvansteenburghnah, not really. i just discovered that he broke something in charm-tools :P00:13
rick_h_doh00:13
rick_h_:)00:13
tvansteenburghbut i'm just gonna fix it since he's not around00:13
tvansteenburghno worries00:13
rick_h_cool thanks00:13
=== revagomes is now known as mGraf
=== mGraf is now known as revagomes
=== revagomes is now known as mGraf
=== mGraf is now known as revagomes
=== revagomes is now known as mGraf
=== sebas5384 is now known as _sEBAs_
=== mGraf is now known as revagomes
=== revagomes is now known as xauz
=== xauz is now known as revagomes
=== _sEBAs_ is now known as sebas5384
=== sebas5384 is now known as _sEBAs_
=== _sEBAs_ is now known as sebas5384
=== CyberJacob|Away is now known as CyberJacob
=== vladk|offline is now known as vladk
=== vladk is now known as vladk|offline
=== CyberJacob is now known as CyberJacob|Away
_mup_Bug #1320135 was filed: No way to configure the archive mirror <pyjuju:New> <https://launchpad.net/bugs/1320135>08:21
=== ev_ is now known as ev
=== vladk|offline is now known as vladk
zchanderAnyone around who can help me a little with a local devel version of a charm which isn’t picked up by juju (??)09:16
davecheneyzchander: i can (try to ) help09:36
zchanderdavecheney: The problem I got, is that any changes I make to my (devel) local config.yaml isn’t picked up when I deploy the charm (using the command juju deploy --repository=/home/madmin/charms/devel local:owncloud --to=6)09:55
zchanderI just got a reply through the list. Seems the problem was/is with the versions I got in my repo (both named owncloud in the metadata.yaml). I am testing this now10:02
zchanderIt seems that was the problem….10:02
zchanderBut the test/deploy is still running10:02
marcoceppizchander: that is probably the problem. Juju should throw an ambiguous error when more than one version of the charm is loaded via metadata.yaml11:29
zchanderGot it working….. The reason I mentioned before (two different versions of the same charm, though different named (folder))11:30
zchanderAlso got the Ceph relation working again, including MySQL ;)11:30
* zchander is going to get some teachers happy11:30
marcoceppi\o/11:36
zchandermarcoceppi: Is any information available to format e.g. metadata.yaml and config.yaml? I am trying to tidy the docs…12:06
marcoceppizchander: what do you mean to format?12:06
zchandermarcoceppi: I want to add some newlines, but it won’t render properly….12:27
marcoceppizchander: can you paste what you've got?12:34
zchanderhttp://paste.ubuntu.com/7472808/12:36
zchanderAlso for the config.yaml, I would like to make proper formatted descriptions12:37
lazyPowerhmm... seems a bit over the top being placed in the description like that.12:41
zchanderlazyPower: I know…. ;) But it is for my own testing (and this way I get to know the proper way to create the documentation)12:46
lazyPowerzchander: something like that is better served in the README. The description is more for the front facing display in the charm store.12:46
zchanderI’ll put it in the README.12:54
zchanderWhat is a charm with a proper metadata I can refer to?12:54
lazyPowerzchander: Postgres13:03
lazyPowerits got a really well formed charm structure, metadata, etc.13:04
zchanderI’ll have a look at it, tonight13:07
jcastrolazyPower, aws-specific things are ok in charms, we prefer they be labelled clearly13:17
jcastroso like "aws-loadbalancer" for example would be ok13:18
lazyPowerjcastro: thing is, its a voilation of the tenants of the charm store. so its not "ok" in the charm store's current guidelines.13:18
jcastrowhat we don't want is something that claims to be generic to have cloud specific bits, so like "haproxy"13:18
=== ted is now known as tedg
lazyPowerand this is very much cloud specific to aws, openstack - as it won't work on azure / maas13:18
jcastroit's a should13:18
jcastronot a must13:18
lazyPowerhmmm13:19
jcastrocharms "should make use of apparmor" too, but we don't reject if they don't13:19
lazyPowermarcoceppi: i followed your lead here, i need backup. am I being too picky?13:19
jcastrowe should certainly clear it up13:19
lazyPowerjcastro: i'm not opposed to the idea that i'm making a mistake, but i'd like clarification on if i'm nacking a charm that would otherwise be included but lower ratings because of it.13:20
jcastroyeah13:21
marcoceppilazyPower: tldr?13:21
lazyPowermarcoceppi: charm uses aws/openstack metadata ip hard coded in teh charm13:21
jcastromarcoceppi, cloud specific charm features13:21
lazyPoweri nacked it13:21
lazyPoweris it ok to ack with lowr rating, or nack it and say "fix it"?13:21
jcastrohttps://bugs.launchpad.net/charms/+bug/125963013:22
_mup_Bug #1259630: add storage subordinate charm <landscape> <Juju Charms Collection:Incomplete by davidpbritton> <https://launchpad.net/bugs/1259630>13:22
jcastroI only noticed because I saw Kapil's comment scroll by13:22
lazyPowerjcastro: the problem I have with acking provider specific charms is its not really apparent that this charm would only work on those 2 substrates. so some guy deploys this on his private maas cloud, and assumes this setup will work13:23
marcoceppilazyPower: well, that fails charm proof13:23
lazyPowermarcoceppi: beg pardon? all i get is I fails stop/start hook13:23
jcastrolazyPower, right, so in this case I think it's a naming issue13:24
lazyPowerwhich si safe to ignore, but it's also got this hoks.d thing13:24
lazyPowerer storageprovider.d13:24
jcastroso like, this is really "s3-storage", not "storage"13:24
* marcoceppi remembers why he punted on this review now13:24
marcoceppijcastro: not even, more like aws|openstack-block-storage13:25
lazyPowerjcastro: ^13:25
jcastrosure13:25
lazyPoweri was just typing that out.13:25
jcastroeither way, it's not generic "storage" at this time13:25
marcoceppijcastro: right, the only way to do that is in core13:25
marcoceppiand that's, as Kapil mentioned, months away13:25
marcoceppilazyPower: well, charm testing will reveal that this doesn't work on X Y and Z cloud13:25
jcastroright13:25
marcoceppiif that was hooked up to the charm store13:25
jcastroI think if we describe it properly, people will realize that13:26
marcoceppiif the readme doesn't CLEARLY call out that it only works on X clouds, that's the only issue13:26
jcastroright13:26
jcastroGeneric storage charm subordinate.  Intended to aid in making charms13:26
jcastro    easier to interface with external storage solutions without having13:26
jcastro    to speak and understand each type.  Presents a single mount point on13:26
jcastro    the unit, and communicates that back to your service through the data13:26
jcastro    relation.13:26
marcoceppiI think "cloud-storage" might be a more apt name13:26
marcoceppias maas and local aren't a real cloud, though that does cut azure out13:26
jcastroright, with clear limitations described in the description and readme13:26
marcoceppistorage, I agree, is too all encompasing, unless then plan on adding azure, joyent, etc in future releases of the charm13:27
jcastroyep13:27
marcoceppiactually, maybe storage is okay13:27
* marcoceppi reviews the charm13:27
jcastroif you read that description I just pasted it makes it sound all encompassing13:27
marcoceppijcastro: right, so it has this storagebroker.d directory in hooks, that allows you to select different broker methods13:28
jcastroshould be something like "This subordinate charm uses the AWS metadata server, which currently is supported by X, Y, and Z. Local and A,B,C are currently not support"13:28
marcoceppithere's a block-storage-broker, which is the aws/openstaack stuff. then local and nfs13:28
marcoceppiwhich would work in a cloud agnostic environment13:29
jcastrowhat does it default to?13:29
marcoceppiI'm inclined to recind some of what I said earlier, this is an interesting attempt at making a plugin based charm13:29
marcoceppilocal13:29
* marcoceppi goes to find what the heck that means13:30
lazyPowerwell thing is13:30
jcastrohey so out of the box it'll work on every substrate?13:30
lazyPowerthe storage broker is whats supposed to be doing the communication with the aws storage services13:30
marcoceppiasciishrug.tiff13:30
lazyPowerteh rest of that data comes in over the wire13:30
lazyPoweri'm wondering if the storage sub didn't get the update13:30
marcoceppilazyPower jcastro I'd nack this soley on the fact they don't seem to document each of the providers13:30
marcoceppia charm this monumental in undertaking needs more than 53 lines of readme13:31
marcoceppiI'm 5 mins in to poking and I still can't figure out what's going on13:31
jcastromarcoceppi, I'm not forming an opinion on the charm itself, doing a normal review of that is fine13:31
jcastroI think we should just clarify what we mean by cloud-specific features in policy13:31
marcoceppiI thought this was a nack or ack13:31
marcoceppijcastro: ah13:31
marcoceppiwell, we currently don't have any cloud-specific charms in the store13:31
marcoceppiall of Kapil's AWS specific stuff is still personal namespaced13:32
jcastrosure13:32
marcoceppiso there's not precedence for this yet13:32
jcastroso like, I think they should be fine, as long as it's obvious13:32
marcoceppiand we'd be potentially making one with this charm13:32
jcastroso if I do "aws-blueshift-workload-thing", that should be fine13:32
marcoceppiI'm inclined to agree, esp. with testing being the true revealer of "does this work on X substrate"13:32
jcastroright13:32
marcoceppi<cloud>-name works great when it's for one cloud13:33
marcoceppithis is potentially n clouds where n is < substrates13:33
jcastrook, I'll propose to list.13:33
marcoceppijcastro: yeah, lets move the discussion there13:33
jcastrothe policy bullet is somewhat ambiguous13:33
lazyPoweri put it in the review13:34
jcastrolazyPower, I'll also post the new trusty vagrant juju boxes13:34
lazyPowerteh verbatim bullet13:34
jcastrobut let's stick to precise for today13:34
lazyPowerjcastro: did those get pushed?13:34
lazyPowerawesome!13:34
jcastroyeah yesterday13:35
lazyPoweri have the precise box up right now13:35
lazyPoweri'm going to release a github repo with the workflow items in it for people to clone and quickstart13:35
jcastrolazyPower, link to my mailing list post for your review13:35
lazyPowerack13:35
lazyPoweroh nice you mailed the list about it yesterday. hi5 jcastro13:38
sebas5384good morning :)13:43
jcastrohmm13:43
lazyPowerMorning sebas538413:43
jcastromarcoceppi, any idea why https://code.launchpad.net/~jose/charms/precise/owncloud/port-change+repo+ssl-support/+merge/215527 is not in the review queue?13:43
sebas5384ben is working in the vagrant box for juju with trusty13:44
lazyPowerjcastro: its listed as needs fixing by mbruzek13:44
marcoceppijcastro: it's not assigned to charmers13:44
lazyPowerif it needs to be in the rev q, it needs the "request a review" button clicked13:44
sebas5384someone knows if there is a repository of the vagrantfile?13:44
jcastromarcoceppi, oh I see, charmers is merely subscribed13:44
jcastrothat kind of sucks doesn't it?13:44
marcoceppijcastro: mbruzek gave the wrong canned message for the re-review13:44
sebas5384because I have one, and i would like to help :)13:44
marcoceppijcastro: yes13:44
jcastrosebas5384, you want to talk to lazyPower13:45
sebas5384nice!13:45
jcastrosebas5384, I was just about to post the address to the trusty boxes13:45
lazyPowersebas5384: the vagrantfile is very barebones. the vagrant image itself has a cloudinit package that sets up the environment.13:45
mbruzekmarcoceppi, Did I?  we talking about the owncloud charm?13:45
jcastroyeah owncloud ssl13:45
mbruzekmarcoceppi, I see now, sorry about that13:46
mbruzekmarcoceppi, So if a MP has tests and they don't pass it is OK to ack it?13:47
jcastrowe should shove them in juju/vagrant on github or something13:47
sebas5384lazyPower: today the box use an externals python scripts13:47
sebas5384having to install bzr and cloning some scripts13:47
marcoceppimbruzek: only if the tests failed before the mp13:48
jcastrombruzek, he updated the branches after from what I can see13:48
mbruzekack13:48
jcastrono idea if it passes tests now though13:48
jcastrobecause marco hasn't fixed the world yet.13:48
jcastrocome on marco, I want test results on the internet!13:48
sebas5384lazyPower: so that means there is code from other places being used into the box, and that difficult the collaboration :)13:49
lazyPowersebas5384: you're hitting me at literally the first corner i've turned with this process. up until now, the vagrant building process has lived witho ur cpc team13:49
sebas5384lazyPower: but hey, it's just an ideia13:49
lazyPowersebas5384: so i dont have an answer for you yet13:49
lazyPoweri'm gonna need you to calm down :P13:49
lazyPowerCURB YOUR ENTHUSIASM SEBAS!!13:49
lazyPowerand i'll follow up with you when i have a better picture of whats where and how i can expose this to community contributors13:50
marcoceppisebas5384: DON'T LISTEN TO LAZYPOWER, UNCURB THE ENTHUSIASM13:50
marcoceppi:D13:50
lazyPowermarcoceppi: i'll push you :|13:50
sebas5384lazyPower: i don't understand what you mean, i'm not hiting you at all hehe13:50
jcastrojust push it somewhere dude13:50
jcastroactually13:50
jcastrolet's do this13:50
lazyPoweri think you guys are missing what i'm saying13:50
jcastrolet's do that after the charm school13:50
lazyPowerand a) i'm already working on a github repo to go WITH the charm school today13:51
sebas5384lazyPower: calm down :) take a big breath :)13:51
lazyPowerb) the vagrant file doesn't control much at present, just which box to use13:51
jcastrosebas5384, http://cloud-images.ubuntu.com/vagrant/trusty/current/13:52
sebas5384jcastro: nice!13:52
jcastrombruzek, tedg is being a squeaky wheel wrt to ssl support, so if you've got time to review that that would be <313:53
* tedg squeaks13:53
mbruzekon it.13:53
tedgmbruzek, Thank you!13:54
lazyPowermbruzek: i rereleased tests, its close but still doesnt pass when being run through the test harness.13:54
lazyPowerand they have not been merged yet afaik13:54
mbruzeklazyPower, They were your tests?13:55
lazyPowermbruzek: according to the bzr history yes - and they passed at one point.13:56
lazyPowerwhich is why i started refactoring, because they do not pass now13:58
lazyPowereither the test infrastructure has changed, or we encountered a 'this shouldnt work but does' situation.13:58
jcastroIncoming! https://github.com/juju/docs/pull/10214:27
lazyPowerjcastro: 1 comment, otherwise LGTM14:29
jcastrofixed and pushed14:29
mbruzekjose are you out there14:30
marcoceppilazyPower: sorry man, sniped that merge14:30
=== vladk is now known as vladk|offline
roadmryou sniper, you14:39
dpb1lazyPower: hey, updated the storage readme.  thanks for the review, let me know what you think.14:44
lazyPowerdpb1: pulling it up now, thanks for the quick turn around14:45
dpb1lazyPower: np! write back on the review, if you have more feedback.  I'll be AFK for today.14:48
lazyPowerdpb1: ack. Should be good though. The hooks looked good when i looked through it, the big thing was the conversation that started with the metadata url14:48
lazyPowerdpb1: perfect.14:59
=== vladk|offline is now known as vladk
=== vladk is now known as vladk|offline
=== tedg is now known as ted
=== Ursinha is now known as Ursinha-afk
=== Ursinha-afk is now known as Ursinha
=== roadmr is now known as roadmr_afk
sebas5384someone haves a guide of how to use ansible into a juju charm?17:18
sebas5384i sow that I need to download some tools for the charm to work17:19
sebas5384https://github.com/absoludity/charm-bootstrap-ansible17:19
whitsebas5384, I used this charm as an example w/ ansible17:20
sebas5384whit: didn't know! what charm?17:20
whitwoops -> http://bazaar.launchpad.net/~michael.nelson/charms/precise/elasticsearch/trunk/view/head:/hooks/hooks.py17:21
whitsebas5384, charmhelpers grabs ansible and installs it on the remote machine17:21
whitala  charmhelpers.contrib.ansible.install_ansible_support17:22
sebas5384whit: exactly!17:22
sebas5384so i have to install charm-tools before17:22
marcoceppisebas5384: charm-helpers needs to be embeded in the charm17:23
sebas5384whit: elasticsearch, sweet!17:23
marcoceppihelpers and tools are two distinctly different things17:23
sebas5384gotcha marcoceppi, because helpers is things that helps the charm, and tools are for the charmer17:24
sebas5384or something like that hehe17:25
marcoceppisebas5384: exactly that17:25
marcoceppi:D17:25
sebas5384:D17:25
sebas5384i'm starting to develop the drupal charm in ansible :)17:25
whitah nice17:25
sebas5384yeah! I loved ansible17:26
sebas5384and that would be to help me to prove a concept17:26
sebas5384then start develop the charm factory :)17:26
whitsebas5384, ansible great for this sort of stuff because it's fairly conscise and makes idempotency easy17:27
sebas5384whit: exactly!!17:27
sebas5384and reusable role are great to extend a charm17:28
sebas5384but something I was thinking about17:28
sebas5384ops, im back :P17:47
sebas5384let say i have a drupal charm, and for now attends what the project needs17:47
sebas5384but17:47
sebas5384now we have a new library dependency, and other things like that17:47
sebas5384so now, i have to "clone" the charm and called "nameOfTheProject"17:48
sebas5384but thats not nice, because now i have duplicated code, etc...17:49
sebas5384and now i have to maintain 2 codes17:49
sebas5384so, thats why i'm going forward to the ansible, so then, i can just update the ansible role17:50
sebas5384marcoceppi: what do you say, whats your advice to that case?17:52
sebas5384when i need to "extend"  a charm, to attend the project needs17:52
sebas5384:)17:52
marcoceppisebas5384: you can use a subordinate charm17:53
sebas5384marcoceppi: so i would have a subordinate charm called with the name of the project, something like that?17:55
marcoceppisebas5384: well you can have a "drupal-project" subordinate, that would just install the themes, plugins, files, etc in the right place, create it's own apache configuration, etc17:55
sebas5384hummm17:56
sebas5384marcoceppi: interesting17:56
sebas5384but I have to deploy it --to the same machine, or subordinated charms executes all into the related charm?17:57
sebas5384https://juju.ubuntu.com/docs/authors-subordinate-services.html reading...17:58
marcoceppisebas5384: subordinates don't use --to, they only exist to co-exist on a service17:59
sebas5384oohhh18:00
sebas5384marcoceppi: gotcha18:00
sebas5384marcoceppi: thanks! I will look at it :)18:01
=== CyberJacob|Away is now known as CyberJacob
=== roadmr_afk is now known as roadmr
=== vladk|offline is now known as vladk
cff"Juju Charm School Vagrant Workflow" starting in 10 minutes https://www.youtube.com/watch?v=qLNPn2rQynM18:49
=== cff is now known as CodePulsar
jcastrojose, ping18:51
=== revagomes is now known as mGraf
=== mGraf is now known as revagomes
sebas5384yeahh!! \o/18:56
jcastro3 minute warning!18:57
jcastrohttps://www.youtube.com/watch?v=qLNPn2rQynM18:57
marcoceppiCHARM SCHOOL! TUBULAR!19:03
josejcastro what's up19:04
jcastronm I figured it out!19:05
josenp19:05
cory_fuDid the video stop for anyone else?19:05
tvansteenburghno19:05
marcoceppicory_fu: not i19:05
cory_funm, back19:05
sysdrumseems fine for me19:05
tvansteenburghi called youtube and asked them to stop it for cory_fu only19:05
cory_fuAlso, is there any way in ubuntu to keep a video full screen on one monitor while typing in the second monitor?19:05
CodePulsar:-)19:06
cory_futvansteenburgh: Jerk.  :)19:06
CodePulsarcory_fu: yes, there should be a way19:06
josesorry I couldnt help, I'm still trapped at university19:06
marcoceppilazyPower: it'd be cool if vagrant init just did that for you, re box_url19:07
sysdrumis it working like docker?19:09
marcoceppisysdrum: similar, vagrant uses different backends (VirtualBox, kvm, vmware) so it's not limited to just linux since docker seems to be only LXC19:10
jcastrohttps://github.com/chuckbutler/juju-vagrant-charmschool19:11
sebas5384yeah i'm waiting, default: Warning: Connection timeout. Retrying...19:11
tvansteenburghjcastro: question: does it take this long to boot the vm everytime I do 'vagrant up' or just the first time it's provisioned?19:12
sebas5384tvansteenburgh: no, thats only the first time19:13
sebas5384tvansteenburgh: after that you can use just 'vagrant suspend' for example19:14
sebas5384but, here, gives me a timeout exceeded to boot19:14
sebas5384so im going to try again19:14
jcastro mine took a bit Retrying with the remote connection19:16
jcastrobut eventually continued19:16
marcoceppilazyPower: mediawiki is spelled wrong19:17
sysdrumQUESTION: Would I be able to manage it from a remote windows clients if I spin it up on another windows host?19:20
jcastrohttp://www.vagrantup.com/blog/feature-preview-vagrant-1-5-share.html19:29
sysdrumDo I have the ability to bridge the network? under VBox?19:30
marcoceppi\o/ thanks jcastro and lazyPower!19:33
=== roadmr is now known as roadmr_afk
sysdrumwas about to ask that.19:36
sebas538_sweet!! jcastro and lazyPower :)19:38
sebas538_jcastro: i remember that before these new vbox they where failing to connect after halting the vbox19:41
sebas538_so i assume this was fixed?19:41
=== sebas538_ is now known as sebas5384
sebas5384thanks!! \o/19:42
sysdrumThanks for answering my questions.19:43
lazyPowersebas5384: yeah that's fixed last i checked.19:43
lazyPowersebas5384: if you have any further issues with it make sure you open a bug against it or ping me and i'll do the due dilligence19:43
sebas5384lazyPower: ohh great! nice work men :)19:43
sebas5384*man19:43
lazyPowero/19:44
sebas5384thanks!19:44
tvansteenburghlazyPower, jacstro: good stuff, thanks guys!19:46
tvansteenburghjcastro even19:46
jcastroomg19:46
jcastrohey guys19:46
jcastrohttp://itty-bitty-cat-4232.vagrantshare.com/19:46
jcastro`vagrant share` is totally awesome19:47
jcastroI did that, logged into vagrantcloud via the CLI19:47
jcastroand then it generated that URL19:47
sebas5384yeah! was introduced in the new version19:48
sebas5384its awesome!19:49
lazyPowersebas5384: So wrt your questions this morning about the Vagrantfile and what not20:15
lazyPowerdid that github repository give you what you were looking for? or was there more to that statement that we haven't covered?20:15
josembruzek: lazypower is working on the tests20:24
mbruzekjose yeah I talked with him on that.20:25
josecool :)20:25
josealso, I meant to set 443, 443 is https20:25
lazyPowerjose: what i've pushed is where the tests are.20:25
lazyPoweri dont have any additional cycles to devote to those atm20:25
joselazyPower: so tests are done for now?20:26
lazyPowerjose: whats there is what i got mang. :| they work when being run direct but fail when run via charm test20:26
mbruzekjose, lazyPower:  Since the tests did not run before the MP I can't nack it for that.20:26
josembruzek: so we need my branch merged and subsequently Charles' branch20:26
mbruzekThe owncloud needs a small fix that I found, I did not rejet on the tests.20:26
joseit's stacked on mine afaik20:27
mbruzekjose, did you fix the 80 problem?20:27
joseI am branching to fix it now20:27
joseliterally *just* turned on my PC :)20:27
lazyPowermbruzek: jose - https://code.launchpad.net/~lazypower/charms/precise/owncloud/refactor_amulet_tests20:28
lazyPowerjose: i can propose this for merging into your branch and you can pick up teh torch from there20:28
lazyPowersound good?20:28
joselazyPower: sounds good to me, thank you20:28
lazyPowernp - sorry i dont have a 100% golden package for you, my week kind of blew up, and next week is looking just as hectic20:29
joseit's good :)20:29
lazyPowerhttps://code.launchpad.net/~lazypower/charms/precise/owncloud/refactor_amulet_tests/+merge/21990320:30
joseack, thanks!20:32
=== roadmr_afk is now known as roadmr
=== vladk is now known as vladk|offline
=== CyberJacob is now known as CyberJacob|Away

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