[00:16] menn0: ok [00:22] thumper: thanks for the review! [00:38] thumper: another one for you: http://reviews.vapour.ws/r/707/ [00:38] thumper: thnx for reviewing my doc :-) [00:39] menn0: ack, doing another right now [00:39] thumper: np. [02:12] wallyworld_ thumper: responded to http://reviews.vapour.ws/r/699/, PTAL when you can [02:12] sure, looking [02:56] morning all [02:57] thumper, morning/ afternoon, you still around? [03:08] axw: done [03:08] mattyw: yep, and reviewed [03:08] cheers [03:09] trunk is blocked again :~( [03:09] thumper, cool - is that an LGTM or a vote of confidence? [03:09] :) [03:09] merge it [03:09] thumper, I don't have permission - but if you do you're welcome to [03:09] ah, ok [03:10] mattyw: done [03:10] thumper, that's great service [03:10] thumper, thanks very much [03:14] axw: i think i may be running into an issue with maas 1.7 - the node is green/ready, and i try and bootstrap. i expect it to go blue/allocated and then i can use the virsh console to start manually. however, juju bootstrap exits with an error FORBIDDEN (You are not allowed to start up this node.) [03:15] maybe maas is complaining there's no power mgmt options set [03:16] wallyworld_: I don't think I've used 1.7, only 1.6... [03:16] can't say that error sounds familiar [03:16] bollocks [03:17] i have a look at the python used to control the virsh console and it appears deficient in that it neglects to pass the --quite flag, so login is deemed to fail [03:17] so there's clearly some things they need to fix to get start/stop working [03:19] axw: rereviewed the storage branch [03:19] thumper: thanks [03:21] thumper: got a sec to chat? [03:21] rick_h_: sure [03:21] thumper: https://plus.google.com/hangouts/_/canonical.com/daily-standup?authuser=1 === kadams54 is now known as kadams54-away [07:09] rvba: hi there, did you have a few minutes for a question or 2 about maas? [08:13] wallyworld: how are you going with maas? [08:13] need a hand with anything? [08:16] axw: well, tl;dr; it kinda sucks. there's a gomaasapi commit after the lastest one in dependencies.tsv that fails unit tests in juju. so i have to get that fixed prior to being able to land my gomaas mods to then propose my juju fix. i have also reverse engineered the api somewhat to figure out how to fix the deployment status issue, but need to confirm a couple of things as the doco is lacking. i'm just finishing adding the new api [08:16] to the gomaas testserver so i can then use that in juju unit tests [08:16] and maas plus kvm in 1.7 is broken [08:16] so hard to test live [08:16] oy :( [08:17] thanks for asking :-) [08:18] axw: also, the testserver was somewhat, shall we say, retarded in how it was written wrt handing constraints, so i had to fix that also [08:35] axw, wallyworld, rogpeppe1, mgz, hey guys do you mind going here https://github.com/orgs/go-amz/people and setting yourself as publicly visible? [08:36] davecheney, ^^ [08:36] dimitern: done [08:36] rogpeppe1, cheers! [08:37] dimitern: done [08:38] axw, btw you should be unblocked now wrt storage work on goamz - api version upgraded to latest, so if you import gopkg.in/amz.v2 you should be all set [08:38] davecheney, thanks! [08:57] dimitern: I noticed, thank you [09:14] alexisb, fwereade: just bringing this feature bug to your attention - https://bugs.launchpad.net/juju-core/+bug/1409639 - hopefully that's not to much of a surprise [09:14] Bug #1409639: juju needs to support systemd for >= vivid [09:14] jamespage, ha, yes [09:21] dimitern: I don't know if you want to have a look at this http://reviews.vapour.ws/r/704/ [09:22] voidspace, sure, I'll have a look in a bit [09:50] voidspace, reviewed [09:50] dimitern: thanks [10:23] voidspace: hey there [10:24] wallyworld: hey, hi [10:25] voidspace: quick one - the last comment to gomaasapi (rev 59) doesn't work with juju unit tests. i have changes i need to make to gomaasapi but can't until those issues are fixed. related to nodegroup changes [10:25] juju deps still refers to rev 58 so juju is not broken at this stage [10:26] wallyworld: right, I have a branch that fixes the maas test that is broken [10:26] wallyworld: I think the code is actually broken [10:26] \o/ [10:26] wallyworld: http://reviews.vapour.ws/r/704/ [10:26] oh, great, a pr already [10:27] and you have a ship it [10:27] that will unblock me tomorrow, so i'm happy [10:27] cool :-) [12:33] wallyworld: branch ready to land, but trunk blocked... [12:34] voidspace: ah, :-( [12:34] wallyworld: the fix you need to get tests passing with latest gomaasapi is change line 183 of maas/environ.go to [12:34] if err != nil || len(bootImages) == 0 { [12:34] (add a check for empty bootImages) [12:34] voidspace: i made that same change locally [12:34] cool [12:34] so it must be right :-) [12:34] I'll try and land this asap [12:34] hah [12:35] I defer to your wisdom [12:35] as always [12:35] barf [12:37] voidspace: i checked critical bug from irc topic, fix for regression was committed over w/e so hopefully it will be verified soon as fix released to unblock [12:37] wallyworld: great, thanks [12:38] the fix is in, last run over the weekend failed on maas, I triggered a rebuild in case it's happy now [12:39] hmmm [12:40] wallyworld: I'm not sure about the fix though, only changes the test, so the api still returns stuff in an undefined order? would prefer if the api return was just sorted [12:43] mgz: why would u prefer to see return sorted? [12:43] anastasiamac: because it's an api - I like having stuff we're exposing for other people deterministic, just so they can't also make a mistake in their code and assume that it's in insert-order or whatever [12:45] there are places where that's quite important, and many more where it's really not [12:48] fwereade: wallyworld: ^^ would the order of returned annotations from bulk call matter (on get)? [12:48] anastasiamac, yes please, preserve the order they were sent in [12:49] anastasiamac, * the requests were sent in [12:49] mgz: the test was failing not on returned order. it was setting annotations and m checking if the collection that m about to throw over the wire contains elements I was after [12:50] fwereade: mgz: will duble check that m returning collection from get in the same order as the request :) [12:50] mgz: thnx for pick up [12:53] anastasiamac, doh, I see, I think; so no, I don't think that order matters, users that need to sort results by alpha or something can do so trivially, it's not necessary that we do that sort of work [12:53] anastasiamac: fair enough [12:55] good morning everyone, sorry for the lateness === ChanServ changed the topic of #juju-dev to: https://juju.ubuntu.com | On-call reviewer: see calendar | Open critical bugs: None [14:27] trunk is unblocked [14:36] dimitern: how did you fix it_ [14:36] ? [14:37] perrito666, anastasiamac fixed it [14:38] perrito666, see https://github.com/juju/juju/commit/66bf4163b9580a1052193bbe8934dc21a4384f11 [14:38] dimitern: ah I saw that on friday [14:39] I though thi was about abentley mail (or perhas his mail was about this)? [14:39] perrito666, ah, the uniter one - no, no fix for that yet - it's intermittent [15:01] ericsnow: ping [15:03] hey wwitzel3 still of for the standup right? [15:03] perrito666: ping [15:03] wwitzel3: pong [15:03] perrito666: just checking if you're coming for standup? [15:04] wwitzel3: I thought I got an email from you saying no standup [15:04] :p [15:04] brt [15:31] voidspace, https://github.com/juju/juju/pull/1382 got merged already btw :) [15:34] dimitern: yeah, I saw [15:34] dimitern: thanks :-) [15:34] voidspace, np [15:34] dimitern: I only saw *after* I tried to merge again of course... [15:34] voidspace, I thought so yeah :) [15:56] wwitzel3: you back? === kadams54 is now known as kadams54-away [18:00] wwitzel3: as OCR, can you try to make sure someone is at least investigating Aaron's 2 reported bugs about CI cursing current trunk? [18:00] (it doesn't have to be you, I just want to make sure someone is following up on it) [18:09] jam: will do, thanks for the ping [18:32] g'night all [18:32] EOD === ChanServ changed the topic of #juju-dev to: https://juju.ubuntu.com | On-call reviewer: see calendar | Open critical bugs: 1409827 [19:35] morning [19:37] hi thumper [19:37] hey perrito666 [19:37] perrito666: all good in your neck of the woods? [19:38] thumper: sure, we seem to be doing better than most parts of the world accoding to the official news, albeit a bit more hot :p [19:39] :) === urulama__ is now known as urulama [19:47] abentley: do we know when did this break? #1409827 that is [19:47] Bug #1409827: TestSetMembersErrorIsNotFatal fails on ppc

[19:49] perrito666: 66bf4163b9580a1052193bbe8934dc21a4384f11 did not exhibit the issue: http://reports.vapour.ws/releases/2219 [19:49] abentley: tx [20:33] bbl [20:33] ericsnow: I'm back, looking at the regression [20:34] thumper: ping, we're trying to get a backport of frankban's polish into 1.22, we've filed a bug and need a review on a backport PR I'm told? https://github.com/juju/juju/pull/1388 [20:41] wwitzel3: k [20:43] thumper: ping, noticed your comment on the regression are you running the PPC test suite? [20:43] thumper: I'm trying to debug that issue now, but not sure how to run the tests. [20:46] ahh nice, I can just use a ppc vm on qemu .. that seems to work [20:49] wwitzel3: thnx for the review!! [20:49] anastasiamac: np [20:49] wwitzel3: I have a question - PTAL if u get a chance b4 :) [20:53] wwitzel3: b4 eod even :) [20:55] anastasiamac: yep, you have a response, let me know if that makes sense [20:58] wwitzel3: i do? :) sorry i must be looking in the wrong place - where can I find it? [20:59] anastasiamac: oops, sorry, was still a draft .. my fault [20:59] how are unit tags generally marshalled/unmarshalled? [21:00] service tags are handled automatically because their member-variables are public; unit has a private member-variable [21:00] wwitzel3: yep. thnx :) that's what I thought I understood :P [21:01] wwitzel3: really appreciate ur thoughts and input! thnx [21:01] anastasiamac: you're welcome [21:05] wwitzel3: do you get the same issue just running with the gccgo compiler on the local machine? === urulama is now known as urulama__ [21:05] wwitzel3: I'm fairly sure it is a timing problem [21:06] wwitzel3: and I think the test should be rewritten to not need time.Sleep [21:06] cherylj: I'm back if you have time to talk now [21:07] thumper: ok, let me try compiling with gccgo. [21:07] thumper: thanks for the hints [21:10] thumper, now works for me [21:11] cherylj: we'll just use the standup hangout [21:11] ok === kadams54-away is now known as kadams54 [21:14] thumper: it passes locally for me when using gccgo [21:15] wwitzel3: otp now [21:32] thumper, waigani: environments watcher here: http://reviews.vapour.ws/r/711/ === kadams54 is now known as kadams54-away [21:41] menn0: lgtm [21:41] waigani: cheers === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [22:45] thumper: when you have a moment can we discuss the machine agent worker stuff? [22:45] menn0: in 10 minutes? [22:46] thumper: sounds good === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === whit is now known as whit|dinner