[00:26] I just noticed that it became possible to remove a subordinate relation *hattip* ty for that gents [00:44] lazyPower: coolio [00:45] rick_h_: i'm probably lagging behind in the features landing in gui. Hows things on your side o the fence? [00:46] lazyPower: going well, antonio pointed out a giant bug today but docs now update and http://juju.ubuntu.com/ is a cool new place now :) [00:46] rick_h_: intredasting. are you aware that the /docs still links to the old docs? [00:47] (it auto completed.... maybe i'm being picky) [00:47] but i dig that its all redirecting to jujucharms.com now, thats groovy baybeh [00:55] lazyPower: the docs should be redirected as well [00:55] might be caching on my side [00:55] so have to bring that up I guess [00:55] i seem to find all those fun "you had something cached" issues [00:56] naw, it's just they just did it today so missed something there [00:58] rt replied to [00:58] so hopefully fixed up tomorrow === TheRealMue is now known as TheMue === johnweldon is now known as jw4|afk === urulama_ is now known as urulama === jw4|afk is now known as johnweldon [12:19] hey guys [12:20] someone familiar with quickstart can perhaps tell me were there any changes needed after the fix for bug 1420403 landed in juju-core (1.21, 1.22, and trunk) [12:20] Bug #1420403: juju-quickstart: bad API server response: 'NoneType' object is not iterable

[12:20]

[12:21] dimitern: no, no changes required on the client side [12:22] frankban, that's what I remember as well - is quickstart using lp:python-jujuclient internally? [12:22] dimitern: it is [12:23] frankban, and the client also didn't need to change right? [12:28] frankban, I doesn't seem like jujuclient changed since before the fix - just looking at the code [12:28] dimitern: correct [12:29] frankban, thanks! [12:29] dimitern: was the fix released? [12:29] frankban, well, 1.21.3 is released, which includes the fix, but jamespage today reported this bug 1425435 [12:29] Bug #1425435: juju-deployer/jujuclient incompatibility with 1.21.3

[12:31] dimitern: I still get "Ports":null with 1.22-beta3 [12:32] frankban, that's expected - 1.22-beta3 was released before the fix [12:33] frankban, hmm.. actually according to the changelog it should have the fix [12:34] frankban, something's wrong here.. I'll investigate more [12:37] frankban, ah, I see the problem - 1.22-beta3 includes my original fix for bug 1418433 which you then improved as part of fixing bug 1420403 [12:37] Bug #1418433: unit ports not populated by API megawatcher

[12:37] Bug #1420403: juju-quickstart: bad API server response: 'NoneType' object is not iterable

[12:37]

[12:40] dimitern: so we don't have a release with the subsequent fix? [12:44] frankban, 1.21.3 should have both, but apparently there's still a problem [12:44] frankban, can you do a quick test with 1.21.1 from the stable juju ppa? [12:48] frankban, sorry, I meant 1.21.3, not 1.21.1 [12:48] (the latest available) [13:06] dimitern: doing it now [13:29] dimitern: it works with 1.21.3-utopic-amd64 from the stable PPA [13:30] frankban, great! thanks for confirming [13:30] dimitern: so new release landed in stable but not in devel? [13:31] frankban, that's one possibility, another one that's bugging me is the issue might be still there if you upgrade from 1.21.1 to 1.21.3 - I'm testing this now [13:32] dimitern: it wasn't working on my local machine with a fresh environment using the latest devel version, so I presume devel does not have the fix [13:33] frankban, oh, that's weird, but easier to check - cheers, I'll look into the devel ppa [13:34] dimitern: ty === kadams54-away is now known as kadams54 [14:07] hey rick_h_ [14:07] have we talked about CIing all the bundles on that page on a regular basis? [14:07] we as in us collectively, not you you. :) [14:10] jcastro: it's brought up. There's the bundletester stuff and charm testing is supposed to look at bundles once the charm stuff is good, but I've no idea where that's at to be honest [14:10] yeah also we have another problem [14:11] bundles and all that jazz will work fine in AWS/Azure, etc. [14:11] then you go onsite with an OB and boom [14:11] boom because? [14:12] whatever issue is local that we can't reproduce in a cloud, for example this weekend we had a squid problem on an OB [14:12] and it didn't matter if the bundle worked or not, the entire thing needed to be hand held [14:12] I guess I am just saying that just testing bundles in the public cloud isn't good enough [14:13] jcastro: ok, good to know. We can look into doing some OB testing on things. I know they've set some up over a company vpn and maybe we can get some time to figure out what's up and correct it if it's a regular isue. [14:14] hatch: ping when you're in, frankban brought up a good point on the promulgated bundles and that the id there is just the bundle name. [14:38] rick_h_: that's not a problem though is it? [14:38] and good morning :) [14:38] hatch: no, not a problem just wanted you to make sure you kept that in mind during dev/qa [14:38] hatch: as we only talked about u/xxxxx/name urls yesterday [14:39] yeah, this is what I came up with https://gist.github.com/hatched/3613d21af62ca9c221ed [14:39] I think that it'll work for all different formats [14:40] hatch: cool will QA that then. I'll work to bring up a storefront/gui together and test out the 'add to demo' and the commands across a bunch of bundles once we're ready [14:41] sounds good - I'm just writing the GUI tests right now for this then I'll be hopping on the storefront [14:41] ty [15:54] rick_h_: the GUI PR is up https://github.com/juju/juju-gui/pull/697 [17:38] uiteam could I get a review on https://github.com/juju/juju-gui/pull/697 thx [17:38] hatch: looking [17:38] thanks [17:38] two cancelled calls in a row ftw! :) [17:39] haha less walking though [17:49] dimitern: any news on the Ports problem? [17:53] frankban, not really - I've investigated thoroughly, couldn't reproduce it and reported my comments in bug 1425435 [17:53] Bug #1425435: juju-deployer/jujuclient incompatibility with 1.21.3

[18:07] dimitern: ack, ty === kadams54 is now known as kadams54-away [18:52] great talk lazyPower [18:53] Thanks hatch :) [18:54] lazyPower: do you find people understand the ansible charms better than bash ones? [18:54] I would have assumed a bash/python charm would be easier to grok [18:57] hatch: Ansible is really well documented and gives you a lot of batteries out of the box in terms of idempotency, and logic structures [18:57] is it the 'popular' one now? vs chef and whatnot [18:57] i'm trying to campaign to get ppl using proper config management rather than spawning these behemoth toolbox libraries (like we have in charmhelpers) where cm tools have already solved this problem. [18:58] eh, its a bit of preference, yaml is easier to learn than ruby + DSL's for chef/puppet - but puppet is still hands down the 800lb gorilla in the room === kadams54-away is now known as kadams54 [19:00] ahh === jcw4 is now known as jw4 [21:25] hatch: do you know if rick_h_ is still around? [21:26] lazyPower: your guess is as good as mine :) [21:27] if he is I'm sure he'll poke his head in now that you have dinged [21:27] ok, just curious if he had an emergency to run out for [21:27] well i feel bad now, i've dinged him in 3 places :( [21:27] my name is going to be synonymous with a curse word before long... [21:27] lol [21:27] man.. chuck that. [21:27] nah you have a long way to go yet [21:27] thats the chuckest thing i've heard all day [21:29] Makyo: you able to do a qa on my PR? [21:30] hatch, sure, I'm spinning my wheels as it is [21:30] break time! :) [21:31] Got the PR number off the top of your head? [21:32] https://github.com/juju/juju-gui/pull/697 [21:32] one better :) [21:32] Thanks [21:38] hatch, LGTM (though maybe in the future, we want the "Location:" link in the header might need to be update din the future) [21:38] Makyo: yeah tbh I have no idea what the purpose of that is anymore [21:39] it's always been a thorn for me [21:39] I think it made more sense in an earlier iteration. [21:39] Oh well, something for another time. [21:39] lazyPower: yes, was on the phone doing manager stuff for a bit :P [21:40] yeah sorry about the ping crazyness :( [21:41] grr, this sysdeps thing is not working for me here [21:42] ah crap, vivid lxc container FML [21:42] hatch: have no QA for you atm, have to start over sorry [21:42] and uploading GB of stuff for lazyPower is making this brutal [21:42] :) np Makyo is on it [21:42] hatch: please get QA from gui from someone else [21:43] ty Makyo [21:43] GIVE ME ALL YOUR BANDWIDTH MWAHAHA [21:43] rick_h_: if its a better option - just hit me with what you're working on now, the others can wait until an overnight xfer [21:44] lazyPower: np, Makyo has my back and I'm going to feed the dog/etc and just let the bandwidth go [21:44] 10-4 [21:44] should see the sharepoint iso showing up now [21:44] and then you'll have to wait forever on my upstream to pull it from guimaas [21:44] run that overnight heh [21:44] i feel like we've reverted to 90's style transfer of files.... [21:44] like, xdcc [21:44] fxp [21:44] haha [21:44] all pre bt options [21:45] eff tee pee [21:45] ess eff tee pee [21:45] :) [21:45] hatch: pleb - distribution hubs used server => server transfers :P [21:45] haha [21:46] i'm not going to say how i know this... but it might be from the scene... http://www.welcometothescene.com/ [22:00] Morning [22:00] morning huwshimi [22:01] lazyPower: ok, best of luck downloading now [22:01] ta rick_h_, cheers! [22:06] hatch, the more I think about it, the more we might want to escalate the Location link, since it's no longer the same as the deploy link. [22:06] hatch, I mean before the design stuff in the pipeline. [22:06] Makyo: I always thought it was supposed to be the bzr location [22:07] morning huwshimi [22:08] hatch, used to be that the charmstore URL ("cs:") was what you could dump into quickstart. Now that they've diverged, I think it should be one or the other - quickstart or bzr. [22:08] Makyo: well both charms and bundles have the location so probably should be bzr [22:08] since you can't use qs for charms [22:09] hatch, sure. I think in that case, that we should promote the deploy tab (or the information) [22:09] Just deemphasize the basket. [22:11] Just a thought, not actionable. [22:12] bundles have the cs: and charms have lp: [22:12] lol [22:12] oy [22:12] Yeah, and some of the future stuff is a unification of styles toward first class bundles, just thinking about iterative approaches. === kadams54 is now known as kadams54-away