[00:49] ericsnow: correct, new providers need not use storage. they *can* if it's beneficial, but from what hazmat described to me it didn't sound like it would be [00:50] axw: just what I wanted to hear :) === negronjl is now known as negronjl-afk [02:57] anastasiamac: you win for providing the most useful PR description I've seen in a while [02:58] menn0: thank you ;-) [02:58] menn0: i was apprehensive of submitting another big change... was hoping to make it easier for u :) [02:59] menn0: i always appreceate ur feedback :) [03:26] menn0: hopefully the tweaks to the unit agent status pr are ok - i added a Stopping state for unit agents (machines continue to use Stopped) [03:27] wallyworld_: ok, i'll take a look after i've finished with anastasiamac's PR [03:27] sure, no hurry, ty [03:36] menn0: I've updated http://reviews.vapour.ws/r/676/diff/# to add an additional test with multiple environments [03:40] anastasiamac: regarding the first question in your PR - about a follow up PR to remove the embedded style annotation support [03:40] anastasiamac: you're just going to change the underlying implementation right? [03:41] anastasiamac: the old style API will need to hang around for a while to support older clients won't it? [03:41] menn0: yes, removing embedded annotator and tweak old client to not use it [03:41] menn0: yes, API will stay ;-) as deprecated... [03:43] anastasiamac: good. in that case I don't there's a need to increment the API version number. [03:43] anastasiamac: the external API stays the same after all. [03:43] menn0: that was my hope ;-) thnx for confirmation! [03:44] anastasiamac: cool. [03:44] ericsnow: ping [03:44] anastasiamac: I'm almost done with the review (looking good). just trying to answer your other questions on the review. [03:45] menn0: u r almost done? amazing!!! thnk u ;-) [03:47] menn0: m pondering if it's worthwhile adding a test to explicitely annotate a local charm... current test creates a "cs" charm... [03:48] menn0: it's just couple of lines of code but will be explicit and closely aligned to the concern in the bug... [03:54] * wwitzel3 dances a jig [03:57] anastasiamac: hmmm not sure. what's different about a local charm in terms of this code? [03:57] anastasiamac: regarding your last question, what do you mean by "stuttering" in the tests? [04:00] thumper: newEnvAsUser? shouldn't that be newEnvWithUser or something? [04:00] menn0: I suppose [04:00] thumper: not a biggie [04:02] thumper: those changes look good [04:02] cool [04:05] menn0: if I change the name, good to land? [04:05] thumper: even if you don't (and I had already said ship it :) ) [04:06] menn0: I don't hold those as binding if the coder does lots of changes [04:06] :) [04:06] wallyworld_: have you pushed the latest version of status branch? RB doesn't have it. [04:06] menn0: i thought so, let me check [04:09] menn0: the github pr has been updated,but rb appears not for some reason [04:09] anastasiamac: i've published the review, leaving your last question unanswered. [04:10] thumper: fair enough. well it look good regardless :) [04:10] menn0: landing now [04:13] ok, I'm off, back later tonight hopefully for meeting [04:14] menn0: sorry - was with kids ;-( [04:14] menn0: thnx for review - will look now-ish [04:15] axw: question about your disk PR: how will a user specify the desired disks via the GUI or command line client? Or is this a charm level thing? [04:15] menn0: re: stuttering - result.Error.Error.Error() ;P [04:15] anastasiamac: no problems. if you can explain the 4th question some more I can try and answer that one too. [04:16] wallyworld_: sometime the RB or GH API chokes on large diffs I believe (not sure which side is the issue) [04:17] menn0: q4... for eg, in apiserver/annotations/client_test.go, l. 131 [04:17] there's certainly something to be said for the extra efficiency that IRC style communication gives you. I was having 3 conversations at once for a while there :) [04:18] menn0: m impressed with everyone's multi-tasking ;-) [04:18] :) [04:19] menn0: from annotations perspective, there is no difference btw local vs non-local charm... i'll leave th test out ;) thnx for the listening ear :0 [04:19] anastasiamac: so by "stuttering" are you referring to to the repeated code snippets? [04:20] menn0: m using Jesse's jargon :p but yes [04:23] anastasiamac: I guess it might be nice to have methods for the most commonly used bits: makeWordpressCharm(), makeMachine() seem like the biggest wins [04:23] menn0: "juju deploy --storage =", and (at least while prototyping), "juju machine add --disks " [04:23] axw: cool. I was just curious. [04:23] menn0: those constraints will be combined with charm-level properties in the former case [04:24] axw: the Disks param to StartInstance - is that still determined from constraints which specify minimum and preferred values? [04:26] wallyworld_: there's no preferred anymore, but yes. on the CLI the user specifies constraints which may include size, count and a storage pool. those are converted to one or more DiskParams by looking up the pool, and (later) combining with charm storage metadata [04:27] wallyworld_: the pool lookup bit will not be in the MVP though; I'm assuming a single, unconfigurable pool for now [04:27] axw: right. so the code comment just needs tweaking to remove "preferred" [04:28] wallyworld_: ah yeah, thanks. [04:28] menn0: are you ok just to look at the github commit diffs for that pr? [04:29] wallyworld_: yep, will do [04:29] ta [04:44] wallyworld_: ok done. my comments are probably more due to my lack of understanding than anything wrong with the PR. [04:45] menn0: great, thank you, looking [04:45] wallyworld_: i've got to stop for a bit now but will check back in again later on [04:45] sure [04:45] wallyworld_: if my comments are silly, please feel free to merge [05:20] ericsnow: ping?.. [06:14] morning all [06:14] morning dimitern [06:15] axw, \o [06:16] dimitern: what's the plan for switching juju over to github/go-amz/goamz? [06:17] axw, right, so I'm having a call with gustavo at 16 utc to discuss this and other related stuff, like collaboration and integration with other forks, maintenance, etc. [06:18] dimitern: ok, cool [06:18] axw, my plan is to do it sooner, even if that means using the "unstable" v2-dev branch for a while [06:18] dimitern: I have a couple of minor changes I need to make to goamz, I'm holding off until that's sorted out [06:18] ok [06:19] sure [06:19] dimitern: o/ [06:19] anastasiamac, hey there [06:20] dimitern: how r u? happy ny [06:24] anastasiamac, cheers :) great to be back to work, eh? [06:25] dimitern: dfinitely :p [06:25] seriously, sometimes holidays seems a wee bit too long [06:26] dimitern: i actuually need a holiday form my holiday ;) [06:27] dimitern: as far as celebrations r concerned, we r still in Xmas mode here - today is Russian Xmas, 13th is Russian old NY, etc... [06:27] :) [06:27] anastasiamac, oh right - until 14th? [06:27] yeah [06:28] dimitern: m working and celebrating ;) [06:28] anastasiamac, that's a nice excuse for a double ny/xmas celebrations [06:29] dimitern: any excuse to celebrate anything ;-P life is a long party in my books :) [06:30] anastasiamac, awesome [06:30] :D [06:35] ah, axw - do you mind if I suggest you as a (candidate) maintainer for goamz as well? [06:36] dimitern: sure [06:36] axw, cheers [08:12] anastasiamac: looks like the annotations PR needs rebasing or something... [08:12] anastasiamac: look at all the commits listed for it on Github. Most aren't yours. [08:13] thanks menn0 :-) [08:13] anastasiamac: Reviewboard is also not happy about it [08:13] menn0: i've rebased, resolved conflicts and got this ;-( [08:13] wallyworld_: np [08:13] menn0: that's why i've added a list of files that I have actually touched into the awesome commit comment ;-) [08:13] anastasiamac: it's almost like you rebased onto the wrong branch or something [08:14] anastasiamac: i'm a little worried that if you try to merge what's there now something bad is going to happen [08:14] menn0: m considering to start another branch and move all my stuff there [08:14] anastasiamac: you might not need to do that [08:14] anastasiamac: are your commits all at the top of your branch? [08:15] menno: k. i'll do this either later or tomorrow [08:15] menn0: looks like they are all the bottom of the list... [08:15] anastasiamac: I was going to suggest that you use "git rebase --onto ..." but that won't work [08:16] menn0: i'll start another branch ;( [08:16] anastasiamac: I see annotations related commits mixed in with other people's ones when I look at Github [08:16] menn0: most of it additons anyway... ;) [08:16] anastasiamac: a rebase onto master *should* sort it out [08:16] anastasiamac: how are you rebasing? [08:16] menn0: when m on my branch 'git rebase master' [08:17] menn0: once master is synced... [08:17] menn0: have to go... i'll do another branch [08:17] anastasiamac: ok [08:18] anastasiamac: I'll send you an email with a suggestion [08:21] menn0: sounds awesome ;-) [08:21] menn0: thnx! [08:37] menn0: branch fixed as per ur suggestion! tyvm!!! scotch was ur price, wasn't it?.. [08:45] morning [08:50] anastasiamac: yep :) [10:46] sorry, freenode seems to keep falling over for me for some reason [10:46] I'm in the canonical ones too though if this falls over again [11:42] morning [11:43] fwereade: I addressed some of your concerns regarding my restore patch and made some action plan for the others since they require more in depth changes of the whole mechanism === frankban__ is now known as frankban [11:52] TheMue, katco, voidspace, can any of you have a look at this http://reviews.vapour.ws/r/686/ please? If fixes bug 1355813 [11:53] Bug #1355813: Interface MTU management across MAAS/juju [11:53] [11:54] dimitern: *click* [11:54] dimitern: I can do, but will have to be after lunch [11:54] TheMue, ta! [11:54] voidspace, no worries [12:00] dimitern: done [12:43] TheMue, thanks! [12:44] so, bfl [12:44] dimitern: yw [13:15] mgz, hey, do you have 5m? [14:07] dimitern: hey, was at lunch, what do you need? [14:08] mgz, i wanted to ask if you're willing to lend a hand from time to time with goamz reviews and generally helping with maintenance/collaboration [14:09] it's going to get more active soon, and there are ideas to merge the existing forks into it gradually over time [14:09] dimitern: yeah, saw you'd been busy over the holiday sorting that all out [14:10] I'll try and jump in on reviews on github [14:10] mgz, yeah, it was surprising how much attention it managed to gather in a short time [14:10] mgz, sweet! thank you! === kadams54 is now known as kadams54-away === negronjl-afk is now known as negronjl === kadams54-away is now known as kadams54 [14:49] dimitern: (in case you haven't been asked yet) what is the priority on implementing networking for the GCE provider? [14:50] dimitern: Basically none of the other providers implement it at present. [14:50] ericsnow, heh :) for gce - i'd say very low [14:50] dimitern: k, cool [14:50] ericsnow, we're concentrating on ec2 and maas until 15.04, then openstack will be next most likely [14:51] dimitern: FWIW, the GCE API makes the networking stuff pretty straightforward so adding it later shouldn't be a big deal === kadams54 is now known as kadams54-away [14:52] ericsnow, that's great to hear! [14:56] mgz, abentley, sinzui: could any of you take a look at http://juju-ci.vapour.ws:8080/job/github-merge-juju/1729/console? [14:56] mgz, abentley, sinzui: it is not clear to me why it can't find the dependency === kadams54-away is now known as kadams54 [15:01] ericsnow: will in just a sec [15:01] mgz: ta [15:03] ericsnow: you see my commits for gce? [15:03] ericsnow: was able to fully bootstrap and deploy mysql/wordpress last night. [15:09] is there a way to get a charm of a unit without the charm url on state? [15:11] ericsnow: golang.org/x/crypto is explictly exclused in the build script - if I recall corerctly it had licensing issues [15:12] or that might have been one of the other ones, and it was just pulled in by mistake, and can now be removed if you actually need to use it [15:12] I'll find out [15:13] mgz, ericsnow. The x/crypto lib was purged because it first appeared in the tree, but wasn't in deps. It appeared there because a subdep pulled it in, but the version of the dep we used did not use it [15:14] wwitzel3: awesome! [15:14] mgz, ericsnow (my recollection). I f the license is good, we can remove the line that purges it now that juju's dependencies state it is wanted [15:14] ericsnow: it looks like you're not actually adding it as a dep in your new code? [15:15] ericsnow: so, probably just the mechanism you used to update dependencies.tsv wasn't right? [15:15] sinzui, mgz: well, the utils repo depends on x/crypto as of a couple revisions [15:16] sinzui, mgz: so it was added as a dep when I updated the rev for utils [15:17] okay, I'll need to update the script to try to produce a sane tarball when deps get pulled in by mistake then, that will mean I can drop my other hack [15:17] mgz: cool [15:18] ericsnow, good. so if the lib's licence is god we can rm the line [15:18] * mgz giggles at god [15:18] mgz: do you recall what the licensing issue was? === kadams54 is now known as kadams54-away [15:19] ericsnow: I think that was the go.net package test data === kadams54-away is now known as kadams54 [15:19] so, not this one [15:20] mgz: oh good, so once the deps mechanism is fixed in CI I should be able to try merging again :) [15:24] mgz: what is the ETA on that? I recognize you probably have higher priorities. :) [15:25] ericsnow: I'll try to do it before your eod so you can try again [15:25] ericsnow, just a few minutes to remove a line and update the machine [15:25] mgz: thanks! [15:25] ah, yeah, and the unblocking of you r branhc is even easier, because it doesn't need the script fix [15:25] just hte blacklist remove [15:26] sinzui: shall I do it, or are you? [15:26] I will. I have it open [15:34] cherylj: ping [15:35] katco: pong [15:35] cherylj: got a few minutes to chat? i would love to meet you [15:35] Give me just a minute [15:35] cherylj: no rush at all [15:57] ericsnow: did you see smosers response in cloud-init? [15:57] ericsnow: I'm going to add that information to the doc [15:57] wwitzel3: not yet :) [16:07] niemeyer, hey, just a reminder - I'm in the hangout now :) [16:09] perrito666, wwitzel3: standup? [16:09] katco, hey, as you've contributed to goamz in the past, are you willing to give a hand from time to time with reviews and maintenance? [16:11] ericsnow, I let you down. I got distracted and didn't complete the deployment to the server. I am doing it now. poke in 15 minutes if I haven't confirmed it is in place [16:12] sinzui: no worries; I'm itching to land the patch, but the only thing it blocks is the new GCE provider (which won't be landing yet) :) [16:12] sinzui: thanks for taking the time to sort this out [16:13] ericsnow: wwitzel3 having some connectivity issues, sorry Ill be there in 10 mins [16:13] perrito666: k [16:19] ericsnow, you can resubmit [16:19] sinzui: thanks! [16:21] ericsnow: wwitzel3 I need to sort a couple of things before the standup, can we do-it in half an hour? [16:21] perrito666: fine with me [16:21] thanks a lot [16:21] wwitzel3: ? [16:22] I wasn't even paying attention, sure [16:22] lol [16:22] :) [16:22] I'm in deep thought, I will be in seclusion in the west wing [16:36] I'm having some trouble with state/ConnSuite. it seems that the charm URL isn't found on the testing units of a testing service [16:36] if anyone can enlighten me, it would be much appreciated! [16:42] dimitern: hey sorry i was on a call. certainly i can, as long as i'm not too pressed for time with other juju priorities :) [16:45] katco, awesome, will send you an invite to join! I promise not to pester you too much :) [16:45] dimitern: :) [16:56] ericsnow: I can see you [16:56] :p [16:56] perrito666: haha [16:57] wwitzel3: can I get you to surface from the depth of your thoughts for a moment? [17:21] fwereade: ping me when you have a moment please :p [17:24] bodie_, a unit's charm url gets set by the uniter when it's actually downloaded a charm, it doesn't otherwise have one [17:24] bodie_, what were you expecting to use it for? [17:25] fwereade, I'm set. didn't realize units weren't created with the charm URL of the parent service [17:25] fwereade, was just discovering that for myself :) [17:26] bodie_, what would you be using it for? [17:26] fwereade: hey when you have a moment, do you have time for a hangout? [17:26] katco, might be able to manage one later, it's a bit loud and approaching-suppery here at the moment [17:27] fwereade: no worries. it can be tomorrow as well... i think i'll likely spend the day doing reviews [17:27] katco, I'll ping you if/when I'm free later, otherwise grab me tomorrow :) [17:27] fwereade: sounds like a plan, and happy new years :) [17:29] fwereade, when an Action is enqueued on a unit, it should validate the params before enqueuing (sp?) it. that way the user immediately gets an error when invoking it via `juju do`. [17:29] s/enqueuing it/adding it to the actual queue on state/ [17:30] otherwise, the user must discover ID10T error via `juju fetch` [17:30] fwereade, if a unit doesn't have a charm URL, then the unit's charm can't be found, AFAICT, meaning it can't validate Actions passed to it [17:31] bodie_, it's certainly rational to fall back to the service's charm [17:31] bodie_, might even be sane to just always validate against the service charm url [17:33] bodie_, not clear which approach will fail more often [17:34] bodie_, but it's also important to keep perspective, the window is fairly narrow: running actions on a service while upgrading that service [17:34] fwereade, seems like if a unit is in the process of being added, it would be better to validate against the service at time of enqueueing [17:34] bodie_, certainly it's a good fallback when no charm url has been set [17:35] fwereade, I think the worker will catch the case you sketched up [17:35] so, the action would be enqueued but fail [17:35] bodie_, yeah, that's fine [17:37] fwereade, thanks :) === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [18:46] g'night all [18:46] EOD === kadams54_ is now known as kadams54-away === kadams54-away is now known as kadams54_ === kadams54_ is now known as kadams54-away [20:21] I feel that today is one of those "start the day with two coffees" types of day [20:24] thumper: I feel you on that. I had to start the day off with a 4 pack of Monster Energy drinks. [20:27] ericsnow: I am going to take some time to review the cloudsigma stuff and then I'll join you back on GCE stuff. [20:27] wwitzel3: sounds good [20:27] ericsnow: feel free to start on any TODOs or refactors or whatever and we can just review [20:27] ericsnow: I feel much better after eating ;) [20:28] wwitzel3: I landed the PortSet patch so now I'm working on merging our branch with master === kadams54-away is now known as kadams54_ [20:28] wwitzel3: :) [20:28] ericsnow: ahh, ok, good idea, been a while since we did that [20:29] wwitzel3: I'm doing a merge instead of a rebase since I think the history is worth keeping [20:29] ericsnow: +1 [20:43] hmm [21:27] anastasiamac: how close are you to landing your annotations work? [21:27] anastasiamac: I have to make some changes which will conflict with yours [21:28] menn0: hopefully in the next 4-5 hrs? [21:28] anastasiamac: hmmm... mine might be quicker than that (and certainly a lot smaller) [21:28] anastasiamac: let's keep each other updated I guess [21:29] menn0: k. thnx [21:34] wallyworld: ill walk my dog and then be back so we can call [21:35] perrito666: great, thanks, was about to ping you [21:35] I could do now but my dog seems to be much more urged than you and is more likely to express her unhappiness with bites :p [21:35] sure :-) [21:44] katco: nice review on the cloudsigma stuff, thank you [21:44] wwitzel3: oh no problem [21:44] wwitzel3: i had to stop part-way through though; that was going to take all day [21:45] katco: yep, I'm picking up where you left off :) [21:46] wwitzel3: oh, thanks :) i thought maybe they could extrapolate a little [21:46] katco: I've never successfully done a full review of this code in a single day .. week [21:46] wwitzel3: wow yipes [21:47] katco: yeah, there are some branches that are broken out, but their latest updates touched all of them [21:47] katco: so it ended up with a big single patch on reviewboard [21:47] wwitzel3: ah i see [21:47] wwitzel3: i'm happy for their contribution, just needs a bit of standards compliance [21:48] katco: ericsnow and I are going to have the same issue when we propose GCE .. figuring out how to make it not a huge monolithic review [21:48] wwitzel3: i think it's a real challenge [21:48] wwitzel3: i always like to have everything working before i submit anything [21:49] katco: you want to give enough context to the review .. but then how much is too much [21:49] wwitzel3: right [21:49] katco: we've tried to learn from some of the abstraction leaks of other providers so that we cen propose it in useful pieces, but it is still hard. [21:50] wwitzel3: i would be really interested if someone wanted to do a hangout or writeup where they demonstrate a methodology for pipelining changes; i.e.: finish everything up, and then break it up after the fact [21:51] katco: yeah, that would be useful. Maybe eric and I can do something like that if we manage to do it successfully :) [21:51] wwitzel3: i have faith! :) [22:20] ericsnow: took your idea and sent a mail to dev, linked our branch too .. figure that way if though it isn't up for review yet, people can start poking around. [22:20] ericsnow: also you want help with the merge? how is that going? [22:21] wwitzel3: it was surprisingly easy [22:21] ericsnow: well thats good :) [22:25] ericsnow: I like that rename of gce -> raw [22:25] wwitzel3: :) [22:26] wwitzel3: I'm trying out moving gce.go into it's own package ("gceapi") [22:26] wwitzel3: so far it makes for a clearer picture [22:27] ericsnow: slightly confusing, but yeah, worth try [22:27] ericsnow: since, google-api-go-client is the gceapi [22:27] ericsnow: our's is the juju-gce-api .. [22:28] wwitzel3: I'm almost done; we can look it over when you're free [22:28] ericsnow: sounds good [22:28] ericsnow: now is good :) [22:28] wwitzel3: brt [22:41] anastasiamac: it turns out those annotation changes I was going to make probably aren't necessary [22:41] anastasiamac: no need to worry about conflicts === kadams54 is now known as kadams54-away [23:12] wallyworld: sorry for the delay, I am back [23:13] perrito666: np, i have ameeting in15 mins, let's have a quick chat, i'll set up a hangout [23:13] davecheney: ping [23:13] perrito666: https://plus.google.com/hangouts/_/canonical.com/tanzanite-stand [23:28] menn0: that's gr8 news! thnx ;-) === marcoceppi_ is now known as marcoceppi [23:44] katco: ack [23:50] wwitzel3, ericsnow kudos on the gce provider. i've got some gce credits to burn, looking forward to trying it out! [23:50] cmars: yeah, we're pretty excited with it :)