[00:57] wallyworld: would you kindly create me a feature branch for azure? [00:57] sure [00:59] axw: done [00:59] wallyworld: thanks [01:28] wallyworld: just once... [01:28] wallyworld: haha RWC 2015 to the All Blacks!!! [01:28] * thumper is done now [01:28] sigh [01:28] i even hopped up at 2am to watch :-( [01:28] Got the family up early to watch it in a pub in Queenstown [01:29] getting the girls up a 4:30am was interesting [01:29] worth it though [01:29] yeah, pub was packed, even one guy with a wallabies jersey on [01:29] poor bastard [01:29] yeah, I thought he was pretty game [01:29] especially if we had won [01:30] heh [01:30] TBH though, I did think that the All Blacks were the better team on the day [01:30] thumper: wasnt it yesterday? did u celebrate this whole time? [01:30] the frustrating times [01:31] are when you feel your team played better but gets pipped at the post [01:31] anastasiamac: it was Sunday morning [01:31] anastasiamac: but I was away for a wedding over the weekend, and was driving home yesterday [01:32] thumper: i watched haka - that's as far as I go :D [01:32] oh and richie mccaw [01:32] anastasiamac: :) [01:50] thumper: do you have an eta on removing mult env feature flag in master? [01:51] or shoudl i say, mrging in the feature branch? [01:51] wallyworld: wat? [01:51] feature branch was merged for 1.25 [01:51] flag being removed for 1.26 [01:53] thumper: yeah, sorry, what you said above is what i meant [01:54] eta for flag removal from master? [01:54] soon [01:54] perhaps this week [01:54] I'm actively working on it [01:55] yay, ok [01:55] i'll take advantage of that for cross mdel stuff [01:56] * thumper nods [01:56] * thumper is heading away from internet [01:57] I have to take Maia to jui jitsu, taking laptop and working from the gym while she does her thing [01:57] soon [02:57] wallyworld: do you know what hotel we're staying at in oakland/ [03:04] natefinch: is it not Marriott Oakland City Center? [03:04] cmars: ping? [03:08] anastasiamac: I don't know. Did it say that somewhere? [03:09] it's on the spreadsheet, hotel tab [03:09] anastasiamac: arg... tabs in the spreadsheet. Honestly, who uses a spreadsheet to disseminate long-form information :/ [03:09] :D [03:09] natefinch: i hear ur pain \o/ [03:10] anastasiamac: I must have looked at that spreadsheet three times and never saw the tabs :/ [03:10] spreadsheets are for making lists of things, duh [03:11] word documents are for distributing screenshots [03:11] *shudder* [03:11] :P [03:12] trying to figure out if I should just fly into SFO and take a cab, rather than trying to get to oakland airport, which takes 2 hours extra and an extra layover [03:14] looks like it's 15 minutes from Oakland airport and 30 from SFO [03:14] it depends - do u like flying? [03:14] u r very lucky to have a choice :D [03:16] i'd imagine flying to Oakland will also be a smaller plane than to SFO (m not gr8 fun of small planes) [03:16] natefinch: marriott [03:16] 1001 broadway [03:18] wallyworld: thanks. realized I was being dumb and missed the hotels tab on the spreadsheet. Someday I'll remember that's there. [03:19] yeah, easy to do [03:24] axw: ping? [03:24] cherylj: pong [03:25] hey axw :) I'm looking at bug 1512399 and it looks like the error they're getting is a TLS error [03:25] Bug #1512399: ERROR environment destruction failed: destroying storage: listing volumes: Get https://x.x.x.x:8776/v2//volumes/detail: local error: record overflow [03:25] and I'm wondering if it's a config error? or if you've seen that in your testing? [03:27] cherylj: I have not seen that. we have a separate API for Cinder, so it's most likely we're not configuring the client for it the same way we do the other clients [03:28] axw: ok, thanks! I'll keep looking [04:20] running the juju/state tets [04:21] i can see mongo using 1/2 a core [04:21] that seems excessive [04:21] can anyone confirm ? [04:22] select(9, [7 8], NULL, NULL, {0, 10000}) = 0 (Timeout) [04:22] billions and billions of these [04:23] davechen1y: yeah, usually 40-50% of a core for me most of the time, except when it jumps to the occasionaly 180% [04:23] Bug #1512566 changed: Panic in deployerSuite unittests while attempting to connect to mongo [04:25] i wonder if mongo 3.0 will be better [04:26] not sure if a faster car will make towing the giant boulder behind us much easier [04:26] Bug #1512566 opened: Panic in deployerSuite unittests while attempting to connect to mongo [04:27] ....or some other analogy saying our tests suck [04:30] any "database" that consumes 50% cpu with a single client [04:30] owns the market on suckage [04:34] oh fucking golf cla [04:34] now mongo used up all my /tmp [04:34] all 8gb [04:41] Bug #1512566 opened: Panic in deployerSuite unittests while attempting to connect to mongo [04:41] Bug #1512569 opened: UniterSuite.TestRebootNowKillsHook fails with: uniter still alive [04:44] Bug #1512569 changed: UniterSuite.TestRebootNowKillsHook fails with: uniter still alive [04:47] Bug #1512569 opened: UniterSuite.TestRebootNowKillsHook fails with: uniter still alive [09:07] axw, http://reviews.vapour.ws/r/3045/ may be relevant to your interests [09:08] anyone else, http://reviews.vapour.ws/r/3032/ is a few days old and could use a review (and should be trivial) [09:14] fwereade_: looking at the second [09:14] voidspace, ta [09:15] waiting for an environment to bootstrap anyway [09:18] fwereade_: my brain is a bit fried, will take a look tomorrow. skimming through it looks good. NewWorker in config bothers me, but will have to ruminate [09:19] axw, yeah, that's the main thing I'm ambivalent about [09:19] axw, I can't think of a good reason not to do it, though :-/ [09:19] axw, (NewFacde less bothersome?) [09:20] axw, (I feel they're as bad as one another) [09:20] axw, anyway don't think about it now :) [09:20] fwereade_: a little :) NewFacade bothers me less because you're configuring an external resource [09:20] fwereade_: whereas the worker is an internal detail [09:20] fwereade_: I guess neither would bother me if the manifold were in a different package [09:21] axw, yeah, I've been feeling some of them pulling that way as well [09:21] axw, and I don't think a given worker type will necessarily always want the same manifold [09:22] axw, but for now, it seems least wrong to put them with the workers [09:22] will think more on it [09:29] fwereade_: LGTM bar one typo [09:33] voidspace, cheers [10:22] http://reviews.vapour.ws/r/3045/diff/# [11:49] frobware: looks like I an upgrade from 1.22.8 to 1.24 tip with upload-tools [11:49] which is good news [11:49] voidspace, yep [12:19] my NUCs et al have just arrived! [12:28] morning [12:29] dimitern: I envy you a lot right now [12:29] perrito666, :) don't - considering the amount of work we need to do in the next 3 weeks on those NUCs [12:51] go go gadget dimitern [12:51] rick_h__, :) indeed [12:51] rick_h__, nice ODS talk btw! [12:51] dimitern: thanks [13:03] rick_h__: are you around for the meetup ? [13:03] frobware, oh sorry, omw [13:05] jam: yes sorry [13:09] dimitern: is the maas meeting on today? [13:18] voidspace, yep [13:30] dimitern: thanks [13:58] frobware: I was wrong by the way - 1.22.8 to 1.24 tip fails. It just fails without error message. [14:07] why is master still blocked? is CI running on it yet? [14:19] Bug #1512718 opened: Leader election fails with "leadership failure: leadership manager stopped" [14:20] ericsnow: did we update the version of reviewboard we're using? Seems like it handles file moves a lot better than it did just a week or so ago. === mwenning is now known as mwenning-wfh [14:38] cherylj, ping - does your vMAAS setup have a static IP range set for the cluster controller / interface? [14:45] natefinch: nope :) [14:48] ericsnow: weird, I swear the last time I looked at your renaming review, it was all screwed up because reviewboard just showed renames as deletes and adds, but now it looks fine. Maybe I'm just crazy. [14:49] natefinch: rb is a sentient being and tries to make you doubt your sanity [14:49] perrito666: I can believe it [14:49] off course you can, it is true [15:07] frobware: I have it set up to use a particular virtual network I created [15:07] does anyone know exactly what is responsible for ~ expansion? I am trying to do an exec.Command with ~ on a path and it is not being expanded [15:08] perrito666: it's the shell that does it not the os [15:09] voidspace: mm, I presume that for this to work the process I should be running with exec.Command its bash [15:09] perrito666: that would work, but beware of windows [15:09] perrito666: or you can do the expansion yourself [15:10] I could totally wear a shirt with that "beware of windows" [15:10] voidspace: true [15:10] :-) [15:18] trivial change to environs/config, so that I don't need to do an error string comparison... anyone want an easy review? http://reviews.vapour.ws/r/3047/ [15:39] why is master still blocked? is CI running on it yet? [15:40] sinzui, mgz_ ^^ [15:41] cmars: ci hasn't tested a pass yet. I believe master will be retested in a few hours. [15:41] sinzui, ok, thanks [15:52] cherylj, do you have time for a quick HO? [15:52] frobware: I have 8 minutes :) is that enough time? [15:52] yep [15:53] cherylj, https://plus.google.com/hangouts/_/canonical.com/juju-sapphire [15:53] I'm there [16:43] Bug #1512782 opened: wget cert issues causing failure to create containers [16:46] Bug #1512782 changed: wget cert issues causing failure to create containers [17:01] Bug #1512782 opened: wget cert issues causing failure to create containers [17:15] mgz_: any chance we might get a CI run on master today? [17:16] rogpeppe: it'll happen shortly [17:16] mgz_: cool [17:16] and then the stampede to get stuff landed before it breaks again [17:17] mgz_: istm that 24 hours to unblock after critical bugfix gets landed is probably too long [17:17] ^ +1 [17:18] for our feature branch it was more like 7 days. [17:19] mgz_: is it a resources problem? i.e. do we not have an environment dedicated to running CI on master? [17:19] natefinch: every ci run is against a bunch of environments [17:20] mgz_: sorry, overloaded term. Do we not have dedicated hardware for CI runs on master? [17:20] mgz_: maybe a better way to phrase it is - it sounds like CI on master can be blocked by other CI runs. Is there a way we can make that not so? [17:20] well, we're still talking about a whole bunch of machines in different cloud provuders [17:21] so, two things [17:21] we're getting new hardware for maas Soon hopefully, which is a major bottleneck [17:21] good [17:22] also abentley has been getting parallel streams working over the past few weeks, which means we can run more than one revision in across clouds [17:25] anyway, the reson this was slow is that the rev was tested and failed (for unrelated env issues, machine ran out of root disk), [17:26] so needed someone to wake up and get it run again, any aussie could have done that [17:26] obviously we need to hire an aussie for the QA team [17:36] mgz_: hey you all had started writing ci tests for the rackspace provider, correct? [17:36] katco: there's not really tests to write [17:37] we just need rackspace creds we can use, and then we can use the existing provider tests [17:37] mgz_: ah great! is anything blocking you on getting those creds? [17:37] well, I have a non-functional personal account, [17:38] ideally we'd have a system of getting creds as a company, but that's never really happened in the past === Ursinha is now known as Ursinha-afk [17:38] mgz_: i'll ping alexisb to see what we need to do [17:43] mgz_: i also need to bug someone about what it would take to get CI testing the LXD provider [19:25] Bug # changed: 1302118, 1392814, 1464304, 1485784 [19:25] Bug #1512847 opened: Unit suffix incremented after fresh post-destroy-service deploy [19:33] haha [19:33] bug 1512847 now complaining about our "correct" behaviour [19:33] Bug #1512847: Unit suffix incremented after fresh post-destroy-service deploy [19:34] * thumper wonders what the original bug number was that outlined the behaviour.... [19:38] Bug #1512847 changed: Unit suffix incremented after fresh post-destroy-service deploy [19:38] Bug # opened: 1302118, 1392814, 1464304, 1485784 [19:44] Bug # changed: 1302118, 1357045, 1392814, 1464304, 1485784, 1494476, 1498982, 1500703, 1500721, 1500769, 1500803, 1501173, 1501559, 1501637, 1501642, 1501710, 1505648 [19:44] Bug #1512847 opened: Unit suffix incremented after fresh post-destroy-service deploy [19:53] Bug # opened: 1357045, 1494476, 1498982, 1500703, 1500721, 1500769, 1500803, 1501173, 1501559, 1501637, 1501642, 1501710, 1505648 [19:56] Bug # changed: 1357045, 1494476, 1498982, 1500703, 1500721, 1500769, 1500803, 1501173, 1501559, 1501637, 1501642, 1501710, 1505648 [20:24] sinzui: master is still blocked? [20:24] it is the same bug as yesterday... [20:25] yes, thumper master is being retested now [20:25] why has it taken over 24 hours? just curious [20:38] thumper: https://github.com/golang/go/issues/13137 [20:38] upstream issue for the ssh blocker that prevents me moving the ssh code to juju/utils [20:39] thumper: i also tried adding dependencies.tsv to juju/utils to pin the ssh version [20:39] by our build bot doesn't understand that [20:40] thumper: I asked the same question earlier, sounds like the build machine ran out of disk space, so needed manual intervention [21:01] ah [21:12] * natefinch needs to write a bot that'll watch master's CI status and auto-$$merge$$ his PR when master unblocks. [21:17] Bug #1512875 opened: juju 1.25.0 using MAAS 1.9-beta2 juju incorrectly reports the private address [21:26] Bug #1512875 changed: juju 1.25.0 using MAAS 1.9-beta2 juju incorrectly reports the private address [21:28] wallyworld: can you review http://reviews.vapour.ws/r/3050/ [21:28] sure [21:29] another difficult one [21:29] Bug #1512875 opened: juju 1.25.0 using MAAS 1.9-beta2 juju incorrectly reports the private address [21:32] Bug #1512847 changed: Unit suffix incremented after fresh post-destroy-service deploy [21:56] Bug #1367962 changed: HTML entities accidentally escaped === natefinch is now known as natefinch-afk [22:26] Bug # changed: 1439375, 1456851, 1476918, 1486254, 1492095, 1497241, 1500283, 1502202, 1506353, 1509097, 1509292 [22:29] Bug # opened: 1439375, 1456851, 1476918, 1486254, 1492095, 1497241, 1500283, 1502202, 1506353, 1509097, 1509292 [22:38] Bug # changed: 1439375, 1456851, 1476918, 1486254, 1492095, 1497241, 1500283, 1502202, 1506353, 1509097, 1509292 [22:39] cmars: ping :D [23:14] Bug #1511717 changed: Incompatible cookie format change [23:17] Bug #1511717 opened: Incompatible cookie format change [23:23] Bug #1511717 changed: Incompatible cookie format change [23:25] anastasiamac: send me the song name [23:26] thumper: upstream told us to go stuffed on the ssh problem [23:26] thumper: where did the go 1.5 discussion go after we talked yesterday ? [23:27] perrito666: tyvm for reminder :D i will as soon as my DJ gets back to me [23:30] davecheney: sorry, dropped the ball on that email, will do it now [23:34] thumper: thanks [23:34] what's going on with the build [23:34] is it unblocked uet ? [23:34] oh goody it is [23:34] waigani: I've just created a feature branch "controller-rename" [23:34] which I'll retarget my landings against [23:34] thumper: that's a good idea [23:35] I have a feeling we may break some CI tests :) [23:35] so best to get it in a feature branch IMO [23:35] especially taking off the feature flag [23:35] that will help me land this destroy*System* work too [23:35] as I can just finish off the logic and then merge in the controller stuff [23:36] waigani: also means we can do final checks to make sure we catch all the system names [23:36] right [23:36] waigani: probably best to see if we can get your stuff landed first [23:36] * thumper hopes [23:36] okay [23:36] on it :) [23:37] sinzui: ping [23:37] sinzui: i have a problem trying to land https://github.com/juju/utils/pull/167 [23:37] davecheney: did you read my comment? [23:37] the build bot doesn't seem to respect dependencies.txt [23:38] mgz_: thanks for replying [23:38] i added a dependencies.tsv to pin the crypto repo to a specific revision [23:38] but the bot doesn't appear to use it [23:38] davecheney: you need to add *all* the dependencies [23:38] ok [23:38] but how does the bot know ? [23:38] I can make the bot get the deps with godeps [23:38] i can add all the deps to that PR [23:39] can you switch the bot to using godeps please ? [23:39] but it's either go get *or* godeps, we can't mix and match [23:39] understood [23:39] i'll ammend the PR to have a complete set of dependencies [23:39] so, I can switch the bot, and the branch will still fail for now [23:39] k [23:39] give me a few mins [23:39] won't take long [23:40] this is baiscally, in the config of the github-merge-juju-utils job (access is in the consoles.txt file in cloud-city), [23:41] s/--go-get-all/--tsv-path $PROJECT_PATH/dependencies.tsv/ [23:42] mkay [23:42] anyway, retry when ready [23:43] mgz_: thanks, i'll be able to fix this on my own if I find deps I haven't pinned [23:52] master be all like: https://i.giphy.com/11Ss9q3F4zf5VC.gif [23:52] anastasiamac, pong [23:53] cmars: send u an email but if u want to HO, can do too :D [23:54] anastasiamac, right, blueprint. i'll try to update it soon [23:54] cmars: \o/ [23:56] cmars: love the pic btw :D indeed feels like when it unblocks :P