[00:02] mwhudson: /me looks [00:03] hmm, it was working fine up til last night [00:03] * davecheney steps into phone box, enables vpn [00:04] davecheney: i thought u were stepping into phone booth to change into ur super outfit... [00:04] anastasiamac_: it's the same thing [00:04] mwhudson: looks like the builder _is_ running [00:05] davecheney: hm [00:05] i'll keep an eye on it [00:05] the proxy might have screwed up [00:05] and migth be throwing away the calls to the dashboard that report results [00:06] yummy proxies [00:08] mwhudson: you saw the note about go 1.5 last night ? [00:08] i'd say it'll ship in august [00:08] well [00:08] before september :) [00:11] davecheney: yeah [00:12] mwhudson: how much of a problem will that be ? [00:12] don't know [00:12] should I be talking to tim and alexis now [00:12] and issuing dire warnings ? [00:13] argh i've forgotten when feature freeze is [00:13] aug 20 [00:13] hmm [00:14] alpha 2 was yesterday [00:14] i guess that's getting close [00:14] davecheney: it might be worth getting alexis to talk to steve l about dates, yeah [00:14] i can see the timeline being an issue [00:16] yeah, it's getting to 'uneasy' i think, a few notches below 'panic' or 'give up and start mowing lawns' [00:20] mwhudson: done [00:20] i've been here before and excessive planning and overcommunication is the path to success [00:21] definitely [00:21] talking of planning, i'll see about doing another rebuild test using real go for arm64 and ppc64le [00:49] mwhudson: what was that site you used that decoded instructions ? [00:49] davecheney: https://www.onlinedisassembler.com/odaweb/ [02:04] davecheney: i like the new contender for most useless bug reporter ever [02:04] (https://github.com/golang/go/issues/11957) === natefinch-afk is now known as natefinch [02:04] ahh the old "it doesn't work" bug [02:05] lol [02:05] insane amounts indeed [02:05] perhaps not as quite as useless as the "can't bootstrap with gccgo" guy [02:06] It's amazing how even people who work in development can fail to post even the most basic information about a bug. I can't tell you how often I've had to tell even the same people in an org inside my own company to include the damn logs when reporting a bug. It's not rocket science. [02:08] Gah, people, if you're going to have a field which is a map[string]someStruct ... you gotta explain WTF the string is [02:09] yup, [02:09] ...this is why I hate using maps for things which should probably just be slices... because the most important field in the whole struct *isn't in the struct* [02:10] natefinch: yup [03:24] lol @ omitempty on non-pointer struct fields [04:04] natefinch: why? omitempty works for non-pointer values [04:07] axw: for non-pointer struct values? I was just seeing it not work [04:08] natefinch: http://play.golang.org/p/Q9C5yKHhqi [04:09] axw: I mean this: http://play.golang.org/p/rVw_82ZOl8 [04:10] natefinch: ah, fields of struct type. gotcha. [04:10] right, sorry, didn't explain it well [08:21] davecheney: i take it back, i think this might be the most clueless reporter ever [08:51] jam, thumper if you guys are watching your screens we could use you upstairs [09:29] hey mgz [09:30] mgz: the 1.24 branch didn't get blessed due to an intermittent failure on the i386 tests [09:30] mgz: can you just run them again to check? [09:33] mgz: failing that, we need someone to pick this up [09:34] dimitern: can you please keep on top of this? [09:34] it is super important [09:37] thumper, ok, will see what I can do [09:37] dimitern: cheers [10:08] thumper, by "super important" I guess you won't mind skipping that test for i386 for the time being - it looks like a timeout issue [10:09] dimitern: if it is obviously broken... [10:09] then we should not run it [10:10] I'm ok with skipping it on "i386" only if we can do that [10:10] as long as we file a bug [10:10] thumper, yes we can :) [10:10] hopefully this test will be rewritten as part of the uniter sprint upcoming [10:10] thumper, my point exactly, yeah [10:12] dimitern: do it! [10:22] fwereade: how great is this, we already have james flagged as the "C" in RACI for min version [10:29] thumper, done - stamp it! http://reviews.vapour.ws/r/2286/ [10:30] dimitern: stamped [10:31] later folks, lunch time here [10:31] then we're done [10:31] thumper, thanks and enjoy :) [11:44] hey dimitern [11:44] dimitern: you have been volunteered to make sure 1.24 gets blessed :) [11:44] we are all done here and people are stopping looking at laptops [11:44] I'm hoping that it will all be good [11:44] the only blocker would be another intermittent failure [11:44] fingers crossed, we're good [11:44] thumper, oh is that right :) [11:45] thumper, I'll keep an eye on the builds and status [11:45] but someone needs to hold the token :) [11:45] if you hit EOD, make sure someone in the US gets the token [11:45] thumper, hopefully should be fine and I'd nudge it if needed [11:45] sound good? [11:45] thumper, yes, wilco ;) [11:45] awesome [11:46] see y'all next week [11:46] safe travels! [11:46] * thumper closes laptop [13:21] Bug #1480298 opened: unknown object type "Charms"

[13:57] dimitern: in your review when I'm registering the watcher you mentioned a possible failure detection and watcher stopping. how to detect an error during registering? or did you mean when above the mapping returns an error? [13:57] Bug #1480310 opened: dbus link request failed for service FOO: Unit name FOO is not valid.

=== natefinch is now known as natefinch-afk [15:43] https://github.com/juju/juju/blob/master/container/kvm/kvm.go#L158 -- This line does not print the constraints [15:43] machine-0: 2015-07-31 15:36:54 TRACE juju.container.kvm kvm.go:158 create the container, constraints: [16:02] I think it's because %+v outputs multiple lines? [16:02] er, %v [16:02] does it? what is the ouput you get there? [16:03] 08:43:32 machine-0: 2015-07-31 15:36:54 TRACE juju.container.kvm kvm.go:158 create the container, constraints: [16:03] the next line is the next log message, and the kvm container is being created with the default constraints [16:03] instead of the ones I passed. [16:03] duh, I am so used to another channel that I ignore all the lines next to links [16:05] I'm debugging the issue by doing: logger.Debugf("Params passing to container. Arch: %s : Memory: %d : Cores: %d : RootDisk: %d", startParams.Arch, startParams.Memory, startParams.CpuCores, startParams.RootDisk) [16:05] before it parses the hardware [16:08] https://bugs.launchpad.net/juju-core/+bug/1399613 [16:08] Bug #1399613: juju-core not using constraints when creating KVM unit on maas machine

[16:10] I've tested the issue with 1.22 through current 1.25 alpha [16:15] http://paste.ubuntu.com/11974200/ [16:16] juju machine add kvm:0 --constraints "cpu-cores=2 mem=2G root-disk=20G" [16:16] is what I'm passing [16:34] anyone want to help me in crafting an error message? I think I'm over thinking things and trying to make things way too complicated. [16:34] it's for bug 1480298 [16:34] Bug #1480298: unknown object type "Charms"

[16:35] basically, we're using a new juju client with an older version of juju [16:35] and we try to register the charm for metering, and can't since it's not supported in earlier versions of juju [16:37] What I have is: "current state server version does not support charm metering" as a warning, but don't return an error. [16:37] thoughts? [16:37] mattyw ^^ (if you're still around) [16:38] cherylj, hey hey [16:56] the maas provider doesn't call CreateContainer? [16:57] * bhundven is stumped [17:00] ah, it uses instanceBroker [17:52] Bug #1479931 changed: Juju 1.22.6 cannot upgrade to 1.24.3/1.24.4

[19:09] Well, the instanceConfig passed to CreateMachine in containers/kvm/kvm.go doesn't contain the constraints. [19:10] http://paste.ubuntu.com/11975150/ [19:11] http://paste.ubuntu.com/11975152/ [19:17] iow, instanceConfig.Constraints doesn't exist [19:58] cherylj: ericsnow wwitzel3 : this befuddles me. I cannot find a dupe, but it looks familar. Is this a new bug? Do we need to fix this in 1.25? https://bugs.launchpad.net/juju-core/+bug/1478762 [19:58] Bug #1478762: juju bootstrap failed to connect to environment with error "discarding API open error" [19:58] sinzui: doesn't sound familiar [20:01] natefinch-afk: status update on our convo - i have the branch built and living locally in a container. I'm looking at way to do this that would apply to any feature branches we want to track, and get those added as nightly builds, and yield a container that isnt 2gb - should have something for you by early next week say tuesdayish === natefinch-afk is now known as natefinch [20:05] lazyPower: cool [20:07] lazyPower: I think sinzui can point you to binaries created by the CI system that could easily be just copied wherever you need them. [20:07] so our CI is already building/archiving branches? [20:08] lazyPower: Ci tests every revision, the binaries built are places in s3, and reports.vapourr.ws can list them for you [20:08] lazyPower: are you interested in 1.24.4 or 1.25-alpha1? [20:09] sinzui: feature-proc-mgmt branch [20:09] lazyPower: http://reports.vapour.ws/releases lists that branch and the last build tested (2928) [20:10] lazyPower: http://reports.vapour.ws/releases/2928 has a link to binaries (faster than scanning the jobs)...http://reports.vapour.ws/releases/2928/binaries [20:17] sinzui: it looks familiar to me too. I can try to find if there is a dup, or if we're both going crazy [20:18] cherylj: maybe I saw issues like this because of maas setup, and the bug was moved from juju-cire to maas? [20:19] sinzui: I see bug 1477920 which was marked as a dup of 1321212 [20:19] Bug #1477920: juju bootstrap failed with "discarding API open error"

[20:22] sinzui: but I thought that thumper had done some work around that to retry the connection while bringing up the bootstrap node... hmmm [20:22] maybe I'm going crazy [20:29] cmars: could you take a look: http://reviews.vapour.ws/r/2287/ [21:07] Well, I'm not sure what to do. It seems that this (https://code.launchpad.net/~thumper/juju-core/kvm-constraints/+merge/197815) was an attempt to fix constraints passing to kvm containers. Without being able to set constraints on the kvm I need to deploy, I can't move forward with my evaluation of juju. I've tried to figure it out on my own. \o/ [21:11] * bhundven stops talking to rocks... [21:48] * perrito666 decides to test a worker with unittests and see if someone gets mad