[00:52] <_mup_> Bug #996857 was filed: Local environment doesn't require opening ports < https://launchpad.net/bugs/996857 > [01:30] I'm trying to make a charm that does some deployment testing, i.e., install from a source, deb, or similar. How should I get the bundle over to the server in the charm? The bundle is local to me when I'm running juju deploy. [01:50] dpb_: what do you mean by the bundle? [02:07] Anyone available for a quick juju deploy question? [02:09] I'm trying to use a local repository instead of the juju store, but I get this: [02:09] http://pastebin.ubuntu.com/977122/ [02:19] jsalisbury: juju -v deploy --repository=/home/jsalisbury/juju/charms local:kernel-build [02:19] jsalisbury: and you must put the charms in a directory named with the ubuntu version [02:19] like precise or oneiric [02:20] avoine, awesome, that was it! I was missing the local:kernel-build [02:20] avoine, this may need to be updated: [02:20] https://juju.ubuntu.com/docs/drafts/charm-namespaces.html?highlight=juju%20deploy%20local [02:21] I doesn't have the local: [02:23] marcoceppi: Like a private build a piece of software. a private deb. tar.gz file, etc. [02:27] marcoceppi: what I'm really wanting is to be able to test deployment from my checkout on my workstation. I can do that with a cloud instance, but I wanted to do it through juju. [02:28] All I need is a way to scp a file over there as part of charm. [02:32] One additional question. Am I using the --config option correctly: [02:32] http://pastebin.ubuntu.com/977144/ [02:34] jsalisbury: cat out your config file if you can. [02:34] jsalisbury: I can use that option, so it may be something with your file [02:36] dpb_, This is it: [02:36] name: kernel-build [02:36] summary: [02:36] description: [02:36]

[02:36] options: [02:36] release: [02:36] default:precise [02:36] arch: [02:36] default:amd64 [02:36] ~ [02:36] This is my first charm, so I may have set this up wrong ;-) [02:37] jsalisbury: oh I see what you are doing... --config is for specifying configuration options *into* your charm [02:37] jsalisbury: you probably don't need it if you are just getting going [02:37] dpb_, yes, I need to supply the release and arch [02:38] dpb_, That tells my script which kernel to build [02:38] jsalisbury: ah, ok. ya, it would be used for that purpose then. [02:38] dpb_, For now, I can just hardcode it to get it working :-) [02:38] jsalisbury: let me give you an example that I have been doing. [02:39] dpb_, cool. In my script I do this: [02:39] RELEASE=`config-get release` [02:39] ARCH=`config-get arch` [02:42] jsalisbury: http://pastebin.ubuntu.com/ [02:42] jsalisbury: http://pastebin.ubuntu.com/977157/ [02:43] jsalisbury: so, you need two files... one called "config.yaml" in your charm root folder [02:43] jsalisbury: one you pass in with the params [02:44] dpb_, ok, so I need a metadata.yaml with the basics. Then create a second file named config.yaml with the config options [02:44] jsalisbury: actually, these go in "config.yaml" in your root dir [02:45] jsalisbury: metadata is separate and distinct [02:46] jsalisbury: then in your charm, you use "config-get name" to retrieve a param called "name" [02:47] dpb_, awesome I'll give it a try [02:47] jsalisbury: cool. :) This has worked for me. I'm using the PPA version of juju, btw. [02:47] dpb_, thanks for the help! [02:47] jsalisbury: yw [04:52] dpb_: Any file that's in the charm room will be included in the deploy, so you can copy items you want to test on to the charm, then deploy, and when you need to interact with them all of the hooks CWD are the root of the charm. so if you have this in a folder called "foo" in the charm you could just do foo/ to manipulate it [05:04] SpamapS: so I was still having issues with lxc local testing but I made the fixes [05:04] for locker ^ [05:35] SpamapS, naxsi is WAF: http://code.google.com/p/naxsi/ === daker__ is now known as daker === Furao_ is now known as Furao [13:26] anyone awake [13:32] koolhead17: just me :) [13:32] flepied: naxsi looks really cool btw [13:32] SpamapS, good morning :) [13:35] SpamapS, I don't fully master naxsi, it was just another try to do a charm ;-) [13:36] flepied: are you at UDS? Did we meet and I didn't realize it? [13:36] SpamapS, no I'm in France ;-) [13:37] flepied: ahh ok. :) [13:39] flepied: I think imbrandon is working on an nginx charm, naxsi might be an interesting add-on to that [14:08] SpamapS, I think it's not an addon to nginx in term of packaging [14:09] nginx is included in the naxsi package [14:28] flepied: It might be an addon for the charm [14:28] as in.. something you can turn on to increase security [15:32] marcoceppi: Hey, I tried my branch, with our internal "canonistack" and S3.. it worked fine [15:32] $ [15:33] SpamapS: interesting, I'll have to give it a try again, maybe using a different AWS account [15:35] If this does work, will it land in the core? or forever be a branch? [15:35] marcoceppi: I hope we can get it into core [15:35] marcoceppi: at least as a stop-gap until we can do the right thing and de-couple storage from compute [15:35] The second it does let me know and I'll update the Ask Ubuntu question and start pushing it out [15:36] in much the same way libcloud does [15:36] I agree, an eventual de-coupled structure would be great [15:37] actually, it would be sweet if I could just say USE THIS FOR THE COMPUTE, USE THIS FOR STORAGE, and have storage providers like dropbox, s3, u1, etc [15:37] +1 [15:38] Though, native OSAPI would be good too ;) [15:39] marcoceppi: right, OSAPI does not define swift though. ;) [15:39] SpamapS: http://www.nooooooooooooooo.com/ [15:39] hp-cloud: { type: ec2, storage-type: swift } [15:40] Oh, that's cool then [15:40] thats kind of how I see it working [15:40] and then have flavors that extend and default to things, so eventually hp-cloud: { type: hpcloud } would be short hand for the above. [15:42] +1 IMPLEMENT ALL THE THINGS [15:48] we need to dive into the review queue tonight [15:50] actually there's no sessions I need to care about until 11:00 [16:04] SpamapS: I'd be be happy to dive in to the review queue, tomorrow <3 [16:04] hah, cute! [16:05] Or at least after I finish ze charms [16:06] jcastro: thanks for +1'ing imbrandon. :) [16:06] imbrandon: welcome to charmers!! :) [16:06] GET REVIEWIN! [16:06] http://jujucharms.com/~david-duffey/precise/ddclient [16:07] isn't that awesome? [16:07] ok from my count we're up to 17 entrants [16:07] +3 from me in an hour [16:07] or a few hours [16:08] Wait, charms or entrants? [16:08] oh snap [16:08] thats awesome [16:09] SpamapS: I've got a few questions about -r and relation-(get|set) [16:09] marcoceppi: fire away. Also if you're in 205 we can just have a G+ hangout w/ the room. :) [16:10] SpamapS: I am in the room! that would be sweet [16:10] the commands failed yesterday: relation-get private-address -r unit/# (literally translated to relation-get private-address -r gluster/2) and it failed out [16:10] what format does -r expect? [16:10] We're pretty light in the room now, but if we start getting a lot of questions I'll put your face on the wall [16:11] SpamapS: also, I poked at the nginx reverseproxy and haproxy charm hooks, it looks like they were using relation-get private-address - unit/# [16:12] Am I missing what the -r does, man pages are a little light on documentation :) [16:13] marcoceppi: -r allows you to reference a relationship from any context === fenris is now known as Guest8617 [16:13] marcoceppi: so you can run relation-set from a cron job for instance [16:14] SpamapS: ok fixed. [16:15] SpamapS: juju charm and juju contributions are the same as other membership requirements [16:15] marcoceppi: sent you a G+ invite [16:15] so, interacting with the community, sustained contributions, etc. [16:15] it counts the same as anything else [16:15] jcastro: sweet === Guest8617 is now known as ejat [16:21] SpamapS: so what would an example of a -r value [16:23] marcoceppi: anything that 'relation-ids' spits out [16:23] marcoceppi: or that has arrived in $JUJU_RELATION_ID in a hook [16:23] SpamapS: I wish i had a second laptop, Otherwise people will see you and you'll be staring at me the whole time [16:23] marcoceppi: so for instance, in joined, you might store $JUJU_RELATION_ID somewhere, and then re-use it to push values in later when a long running background job is finished. [16:23] SpamapS: interesting that's what I currently do [16:24] marcoceppi, will you have sometime today? i would like to sit with you and know what all new things have got integrated with juju. i missed last evening session :( [16:24] i want to write a simple blog putting LXC as well as ec2 in mind [16:26] marcoceppi: if you run outside the juju process, you also need to save off the socket and stuff. Check the env for the right JUJU_* variables [16:26] SpamapS: I keep a directory or $JUJU_RELATION_ID as the file name, and it's remote address in there. So to build that I'm looping through relation-list and performing some simple tests, etc in the first few lines. It's the relation-get portion that fails http://paste.ubuntu.com/978195/ [16:26] SpamapS: It's running inside a relation-changed hook [16:26] s/directory or/directory of/ [16:27] ah ok [16:27] the -r command fails [16:27] yeah thats wrong [16:27] you're passing unit names [16:27] bahhhhhh [16:27] So relation-ids is what I want to loop against? [16:27] Youmay just mean 'relation-ids' there.. but I'm not entirely sure [16:28] SpamapS, marcoceppi - you want to use relation-ids [16:28] You'll have a relation ID per *service* you are related to [16:28] SpamapS: jimbaker gotchya! [16:28] so if you want to then also loop through the units inside each service you're related to, you need another loop for relation-list [16:31] SpamapS: awesome jimbaker gave me the lowwwwww down [16:31] KEEPING CALM, AND CARRYING ON [16:34] werd [16:46] reviewed ddclient [16:49] DING DING promulgated SOLR! [16:50] Woo! [17:13] SpamapS: bumped service-stacks to not conflict with bug triage [17:14] m_3_: thank you! [17:14] m_3_: IMO stacks is not something we can talk about this cycle anyway. :( [17:15] m_3_: unless we're going to make a straw-man implementation in jitsu [17:15] SpamapS: yeah, I was more gonna ask if it was a core feature or not [17:17] m_3_: no features this cycle [17:18] The resource map discussion was even a bit reaching. :-/ [17:18] gotcha [17:18] Important to discuss, but I can't see it getting implemented. [17:18] what about tomorrow? any we should combine/rearrange/delete? [17:19] SpamapS: when you get a chance... no hurry, just wanna give michelle as much notice as possible [17:20] yeah I'll review tomorrow [17:21] integration's still worth keeping imo [17:21] but yeah, I'll let you listen away :) === Furao_ is now known as Furao [17:44] heh.. we need an environment option.. "spare machines" [17:44] I've taken to always having one machine deployed using the 'ubuntu' charm [17:44] just sitting there, waiting [17:46] marcoceppi: you know what would be awesome in charm-helper-sh? If you could build in support for scraping MD5's from sourceforge right into ch_file_get [17:46] SpamapS: I tried that, I ended up bailing to use php :\ [17:46] Since you have to consume an XML file [17:47] Unless there's another way to get MD5 sums from SF [17:48] marcoceppi: perhaps that is worth putting in a little helper script (I'd suggest python, not php) that gets included w/ charm helper tho [17:48] SpamapS: if that's okay with you, I think it'd be great [17:49] (I agree with Python for this instead of PHP, shazzer IIRC moved my php script to python in one of his charms) [17:50] marcoceppi: part of me wants to start pushing for more charms to just embed a tarball tho [17:50] marcoceppi: when we add this mythical tagging thing.. we will be tagging most of these "download the latest" charms as non-repeatable. [17:53] https://bugs.launchpad.net/charms/+bugs?field.searchtext=&orderby=-importance&field.status%3Alist=NEW&field.status%3Alist=FIXCOMMITTED&assignee_option=any&field.assignee=&field.bug_reporter=&field.bug_commenter=&field.subscriber=&field.structural_subscriber=&field.component-empty-marker=1&field.tag=new-charm&field.tags_combinator=ANY&field.status_upstream-empty-marker=1&field.has_cve.used=&field.o [17:53] mit_dupes.used=&field.omit_dupes=on&field.affects_me.used=&field.has_no_package.used=&field.has_patch.used=&field.has_branches.used=&field.has_branches=on&field.has_no_branches.used=&field.has_no_branches=on&field.has_blueprints.used=&field.has_blueprints=on&field.has_no_blueprints.used=&field.has_no_blueprints=on&search=Search [17:54] Is anyone here currently using canonistack? I seem to be having an issue starting an instance. [17:56] http://bit.ly/lp-charms-review-q [18:18] SpamapS: $JUJU_RELATION_ID is what you would use against -r correct? [18:19] marcoceppi: correct [18:33] SpamapS: ping [18:34] SpamapS: I approved puppetmaster and puppet. Promulgate :) [18:35] hello, someone tried to deploy openstack with juju? [18:42] SpamapS: looking at bug https://bugs.launchpad.net/charms/+bug/996576 ( reviewing ) and I see that you have this bug assigned to you. Are you reviewing this already ? [18:42] <_mup_> Bug #996576: VIVO charm < https://launchpad.net/bugs/996576 > === hazmat` is now known as hazmat [19:29] * popey wonders if someone could review my (first attempt - be gentle) charm, that would be lovely ☺ https://bugs.launchpad.net/charms/+bug/996361 [19:29] <_mup_> Bug #996361: Charm needed: Yacy < https://launchpad.net/bugs/996361 > === JoseeAntonioR changed the topic of #juju to: Coming to UDS? Write a charm for a chance to win one of 3 Dell XPS 13 ultrabooks! http://goo.gl/uTGU2 || Charms at http://jujucharms.com || Want to write a charm? http://juju.ubuntu.com/Charms || OSX client: http://jujutools.github.com/" === JoseeAntonioR changed the topic of #juju to: Coming to UDS? Write a charm for a chance to win one of 3 Dell XPS 13 ultrabooks! http://goo.gl/uTGU2 || Charms at http://jujucharms.com || Want to write a charm? http://juju.ubuntu.com/Charms || OSX client: http://jujutools.github.com/ === fenris is now known as Guest13390 [20:47] * marcoceppi reviews all the charms === gmb` is now known as gmb === _mup__ is now known as _mup_ [21:31] Hey SpamapS if you wget an asc does it *have* to be over SSL? I assume yes, but I'm not 100% certain [21:40] * popey hugs marcoceppi [21:41] marcoceppi: yes it would have to be [21:41] o/ popey got a few things for you [21:41] oh goody [21:41] is it massively broken? :D [21:41] marcoceppi: best to embed a trusted key in the charm [21:41] SpamapS: that's what I figured, cool [21:42] marcoceppi: then you don't have to have https for the .asc, you just need to verify the key [21:42] popey: not that bad! just a few key items and some nitpicks [21:42] ok [21:42] you gonna braindump on the bug marcoceppi ? [21:43] popey: yeah, just finishing up my reply [21:46] just -1'd VIVO [21:46] If anybody sees Nicholas Skaggs let him know he has some TODO's waiting for him in bug 996576 [21:46] <_mup_> Bug #996576: VIVO charm < https://launchpad.net/bugs/996576 > [21:48] Same for David Duffey for ddclient bug 996905 [21:48] <_mup_> Bug #996905: New dynamic dns charm ddclient < https://launchpad.net/bugs/996905 > [21:48] thanks marcoceppi [21:52] imbrandon: hey, you going to finish drupal? === fenris is now known as Guest18607 [22:33] 90 minutes left for the laptop contest. Anybody freaking out? ;) [22:36] SpamapS, is there a document somewhere that explains how to configure environments.yaml to use canonistack instead of AWS or local? [22:36] SpamapS: yes [22:37] jsalisbury: yes but Its just for Canonical people. [22:37] marcoceppi: Could we test my charm on Openstack: https://code.launchpad.net/~komputes/charms/precise/remote-assistance/trunk [22:37] komputes: I can do it on AWS [22:37] * SpamapS will test juju against trystack soon [22:38] SpamapS, Is there info on the internal wiki? [22:40] SpamapS: I'm getting the idea that jsalisbury may be 'Canonical people' :P [22:41] lifeless: indeed, I just don't want others in here thinking we're using some secret cloud. ;) [22:41] which.. canonistack sort of is. :) [22:42] looks like there are no more juju related sessions for today [22:43] at UDS I mean [22:43] so, anybody making last minute tweaks, lets get 'er done [22:52] * tumbleweed has no energy left for last minute tweaks (such as all the things I actually intended to do :P ) [22:53] tumbleweed: reviewing ibid right now btw :) [22:53] UDS-exhaustion :) [22:53] SpamapS: yeah, just saw that [22:54] tumbleweed: you should move most of the config-get stuff to config-changed [22:54] tumbleweed: so it can be changed post-install [22:55] oh, duh [22:55] * tumbleweed quickly does that [22:56] tumbleweed: otherwise, it looks pretty cool. I've thought for a while that we need an IRC bot charm. :) [22:56] * SpamapS deploys w/ --constraint mem=64M curious if that will end up with a t1.micro [22:58] SpamapS: I realized earlier that i collided with someone elses work [22:59] marcoceppi: gitolite? [22:59] SpamapS: pushed, but entirely untested :) [22:59] tumbleweed: mmmm.. fresh crack [23:06] tumbleweed: another nice thing would be to open the ident port [23:09] SpamsPet: are you a real person? [23:09] SpamapS: *blink* [23:09] SpamsPet: what is ibid? [23:09] SpamapS: I'm afraid I have no idea [23:10] SpamsPet: you respond awfully fast [23:10] SpamapS: What? [23:10] SpamapS: yeah, not sure what do do about that [23:10] marcoceppi: shazz has first dibs its not his fault you forgot to follow the process :) [23:10] marcoceppi: he's still within 30 days of his last activity [23:10] right, so mines ready for submission [23:10] just wait it out? [23:11] tumbleweed: I'm sure you've already guessed, but SpamsPet is an ibid deployed w/ your charm :) [23:13] tumbleweed: I manually installed oidentd, and just used the juju-jitsu 'open-port' command to open port 113 [23:14] SpamsPet: hi [23:14] lo [23:14] tumbleweed: By the way, SpamapS on irc told me "tell tumbleweed you are a genius" 6 minutes and 27 seconds ago [23:14] tumbleweed: By the way, SpamapS on irc told me "tell tumbleweed you are a genius on #juju" 6 minutes and 17 seconds ago [23:15] tumbleweed: also in config-changed, just take the existence check out for ibid.ini ... just overwrite it every time. [23:15] already fixed that :P [23:15] oh good :) [23:16] hah rapid iteration eh? :) [23:16] yeah, there are many situations where one doesn't want such a simple config [23:16] but the hacky temporary solution is to stick a local.ini in [23:16] I see some other charms have a way to download a config / provide it at deploy. That seems the way to go for non-trivial configs [23:17] tumbleweed: perhaps look at dotdee too [23:17] tumbleweed: it allows you to emulate .d dirs for things that don't have them. [23:18] tumbleweed: automation always gets in the way of customization. Such is the nature of the beast. [23:18] ejat: hey btw, I didn't get a c hance to say hi yesterday, but, HI! [23:21] SpamsPet: Juju is DevOps DistilledTM [23:21] tumbleweed: I'll remember that [23:22] negronjl: thanks for +1'ing the puppet charms. PLEASE do add things to them. I think there's a ton of interesting stuff left to do. [23:23] * SpamapS rings the bell twice [23:23] puppet(subordinate)/puppetmaster promulgated [23:37] fuuuuuuuuuuuuuuuu [23:37] heh [23:37] running out of time? [23:37] just realized one of my charms depends on charm-helpers and I never merged the dang code for that [23:37] SpamapS: its ok .. hi back [23:39] i guess .. i didnt have time to separate the package to make it scalable .. need to extract the package and see the create db file .. [23:39] marcoceppi: its ok to embed that in your own charm. :) [23:39] SpamapS: yeah I am, I was just ready to start celebrating :) [23:39] marcoceppi: charm-helpers is just an efficiency play. But its more important that charms work than that they are super efficient. [23:39] * ejat less hope to fullfill the 5 criteria .. [23:40] jcastro: where can people look to be sure their charm is submitted to the contest btw? [23:40] http://paste.ubuntu.com/978937/ [23:40] marcoceppi: dont forget to donate ya dv7 [23:40] 0/ [23:41] marcoceppi: http://www.quickmeme.com/meme/3p7ol1/ [23:41] jcastro: ROFL [23:42] :) === elmo_ is now known as elmo [23:44] jcastro @ anyone : can u answer SpamapS Q .. so i also can look into :) [23:45] SpamapS: launchpad? [23:45] SpamapS: all the new-charms ones since april 26. [23:45] jcastro: so, tagged w/ new-charm, commits to a linked branch by 1700 ? [23:45] yeah [23:45] ok, lets make sure to snapshot that at 1701 [23:46] SpamapS: why? Are you working on a list of entries? [23:46] jcastro: I want to make sure we capture it is all [23:46] k [23:46] so how can I do that? [23:46] I'm guessing I can't use the gitolite one for the contest? [23:46] I mean, the date time is in LP already [23:46] I'm sure we can extract it w/ the lp api .. but simple to just look at the page at the entry time [23:46] oh ok [23:46] right right, that's easy, I can do that. [23:47] marcoceppi: I'd -1 it pretty hard since you missed the process entirely. ;) [23:47] Cool, so I have glusterfs and no way to show it's power :) [23:47] :( [23:47] There is no way I can finish the reddit stack charm and openphot charm by 5pm [23:47] le sigh [23:47] marcoceppi: 13 minutes to add it to wordpress ;) [23:48] marcoceppi: in that case, you can use your gitolite just to support gluster's awesomeness [23:48] ಠ_ಠ [23:48] okay, whew [23:48] ok thanks [23:49] marcoceppi: but yeah, it would have been more awesomer to have a charm that uses gluster up for consideration :) [23:49] * marcoceppi is the saddest [23:49] 10 MINUTE WARNING [23:49] marcoceppi: contact shazzner, maybe you can get him to relinquish his claim :) [23:49] don't think that will happen in 10 mins :\ [23:49] * ejat ...... scratching .... [23:54] SpamapS, thanks for reviewing my charm. I implemented all your suggestions except using secure git or implementing some form of cryptographic verification(I'm not sure I'l be able to do that in the next 5 minutes, lol). [23:54] marcoceppi: you love me [23:55] I wish I would have checked earlier [23:55] marcoceppi: https://bugs.launchpad.net/charms/+bug/906176 [23:55] <_mup_> Bug #906176: Gitolite charm < https://launchpad.net/bugs/906176 > [23:55] I pushed my branch and emailed him [23:55] here's to helping [23:55] marcoceppi: I got a hold of Chris [23:55] LOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO [23:55] marcoceppi: note, unassigned [23:55] :OOOOOOOOOOOOOOOOOO [23:55] * marcoceppi owes SpamapS ALL OF THE BEERS [23:56] * SpamapS deflects at least one to shazzner [23:59] * popey crosses his fingers