[00:00] thumper: yep naked send [00:00] * menn0 checks master [00:01] thumper: fixed in master but not 1.24 [00:01] * menn0 backports [00:42] cmars: are you going to merge your branch? [00:42] thumper, i'd like to [00:42] thumper, master's blocked [00:43] I've committed a fix [00:43] cmars: please JFDI it [00:43] thumper, with pleasure [00:51] menn0: thanks for the backport [00:52] thumper: np. hopefully my PR will make it through once this is in [01:09] cherylj: still there? [01:12] wallyworld: I'm having a lateish lunch with Rachel [01:12] wallyworld: will probably be late for our 1:1 [01:12] sure, np [01:29] evening all [01:39] man I really hate linux utilities' default of "don't tell the user wtf is going on" [01:46] wallyworld: back [01:46] wallyworld: good now? [01:47] sure [02:00] axw: just finishing up, be there in a sec [02:00] wallyworld: thanks for reminding me ;) [02:12] cmars: http://reports.vapour.ws/releases/2890/job/run-unit-tests-win2012-amd64/attempt/908 [02:12] cmars: another intermittent metics collection failure [02:13] * thumper sighs [02:13] cmars: I'm wondering if all the extra tests in uniter also pushed the ppc timeout over 10 minutes [02:20] * thumper is looking at http://reports.vapour.ws/releases and feeling sad [02:31] menn0_: got a minute to talk through another intermittent failure? [02:31] thumper: yep [02:31] menn0_: standup hangout [02:32] menn0_: click "join" :) [02:33] thumper: it's not that. I have to re-auth and it wants a new password [02:33] ah [02:50] anyone seen a gccgo error where it thinks you've multiply defined things that are not multiply defined? [02:53] thumper: ^ ? I seem to remember a bug we've hit with gccgo in the past where like embedding things the wrong way will make gccgo mad. Do you happen to remember what caused that? [02:56] nvm, pretty sure it's this: https://github.com/golang/go/issues/7627 which is fixed in go 1.3 .... which does not help us [03:04] decisions, decisions [03:05] wanting a new laptop bag [03:05] trying to decide between black or red [03:05] if not-black is an option you like, go with not-black. There's too much black in everything tech [03:07] * natefinch looks at his black keyboard, black monitor stands, black headset, black camera strap, black camera, black CD player... all just on his desk === axw_ is now known as axw [03:54] * thumper chose black [03:55] it was pointed out how grubby red would look when a little dirty === Spads_ is now known as Spads === TheRealMue is now known as TheMue === ev_ is now known as ev [05:40] ok, I'm done [05:40] drink time [05:40] laters folks [07:05] Bug #1475509 opened: upgrade-charm --force behavior causes races === menn0_ is now known as menn0 [09:04] dooferlad: hangout, and do you know where dimitern is? [09:04] TheMue: No, I don't. I think we should wait for him. [09:04] here I am [09:04] ah [09:04] sorry guys :) [09:04] hehe [09:04] no worries. Joining [10:06] tgif :) why do meteors always land in craters :D http://9gag.com/gag/aPGB9YV?ref=fbp [10:08] ^^ if I heat my SSD drive until it becomes a gaseous state hard drive would that enable cloud computing? <<- good idea for a juju plugin [10:17] Bug #1475565 opened: juju expose should allow ip/port restrictions [10:29] Bug #1475565 changed: juju expose should allow ip/port restrictions [10:35] Bug #1475565 opened: juju expose should allow ip/port restrictions [11:36] perrito666, ping [11:48] dooferlad, check out bug 1474946 btw and thumper's fix http://reviews.vapour.ws/r/2190/diff/# [11:48] Bug #1474946: kvmBrokerSuite worker/provisioner: tests are poorly isolated

[11:49] dooferlad, I believe the kvm broker shouldn't need to call EnsureRootFS.. as it only applies to lxc broker tests with lxc-clone: true (when we have the template) [11:54] perrito666, what's the source of the differences in SetStatus on unit vs service in state? [12:00] perrito666, and for that matter unitagent? [12:17] fwereade: sorry was @gym [12:17] perrito666, no worries, I expect asynchronicity :) [12:18] fwereade: I feel your comment a bit context lacking [12:19] perrito666, so, we have 4 SetStatus methods in state [12:19] perrito666, they're all basically setting the same document, and they're all different [12:19] perrito666, do you know what the source of the differences is, and/or if there are any plans to fix this? [12:22] perrito666, some gate on entity life, some don't; some error out when the entity disappears, some don't; some use buildTxn functions, some don't; some follow the very few documented guidelines for state methods, some don't... === psivaa is now known as psivaa-lunch [12:31] fwereade: there are different rules for each entity, they satisfy state.StatusSetter, they could be a bit smaller and cleaned of redundant code tho [12:31] perrito666, what are the differences that aren't encoded in the newFooStatusDoc funcs? [12:34] fwereade: little to nothing, there is some extra resilience around service status and machine status has no history [12:35] oh and unit agent has a slightly different op [12:36] Bug #1475635 opened: Subordinate can break juju run for principal service [12:37] omg you've got to be kidding me one of my monitors will do a humming noise unless brightness is set in 56-60 or >95 levels [12:37] it has been driving me crazy since last night [12:38] katco: ping? [12:41] perrito666, do you know the reasons behind the various differences? [12:41] perrito666, eg, when implementing history, why *didn't* we do machine status? [12:41] partly [12:42] perrito666, why does service status need extra resilience, but not the others? [12:42] perrito666, why doesn't the unit agent check its entity? etc [13:09] Bug #1475641 opened: Bug in hooks can make jujud unresponsive [13:27] Bug #1475386 opened: unit not dying after failed hook + destroy-service

[13:37] is anywone working on https://bugs.launchpad.net/juju-core/+bug/1474946 ? [13:37] Bug #1474946: kvmBrokerSuite worker/provisioner: tests are poorly isolated

[13:37] says thumper there but I think he is no longer here === psivaa-lunch is now known as psivaa [13:51] perrito666, I've figured out why RB2148 is making me uncomfortable [13:52] perrito666, lots of the changes are to the tests, changing the way they reach into the internals of state to "test" it [13:53] perrito666, which means that none of them are testing the important things -- what happens in real life when people call particular state methods -- they're *all* testing how the methods work when someone makes a specific change to the db [13:54] perrito666, so those tests become less accurate, and potentially have to be rewritten, every time someone changes the status code [13:56] perrito666, if you test to the exported interface you can change the implementation and have some confidence that it works how it used to [13:58] TheMue, dimitern: http://reviews.vapour.ws/r/2197/ please! [13:58] dooferlad: *click* [13:58] wwitzel3: here now [14:07] dooferlad, looking [14:08] dooferlad: reviewed [14:12] dooferlad, one significant comment [14:13] TheMue, I am a bit disappointed you didn't spot that... [14:13] fwereade: pardon? [14:14] TheMue, non-bulk api call [14:14] fwereade: iiirks, yes [14:15] TheMue, ehh, it's friday afternoon, these things happen [14:16] fwereade: but should not [14:18] fwereade: too bad you realized after I pulled it :p [14:18] I change it [14:20] is anyone feeling up to speed on the ugly details of destroy-environment in JES? [14:22] dimitern: katco : can you ask someone to triage this issue? I cannot say if the issue is the local machine or a juju bug, and if it is a juju bug, is it something that needs fixing in 1.25 https://bugs.launchpad.net/juju-core/+bug/1474885 [14:22] Bug #1474885: juju deploy fails with ERROR EOF

[14:24] cherylj: perrito666: dimitern: moonstone is in meetings for the next 4 hours. can one of you TAL? [14:27] sinzui, katco, will do [14:27] dimitern: ty [14:27] dooferlad, I have not finished yet, but do have a few comments already, please wait for my review first [14:33] ericsnow: meeting time [14:33] katco: omw [14:37] akhavr: o/ [14:40] dooferlad, reviewed [14:41] dimitern: thanks [14:41] fwereade: thanks as well [14:43] dimitern: guess I should have tried harder for a pair programming session earlier! [14:43] dooferlad, :) no worries [14:43] we can still do it, but it's getting a bit late for me ;/ [14:44] dimitern: can you do 30 minutes? Clearly there was some stuff I just hadn't encountered before and combining that with a little too much copy and paste has set me off in the wrong direction. [14:46] dimitern: though I guess it is mostly that I should have gone back and looked at the spec again and need to know about the significance of bulk API calls. [14:46] dooferlad, sure, let's use the standup g+ [14:48] dooferlad, juju-sapphire [15:00] Bug #1475565 changed: juju expose should allow ip/port restrictions

[15:06] davecheney: about that bug I emailed you about. It was marked as fixed in 1.3, but we're running 1.2.1 officially. I'm not sure how that interacts with the gccgo version, but it seems like maybe we're not yet running a version that has that fix? [15:26] sinzui: davecheney says that enabling trusty updates should fix that ppc issue... is that something that we're allowed to do on the stilson machines? [15:26] natefinch: they are supposed to be on...ubuntu doesn't not support machines that do not accept updates [15:27] natefinch: I think they must have been turned off when diagnosing other issues. [15:27] sinzui: my local machine has gcgo 4.9.1-16ubuntu6, but stilson-07 has 4.9.1-13ubuntu1 [15:28] :/ [15:28] sinzui: oh, forgot I'm not on trusty anymore [15:28] sinzui: just utopic, though [15:30] natefinch: trusty-updates are already enabled on stilson-09 [15:31] natefinch: wwitzel3: planning time [15:32] katco: brt [15:32] me too [15:32] sinzui: weird, that one has gccgo 4.9.1-1ubuntu3 [15:33] at least we're consistently inconsistent [15:37] natefinch: yes, stilson 6-8 permanently dirty. That is why we want the tests to pass in lxc so we can run them in a clean disposable env [15:46] sinzui: can I mark the bug failure as "invalid because it's running on a dirty machine"? https://bugs.launchpad.net/juju-core/+bug/1471657 [15:46] Bug #1471657: linker error in procsPersistenceSuite unit test on ppc64

=== meetingology is now known as Guest99235 [15:46] natefinch: 09 is not dirty [15:47] sinzui: oh, sorry, misread, thanks for the clarification === meetingology` is now known as meetingology [16:40] natefinch: you might want to take a look at this - it's a little cache package that i've just moved out of charmstore internal into juju/utils: https://github.com/juju/utils/pull/144 [16:44] rogpeppe: nice. Needs a package level comment, though. [16:44] natefinch: good point [16:45] Bug #1474382 changed: MeterStateSuite teardown failure on windows

[16:45] Bug #1474946 changed: kvmBrokerSuite worker/provisioner: tests are poorly isolated

[16:45] Bug #1475724 opened: UniterSuite.TestUniterRelations

[16:51] wallyworld: ping [16:51] oh, sorry, it's saturday [16:51] davecheney, it is also very early for him [16:54] alexisb: early as in friday night :p [16:54] its around 3AM for ian [16:55] davecheney: you people from the western hemisphere wouldn't understand. [16:55] AM == morning [16:56] alexisb: I place the line between late and early around 5AM [16:56] 4am == night time, I agree with that :) [16:56] which is the time when, if I wake up Ill go with a coffee instead of more sleep [16:57] Bug #1475724 changed: UniterSuite.TestUniterRelations

[16:57] Bug #1474382 opened: MeterStateSuite teardown failure on windows

[16:57] Bug #1474946 opened: kvmBrokerSuite worker/provisioner: tests are poorly isolated

[16:59] Does anyone have a quick tldr on Juju/LXD support? I'd like to answer this question: http://askubuntu.com/questions/643658/lxd-and-juju-management [17:04] Bug #1474382 changed: MeterStateSuite teardown failure on windows

[17:04] Bug #1474946 changed: kvmBrokerSuite worker/provisioner: tests are poorly isolated

[17:04] Bug #1475724 opened: UniterSuite.TestUniterRelations

[17:12] natefinch: that bug you emailed me about earlier [17:12] i'm pretty sure that is fixed [17:12] and has been fixed for a long time [17:12] _but_ there is a catch [17:12] it's not available for 14.04 out of the box [17:12] even in 14.04.2 [17:12] you _have_ to enable trusty-updates to get it [17:14] davecheney: sinzui says trusty updates is enabled on the ppc machines [17:14] sinzui: silson-09 says that gccgo is installed and set to "manual"... does that mean it's not getting updates? [17:14] the magic package is libgo5 [17:15] * davecheney struggles to contain his rage [17:15] WE OWN THE DISTRIBUTION [17:15] WHY CAN WE NOT ACTUALLY SHIP UPDATES FOR SHIT WE MAINTAIN [17:15] s/silson/stilson/ [17:15] phase 1. fix bug [17:15] phase 2. ??? [17:15] phase 3. nobody actually gets the bug fix [17:15] ffs [17:15] afk for 15-ish or so, sorry. Kids need naps [17:16] davecheney: the machine always had updates on, but the issue might be that trusty didn't officiall support ppc4el. Its packages come from ports.ubuntu.com [17:16] natefinch: just emailed you the background [17:17] actually no [17:17] that was the wrong email thread [17:19] natefinch: i cannot find it now [17:19] but the summary was [17:19] "no versions of our LTS release ship with a working gccgo, you must be using the updated versions from trusty-updates" [17:22] davecheney: i'd appreciate a glance at https://github.com/juju/utils/pull/144 if you have a few moments. [17:22] * davecheney looks [17:22] rogpeppe: safe flight back to blighty ? [17:22] davecheney: yup, no probs at all [17:22] davecheney: you still in US ? [17:27] * rogpeppe leaves [17:55] wwitzel3: i'm ready to party. [18:12] katco: ok, I'm running a bit behind, lots of plates spinning [18:12] wwitzel3: np, lemme know when you're ready [19:14] perrito666, ping [19:14] alexisb: pong [19:14] are you still working this bug w/ menno: [19:14] https://bugs.launchpad.net/juju-core/+bug/1474606 [19:14] Bug #1474606: $set updates may clear out the env-uuid field

[19:15] alexisb: I added a fix for the original issue, but as I understand menn0 was working on some additional code to do a more general fix [19:15] perrito666, ok [19:16] sinzui, we are still blocked on 1474606 [19:16] alexisb: understood [20:52] Bug #1475779 opened: EnsureJujudPassword fails if the registry keys are not present [21:45] jw4: hey didn't you do a relation state diagram at some point? [21:53] katco: yeah - a first pass [21:53] it should be in the docs folder [21:54] jw4: do you remember where that is? i couldn't find it [21:54] jw4: oh, hm. nothing jumped out at me there [21:54] katco: I think it's doc/uniter-model.txt [21:55] katco: pretty thin [21:55] jw4: no worries, ty [21:55] :e [21:55] um... that was supposed to be :) [21:55] jw4: haha thought you were in vim for sec [21:56] exactly [22:35] rick_h_: ping [22:37] I have a bundle question rick_h_