=== timrc-afk is now known as timrc === CyberJacob is now known as CyberJacob|Away [06:44] lazypower: hey, thanks for reviewing! I was checking and don't see a MP on the branch atm [06:51] oh, there we go [07:09] Yeah, I was still composing :) [07:09] There's a few things I had to do to bring it up to spec for continued review, most of which was mentioned previously by Ben [07:10] If you like i can include those in the MP for you to use as a baseline for the fixes [07:12] lazypower: what are those? I can try and fix them now [07:12] (already merged the branch) [07:14] in config-changed you compare using =! instead of -n [07:14] it should read if [ -n "$SSL" ]; then [07:14] fixed that [07:14] (fixed everything that's on the comment already) [07:14] and your PWDIR variable is unbound [07:14] ah then you should be up to snuff for the charm-run completing, just validate the intent is being performed :) [07:15] lemme check then :) [07:16] I'm going to head out and get some sleep. I'll be around closer to 9AM EST for a review update if its ready [07:17] sure, I should be pushing in a minute :) === CyberJacob|Away is now known as CyberJacob [09:59] rogpeppe: Hi there. RE you question last Friday, yes requests is a python library you'd need to install. Sorry, I didn't mention that in my paste. I saw you were able to repro the issue in the end though...(with the other persons steps). Great! [10:00] noodles775: the issue should be fixed in trunk now [10:00] woot :) === timrc is now known as timrc-afk [10:29] hey all :) [10:29] hey fwereade [10:38] fwereade: are you around? [10:38] eagles0513875_, yeah, but meeting in a mo [10:38] fwereade: ok hit me up when you return then it can wait. [10:39] eagles0513875_, better to ask [10:39] fwereade: just a quick question for using lxc containers or using juju on my vas i need to use manual provisioning right [10:39] eagles0513875_, I might see and answer, so might someone else [10:40] eagles0513875_, those are different things really -- you can experiment locally with lxc/kvm in the local provider, you can experiment on arbitrary machines with manual provisioning [10:41] ok fwereade but is it going to be a bit tricky though to ensure that even though I'm using manual provisioning that it won't spawn another instance? [10:41] or want to spawn another instance [10:46] eagles0513875_, if you're using a manual environment juju never provisions any other instances itself [10:46] eagles0513875_, you have to add-machine them manually [10:46] fwereade: ahh ok [10:47] so basically if i need to deploy another instance of lets say wordpress on a different vps i would need to add the machine but then would it ask me which machine I would like to put it on? === timrc-afk is now known as timrc [10:48] eagles0513875_, if you care in detail about placement, you should probably specify it by hand with --to [10:49] eagles0513875_, otherwise we'll put new units onto clean empty machines hat match the service/environment constraints [10:49] got it [10:50] fwereade: it should i think allow you to specify which machine you want to deploy it on even if you already have an instance on it [10:59] hey all is it possible to install juju on debian? [11:02] marcoceppi, not having 'default' config is no longer acceptable? [11:03] eagles0513875_, there's a --to param for deploy/add-unit [11:03] fwereade: ? [11:03] oh ok [11:03] eagles0513875_, that's how you specify what machine to deploy a unit to [11:03] ahh ok [11:04] heya all, finally figured out my local lxc issuesm in case anyones interested: https://bugs.launchpad.net/ubuntu/+source/lxc/+bug/1205086 [11:04] fwereade: you can't really do that from the web interface can you [11:04] <_mup_> Bug #1205086: lxc-net dnsmasq --strict-order breaks dns for lxc non-recursive nameserver [11:04] fwereade: also is it possible to work with juju on debian [11:04] eagles0513875_, yeah, I don't think that's exposed there [11:04] fwereade: which is a shame tbh [11:04] eagles0513875_, work on machines in the gui is ongoing [11:04] everything you do on the command line imho should be exposed and usable on the web interface [11:05] eagles0513875_, they're separate projects, they don't completely lockstep [11:05] btw fwereade sorry if I'm interrupting your meeting [11:05] eagles0513875_, but yeah I understand that position and generally support it ;) [11:05] fwereade: what would be interesting to see is if juju would become platform agnostic meaning will work on debian or ubuntu [11:05] eagles0513875_, depends which bits tbh [11:06] eagles0513875_, you can currently only deploy to ubuntu, but can deploy *from* from win/mac as well [11:06] eagles0513875_, you never know, the `juju` binary might just work unmodified on debian [11:11] fwereade: what the person I'm talking to is interested in knowing is if we can deploy stuff to debian based machines. [11:11] it would be great if that could be done [11:11] tbh [11:12] eagles0513875_, it can surely be done, but not without some work [11:12] fwereade: would be more then willing to help out with that but i guess i should get familiar with the code first [11:32] fwereade: what would be the best and first steps to get things goign with contributing plus working towards joining your team [11:33] eagles0513875_, I would start by dpeloying an environment and fiddling with the charm [12:29] fwereade: regarding the above as i just saw this would you recommend I do it with the local provider? [12:30] fwereade: out of curiosity to get juju to work on an ubuntu private cloud [12:30] ubuntu server would need to be setup as MAAS right [12:36] fwereade: i think there is an issue with the documentation for the local provider [12:36] getting an error when i try to bootstrap to the local provider about an ssh key [12:37] and there in the documentation there is no mention of generating an ssh key there [12:37] fwereade: https://juju.ubuntu.com/docs/config-LXC.html [12:51] eagles0513875, https://juju.ubuntu.com/docs/ has it right at the top [12:51] eagles0513875, but fwiw the trunk version generates one for you if it can't find one [12:52] in terms of the juju site where would i file a suggestion [12:52] getting started the way it is i wouldnt think to click it as it seems to be a section name for the menu [12:53] i would put getting started as a first menu [12:53] fwereade: brb [13:17] jamespage: yes, if you want a "null" default just use a empty string `default: ""` [13:17] jamespage: It's a W so it's not a critical Juju error, but rather it affects juju tools (ie: GUI, deployer, etc) [13:17] * marcoceppi should document what each level of error is for proof [13:43] marcoceppi, that does not work [13:43] "" != None with config-get --format=json [13:47] marcoceppi, btw do you have a mysql charm for trusty yet? we're still using a butchered one in the openstack test lab [14:05] fwereade: i find it a bit strange if one is using a local provider an SSH key is required [14:06] fwereade: it make sense if you are testing on a remote server but in my case i am not [14:06] i guess it should be presented as an option to the user when setting up the local provider [14:28] hey guys is it possible if im running the local provider on my laptop to bypass the need for ssh keys [14:36] eagles0513875, it's not really -- every difference between the local provider and the others is a source of potential nastiness, and adding a special case to allow people not to use ssh isn't worth it [14:37] eagles0513875, is there a problem with just running ssh-keygen, or setting authorized-keys[-path] directly? [14:37] fwereade: non no i just find it a bit useless if im running the local provider on the same laptop im on now [14:37] here is the use case im thinking of [14:38] fwereade: what im trying to get at is why not provide users the ability if they are using the local provider to use local host which doesntn require ssh [14:38] and if they want to use the local provider obviously they would need ssh there [14:38] eagles0513875, because you're not sshing tolocalhost? you're sshing to the containers [14:39] ahhh ok [14:39] that isnt that clear in the documentation [14:39] eagles0513875, "you need a keypair" is pretty clear, I think [14:39] sorry for the confusion [14:39] eagles0513875, no worries :) [14:39] fwereade: what isnt clear is what its used for. it just says generate a key pair === freeflying is now known as freeflying_away [14:40] eagles0513875, the reasons could be better explained, but IMO that's not really appropriate for step-0 introduction [14:40] just have to figure out a way to keep my production private key seperate from the one im going to be generated [14:40] eagles0513875, what's the concern there? [14:41] with the id_rsa file be appended to? [14:41] eagles0513875, I don't follow -- you have a private key set up already but haven't got the public key there? [14:41] i think my mind is just out to confuse me today [14:42] eagles0513875, is it that you have id_rsa in ~/.ssh, but not id_rsa.pub [14:42] ? [14:42] fwereade: correct [14:42] eagles0513875, hmm, why would you do that? [14:43] fwereade: i have my private key in there so when i ssh into the vps's i have in production i just have to give the key's password [14:43] eagles0513875, sure, but where do you get the public key from when you're setting up one of those machines? [14:43] i have the public key saved in a secure location so i then upload it to the server [14:44] eagles0513875, it's the public key [14:44] eagles0513875, you can hand it out willy-nilly [14:44] fwereade: ok i have my mind in knots today lol [14:44] eagles0513875, it's the private key you need to take care of ;p [14:44] id_rsa the private key [14:44] eagles0513875, yeah [14:44] ok [14:44] i have one of those already in there the public key is on the servers [14:45] if i generate another one will it apend the new private key to the already existing id_rsa [14:45] eagles0513875, no, I would recommend sshing into one of your servers and extracting your pubkey from authorized_keys [14:46] eagles0513875, I don't think it's very meaningful to have a private key appended to another? [14:48] fwereade: ok i have both my public and private key pair for my prodcution servers in a secure location [14:48] let me see what happens [14:48] eagles0513875, so you've got an id_rsa.pub right next to the matching id_rsa now? cool [14:49] ya [14:49] eagles0513875, note: first time you spin up a local provider will take a long time [14:49] fwereade: whats not making much sense for me is that with out it and just the private key on my laptop ssh keys work fine im just trying to understand why i should have the .pub key there [14:49] especially if i have two private keys in .ssh [14:50] eagles0513875, I can't see a reason offhand *not* to keep your public key alongside your private one [14:50] fwereade: i dunno why im worrying it gives you a default but also allows you to specify a different file name for the private key when generating [14:50] good point [14:51] also fwereade [14:52] im an idiot lol [14:52] i didnt notice the path when generating the key from within .juju is that it creates the pair in .juju [14:52] and not in .ssh [14:52] eagles0513875, oh, did we already release a version that creates a keypair? [14:52] fwereade: seems like it [14:53] eagles0513875, hmm, in that case we ought to be using it [14:53] fwereade: its 1.16 [14:53] * fwereade grumbles a bit [14:53] fwereade: Installed: 1.16.5-0ubuntu1~ubuntu13.10.1~juju1 [14:53] thats whats showing up in my apt-cache policy [14:53] fwereade: probably because im using the ppa [14:53] eagles0513875, weird, I'm sure that was before we added that, let me take a look [14:53] and installed from the ppa instead of directly from the repo [14:53] anyway i gave it a name to differentiate between the two anyway if it did land in .ssh [14:54] eagles0513875, just a mo, what did you do? didn't you just get the pubkey from your servers? [14:55] fwereade: no no i was just worried of the key being overwritten when i generated the key for the juju local provider [14:56] eagles0513875, sure, I'm just wondering why you'd need to generate a new one when you have one already [14:56] good point just like keeping my testing and production stuff seperate [14:58] fwereade: there is something missing in the documentation [14:59] fwereade: do i need to add the pub key to the environments.yaml file for local cuz im getting this error ERROR error parsing environment "local": no public ssh keys found [15:00] eagles0513875, please tell me exactly what you did [15:00] sudo juju bootstrap [15:00] it doesnt seem to pick up the keys [15:00] no, what you did to your ssh setup [15:00] fwereade: for juju [15:02] fwereade: when generating the keypair should it end up in .juju if you dont use the default path to .ssh [15:03] eagles0513875, AFAICT you should not be generating a keypair at all [15:03] ok [15:03] eagles0513875, you should have got your pubkey and saved it in ~/.ssh/id_rsa.pub [15:03] ok [15:09] fwereade: when one learns to untangle their complex mind setting up juju as the local provider is very very simple [15:10] I managed to remove my lbox authentication from launchpad in haste, and I'm not positive where it caches its copy of the oauth token. Can someone point me where I need to look to clear its cached authorization? [15:11] lazypower, I think it's ~/.lpad_oauth? [15:11] eagles0513875, glad it worked out -- I hadn't considered the private-key-only situation [15:11] that was it, thank you [15:11] lazypower, yw [15:12] fwereade: should i file something on launchpad [15:12] eagles0513875, yes please, that would be very helpful [15:12] fwereade: under juju core? [15:12] eagles0513875, please [15:12] ok fwereade also any bugs i can help or work on in terms of fixing? [15:16] fwereade: should i run a test on what we just discussed [15:16] and maybe its something i could be mentored in fixing [15:16] fwereade: if you are interested in following this bug https://bugs.launchpad.net/juju-core/+bug/1270858 [15:17] <_mup_> Bug #1270858: possibility of .ssh not having a public key with the private key [15:18] eagles0513875, so, I think the fix is to scan for files with/without .pub suffixes, ignore the ones that don't, and complain about the lack of a keypair (rather than just of a public key) [15:18] fwereade: wouldnt it be good i test with out having the public key first [15:18] to see if it works or does not work cuz right now the issue i filed seems rather vague and not useful [15:19] fwereade, if lbox fails to propose due to a non existant [branch|repository] is this a sign of a larger problem? [15:19] eagles0513875, well, the ideal bug report is generally along the lines of "what I did; what I expected; what actually happened" [15:20] fwereade: ok will get more information when i relocate for my ccna class [15:20] lazypower, hmm, nothing really springs to mind there [15:20] anyway im out for now just deploying wordpress to try things out :) [15:21] lazypower, is it saying the source or the target doesn't exist? [15:21] fwereade: i have any idea for a really good charm setup. how can I see lets say what the apache charm installs for the web server as apache has about 4 different versions [15:21] fwereade, http://paste.ubuntu.com/6786390/ [15:22] eagles0513875, that would be in the install hook of the charm you're looking at. [15:22] lazypower, hmm, I'm not completely sure there -- marcoceppi, have you ever seen that? [15:22] eagles0513875, lazypower: or more generally inferrable from a variety of hooks, and/or from the config file [15:22] ok [15:23] Fair enough :) [15:23] fwereade: lazypower not to but in here but could it be the directory /charms/ was moved and the path you are looking for doesnt exist anymore [15:23] eagles0513875, lazypower: several charms let you choose different versions to run [15:24] Its complaining about the remote. My charm is still here locally, but i believe its trying to push to the wrong remote. [15:24] lazypower: check which branch you are on [15:24] lazypower, hmm, should you be pushing to ~lazypower/charms/mediawiki/trunk perhaps? [15:25] yeah, i've got an open MP that I was trying to get on codereview [15:25] marcoceppi, do yu have an immediate answer for lazypower by any chance? [15:25] thats why i include the parent branch in the config, and i think marco is afk for now. [15:25] s/config/output listing on pastebin [15:25] fwereade: i have some enhancement suggestions but will need to discuss those with you when i return [15:27] lazypower, ah, hold on, surely you just need to push to that branch before you first lbox propose -- right? [15:27] fwereade, https://code.launchpad.net/~lazypower/charms/precise/mediawiki/tests [15:29] ah ok, now i've got additional legwork. My MP branch and local copy have diverged [15:29] lazypower, ok, that's pretty unambiguous :) [15:29] sarcasm? :) [15:30] and i apologize for my ineptness, i'm still getting comfortable with the toolchain [15:30] lazypower, no, not at all, the existence of your branch is 100% clear [15:30] eagles0513875, if I was digging at anyone it was myself ;) [15:31] haha [15:31] lazypower, er^ [15:31] You get my seal of approval gustavo [15:31] lazypower, I'm not gustavo, but I do observe that your error message doesn't include the "precise" I'd expect [15:31] lazypower: Heya [15:32] lazypower, and I was just about to direct you towards the man himself :) [15:32] O_O [15:32] lazypower: Thanks, I suppose :) [15:32] mind=blown [15:32] i ran across a post somewhere from you that referenced gustavo and made the mental note it was you... whooops! [15:33] clearly i need to go back downstairs and not return until i've finished my first cup of coffee [15:34] niemeyer, gustavo i presume? [15:38] * marcoceppi reads scroll back fwereade lazypower [15:39] lazypower fwereade use the -for flag [15:39] lazypower: lbox propose -cr -for lp:charms/mediawiki [15:39] marcoceppi, lbox propose -for lp:~blah blah [15:39] lazypower: Yeah :) [15:39] ooohh [15:40] and good morning [15:40] * marcoceppi still has reservations about using lbox for charm reviews, but those are on hold [15:40] huh, I hadn't read that paste as a problem with the target [15:41] lazypower: can you run bzr info as well? [15:41] lazypower: hey, got to fix the postfix charm [15:42] fwereade: I re-read it looks like it's trying to push to lp:~lazypower/charms/mediawiki instead, so maybe the -for flag isn't the issue [15:42] maybe he doesn't have a push branch in bzr yet? [15:43] jose, great news. Did you put it back in the rev queue? I'll fish it out after i've solved my lbox issue. [15:43] lazypower: yeppers! marked it as new [15:43] marcoceppi, http://paste.ubuntu.com/6786495/ [15:44] lazypower: okay, looks good to me. did the -for flag fix this? If not just do `bzr lp-propose lp:charms/mediawiki` and we'll look at lbox for charm reviews tomorrow [15:45] marcoceppi, i'm working on getting my branch undiverged from whats open as the MP - one moment. Reading up on bzr [15:46] fwereade: will ping you veyr soonn [15:46] i have some nutty but useful ideas :) [16:10] hey fwereade im back [16:28] hi all, got a problem with 1.17 and MAAS, br0 isn't being brought up by Juju when adding a new MAAS machine [16:44] teknico: is the juju agent running on these new machines though? [16:54] marcoceppi, at the end I got this: https://pastebin.canonical.com/103229/ [16:57] teknico: is that from cloud-init log? [17:00] marcoceppi, nope, it's the end of the output from trying to start the container manually with "sudo lxc-start --name juju-machine-1-lxc-0" [17:01] in the log there was nothing more than the original error itself [17:01] which was: agent-state-info: '(error: error executing "lxc-start": command get_init_pid failed to receive response)' [17:02] which in turn was caused by br0 missing, at least according to bbcmicrocomputer :-) [17:03] teknico: makes sense, can you paste the cloud-init log file? [17:03] it should be done during lxc install tbh [17:05] marcoceppi, the whole of cloud-init.log? it's 45KB, it'll take a while pasting it :-) [18:07] marcoceppi: I had a question on that swap charm before I proceed with fixing the review feedback. You want me to document config options in the readme *and* in the config.yaml? Seems like I would just be cut-and-pasting descriptions? Or am I missing what I should be doing there? [18:33] dpb1: config.yaml should have a brief description, readme should have a more detailed explaination as to what the config options does and provide example values [18:37] marcoceppi: OK [18:37] thx === zz_mwhudson is now known as mwhudson === mwhudson is now known as zz_mwhudson === zz_mwhudson is now known as mwhudson [19:52] marcoceppi: in amulet, how do I make use of an existing deployer file? [19:53] dpb1: you can load a deployer config in your script [19:53] it's something like... sec [19:53] mgz... thx. [19:53] * dpb1 waits [19:54] d = amulet.Deployment() [19:54] with open(path): [19:55] script = yaml.safe_load(f.read) # as f above, whoops [19:55] d.load(script) [19:55] d.setup() [19:55] dpb1: hope that helps [19:59] mgz: thx, trying now [20:54] lazypower: hey, did you get to review the charm? [20:57] jose: I did not, i was sidetracked upvoting on askubuntu and their rowdy crowd. Let me promote this in my queue and have a look [20:58] apologies for not getting to it sooner [20:58] sure :) [20:58] no worries [21:53] Great work on the revisions jose, I'm going to add my +1 to this charm and place it in another charmers rev queue [21:54] awesome, thanks! [21:59] hey thats good info to have handy mgz, that loads a deployment yaml, ergo a bundle, into amulet? [22:05] lazypower: yeah, if you have a deployer file already in your tests dir, you can bypass all the d.add() stuff and just do d.load() [22:05] amazingness [22:05] hey marcoceppi, if you could also take a look at the postfix charm it'd be great [22:05] there you pop up :P [22:05] I'm writing up my +1 review right now, give me another 5 and i'll have it complete [22:06] cool! [22:06] jose: technically a day off today, I'll be in th queue still since lazypower is +1 reviewing it as a "jr charmer" [22:06] so someone from ~charmers will cycle on to it sometime this week [22:06] cool thanks! [22:48] Does juju/maas use a special configuration for DHCP IP Assignment? I know that the local provider allows me to assign a network interface in the environments.yaml if i've opted into configuring my own bridge device. [22:54] lazypower: maas has it's own DHCP server === freeflying_away is now known as freeflying === timrc is now known as timrc-afk