=== kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [00:18] thumper: did you see that question in #juju-dev and ask ubuntu about local provider and not being able to mount /var/log/juju--local? Do you recall what might cause that? [00:18] yeah [00:18] sadly [00:18] can we chat after I have some food? [00:18] catch up should be good [00:19] sure, or you could answer his question :-) [00:19] catch up sounds good anyway [00:19] i might grab some food too [00:27] wallyworld_: btw, I wanted to ask you so Ill leave the question here [00:27] there are descriptions for most of the other statuses describing where they are used and what for, do we have those for the new ones? [00:27] waigani: tx for the review btw [00:27] perrito666: yep, all in the spec [00:28] the ones in code now are wrong [00:28] thy are really only there as placeholders [00:28] wallyworld_: I guessed so, but abstained from changing them until it was set in stone [00:28] perrito666: np, I agree with wallyworld_ on the func name, on second thought [00:28] anastasiamac: http://reviews.vapour.ws/r/888/ [00:28] waigani: just saw :D tyvm!! [00:28] anastasiamac: sorry for the wait, running low on fuel, need food! [00:28] wallyworld_: since they have to change all over in tests too which is a bit of a pain to do more than once :p [00:29] waigani: np. eat!! === kadams54 is now known as kadams54-away [00:30] perrito666: agreed. i didn't want us to do anything with the actual values until the spec is finalised [00:30] even the ones there now aren't complete at the time they were added [00:31] wallyworld_: well I finally changed then or you would add a comment that I those are nto the right ones each time I pushed to go from one computer to the other :p [00:32] wot? did i do that? i don't recall. i didn't mean to [00:33] ts ok, you reviewed each of my pushes, even the ones I did forgetting that they changed rb lol [00:40] davecheney: thnx for review :D [00:40] davecheney: there is no better feeling than seeing Shipit :D [00:48] wallyworld_: ready when you are [00:48] anastasiamac: really? no better feeling? [00:48] one minute [00:48] thumper: at work? yes - no beta :D [00:50] thumper: outside of work - driving fast cars does it :D [00:50] anastasiamac: I wasn't even going to go there :) [00:50] thumper: where?.. [00:50] outside work [00:53] wallyworld_: I'm in our 1:1 hangout when you are ready [00:53] why, thumper... what gives u a good feeling at work? [00:53] and reviewing waigani's branch while I wait [01:38] waigani: getting down to relatively small issues now [01:38] waigani: I think most comments were about improving readability [01:39] thumper: okay, thanks [02:23] hello reviewers: http://reviews.vapour.ws/r/912/ [02:35] menn0, waigani, davecheney: http://reviews.vapour.ws/r/912/ [02:38] thumper: looking [02:53] thumper: shit it [02:53] thumper: or ship it :) [02:53] heh [02:53] I was going to ask if that was a slip [02:54] only wrt to how i'm currently feeling, not the code :) [02:56] thumper: only one more component to go with this logging POC. with a bit of luck i'll have logs streaming through the API end-to-end [02:56] soon [02:58] menn0: \o/ [02:58] thumper: it's a bit rough and has no tests and there's a few details to iron out but it should work [02:58] menn0: sounds like a spike to me :) [02:59] thumper: the code is mostly reasonable quality. writing tests and cleaning it up will probably add a day or 2 [03:02] waigani: shipit... when the queue is unblocked [03:03] thumper: great, thanks. I'm going to have some merge conflicts. [03:03] waigani: best get them sorted now then :) [03:03] thumper: yep [03:03] * thumper needs coffee [03:03] bbs [03:08] waigani: thumper: environments can be annotated [03:08] waigani: thumper: for all other entities that can be annotated, we remove annotations on entity destruction/removal [03:09] waigani: thumper: i could be blond but m not seeing it for environment destruction [03:09] waigani: thumper: s/blond/blind [03:09] anastasiamac: when an env is destroyed, all docs associated with that env are removed [03:10] anastasiamac: at least for non state server envs [03:11] waigani: to ease my mind, would it b possible to have a test for it :D [03:11] anastasiamac: yep, sounds like a good idea [03:11] waigani: much appreciated! [03:12] waigani: plz let me know when env desruction lands [03:13] waigani: i may need to go over old annotations impl too as environments r special l:D [03:13] anastasiamac: just got a shipit, so once the IC blocker is moved ... where's menno? [03:13] waigani: menno is blocking ci? [03:14] anastasiamac: no, he is THE UNBLOCKER [03:14] anastasiamac: so in your annotation's tests your using RemoveEnvironment (or something like that) [03:14] anastasiamac: that just removes the env doc [03:14] waigani: yes ;D I had to despite the comment that u guys were not happy with it [03:15] anastasiamac: once destroy lands, there will be a RemoveAllEnvironDocs func [03:15] waigani: hence m happy to go over annotation stuff once ur branch lands :D [03:15] waigani: awesome! m look n forward to it [03:15] anastasiamac: that func will remove the env doc and ANY doc that has a env-uuid field with a matching UUID [03:16] waigani: sounds exactly like what i want [03:16] anastasiamac: so that should clear out any annotation docs associated with the env [03:17] anastasiamac: but yeah, I'll land the branch and then we can take a look at the annotation tests and make sure it's all covered. [03:18] waigani: sounds gr8 :D [03:18] waigani: annotation impl treats env differntly than other entities for historical reasons [03:19] waigani: so i'll fix this once JES is capable of destroying env D [03:20] anastasiamac: okay. I haven't looked into annotations at all. But happy to help from the JES side [03:20] waigani: np [03:20] waigani: i'll ask if i'll get lost with JES :D [03:20] sounds good [04:18] * thumper sighs [04:18] seems like a bad test in state [04:18] intermittent failure [04:18] stopped eric's fix branch landing more than once [04:18] http://juju-ci.vapour.ws:8080/job/github-merge-juju/2093/? [04:20] FAIL: machine_test.go:1038: MachineSuite.TestMachinePrincipalUnits [04:20] * thumper must write the spec [04:20] gah, distractions === fuzzy__ is now known as Ponyo === mup_ is now known as mup [04:53] wallyworld: I'm not going to go there yet, but I'm starting to wonder whether the machine provisioner should be made more generic, to handle de/provisioning storage as well [04:54] we're going to need to do pretty much all the same things... e.g. "harvesting" volumes that we fail to record in state [04:55] and the machine/container provisioning is very similar to env/machine volume scope [05:11] davecheney, ping? [05:12] mattyw: pong [05:31] axw: yes, it may com to that [07:58] morning [08:01] TheMue: morning :D [08:01] anastasiamac: o/ === liam_ is now known as Guest3630 [09:01] ericsnow: about ci blocker, bug 1420426 [09:01] Bug #1420426: backups_nonlinux.go import error

[09:01] ericsnow: i just got trunk and it looks like the offensive import is not there... [09:02] ericsnow: should the bug be marked "fix committed" and hopefully unblock CI [09:16] fwereade: greetings [09:16] jam, ahoy [09:16] fwereade: How's Leia doing? [09:21] fwereade: are you coming to leader-election? [09:22] morning all [09:32] anastasiamac: Block ok pr 916 refers to the action of blocking? [09:33] perrito666: m lost... could u re-phrase? :D [09:33] sure [09:33] http://reviews.vapour.ws/r/916/diff/# [09:34] Block, means blocking as in preventing something to be done? [09:35] perrito666: yes :) PR 916 is about adding env blocks to prevent some operations from runnng [09:35] perrito666: currently, u can 'juju block' whatever [09:35] perrito666: however, it just adds env variable [09:35] anastasiamac: :p as a non native english speaker let me drop a word of advice :) when seen in the code, "block" is a veeery ambiguous word [09:36] perrito666: we want to be able to desribe why we are blocking smth [09:36] perrito666: not my choice of vocabulary :D [09:37] I first read the pr backwards and was thinking on block as "piece of something" :p [09:37] perrito666: altho in this instance it is actually descriptive of what it does... [10:02] voidspace: ping [10:02] TheMue: omw [10:04] TheMue: dooferlad: firefox crash! [10:04] rejoining [10:06] dooferlad: tried chrome, video still didn't work... [10:07] jam: you around? === benji is now known as Guest55003 [10:55] getting maas running locally again [10:55] switching off "disk erase on release" helps... [10:56] I'd switched it on for bug hunting previously [10:56] desperately hoping I get to successful bootstrap before timeout [10:56] and fail... [10:59] ooh, now pxe boot is working again for no specific reason [10:59] that's nice [10:59] possibly due to an update [11:00] nope, it boots then shuts down [11:00] dammit [11:01] ah, the wrong machine pxe booted [11:01] starting the second node kicks things off [11:01] how odd [11:01] race against timeout again [11:01] must increase the timeout... [11:03] bootstrap! [11:05] wwitzel3: ping [11:06] wwitzel3: I can (finally) bootstrap and deploy to MAAS with no problem [11:06] wwitzel3: can't reproduce the x509 cert issue [11:06] wwitzel3: I've added a note on the issue [11:31] fwereade: o/ [11:32] anastasiamac, heyhey [11:32] fwereade: how r u? [11:33] anastasiamac, good thanks, and yourself? [11:33] fwereade: m good... :) [11:33] fwereade: i need to distract u [11:33] fwereade: and beg for a favour... [11:34] anastasiamac, np :) [11:34] fwereade: could u plz chnage my subscription to juju-dev from digest at one stage?.. [11:34] fwereade: m getting delayed correspondence [11:34] fwereade: it'd b gr8 if I could do it myself [11:35] fwereade: but when I tried, i was sternly told that m already subscribed and need to contact u :D [11:35] anastasiamac, hummmm I am technically an administrator there aren't I? :/ [11:35] fwereade: it's not urgent but would b gr8 [11:35] fwereade: yes u r :) [11:36] anastasiamac, figuring out what to do is not going to be a 5-min thing so I need to put it on the stack -- can I suggest unsub/resub as a possibly quicker method? [11:36] fwereade: i'll try :D thnx :) [11:42] fwereade: i bow to ur wisdom [11:42] fwereade: when i went to unsubscribe, i had to log in [11:42] fwereade: which allowed me to change from digest subscription :D [11:42] anastasiamac, excellent :) [11:42] fwereade: all sorted [11:43] fwereade: tyvm!! [11:43] * fwereade likes thanks for doing nothing at all ;p [11:44] fwereade: pointing ppl in the right direction is doing smth :D [11:45] anastasiamac, when you look closely enough, I suppose :) [11:54] wallyworld: what's up ? [11:55] jam: hey, re: running-hook in status spec. i'm torn between "executing" as per mark's suggestion, and running-hook and running-action [11:55] did you have an opinion [11:56] I haven't read that particular bit lately, but just from what you bring up, I'm not sure there is a worthwhile *user* distinction there, is there ? [11:56] wallyworld: as in, someone watching "juju status" will wait on running-action but go ahead if it was running-hook [11:57] jam: not so much for scripts, but humans, but i guess that's what message is for. so seems like "executing" might be ok [11:57] wallyworld: yeah, if its just humans than that's what the message is for. [11:57] mark also didn't want like the recent hooks [11:57] run it by fwereade as well [11:57] wallyworld: that was a gustavo thing that we didn't get a chance to run by him [11:58] IIRC [11:58] yeah, i think that was the case [11:58] I think being able to introspect it is good, I don't think *I* would prefer it to be in default status for sure [11:58] possibly when digging deeper [11:58] i suggested using a --verbose flag or some such [11:58] i didn't want it output by default [11:58] but i could see the value [12:01] wallyworld: yeah, I feel like this sort of thing falls more into a "debug-log" sort of syntax [12:02] juju debug show-recent-hooks $ID [12:03] yeah, sounds fine [12:03] i could see how it fitted into a status deep dive also [12:05] wallyworld: yeah, that *might* be a way to get it, though I'd be hesitant to do it on a command that would try to do it against all 1000 units in the env [12:05] true [12:16] fwereade: an interesting development, I accidentally messed up while transcribing the tests and I wonder if I'm running into something. [12:16] jam, oh yes? [12:16] specifically, "Set the charmURL to trigger config events" has an assertChange() [12:16] and then we change the config [12:16] and assert that we get one more change [12:16] fwereade: which seems fine, *but* [12:16] if I remove the first assertChange [12:16] then I get 2 changes on the channel [12:16] should those actually be coalesced? [12:18] concretely, filter_test.go, line 327, if you comment that out the next "assertChange" fails because it has 2 events in the queue [12:18] fwereade: anyway, have to take the dog out, bbiab [12:20] jam, in general, yes, but... is it plausible that the StartSync (which is not, uh, sync) only causes the second change to be delivered just after the filter sends the event for the first one? [12:20] jam, which could indicate the potential for races in the tests that *do* check coalescence... hmm [12:23] fwereade: I have the feeling it is more that maybeSendMessage actually causes us to send 2 [12:24] anyway, good that it happened because I "fixed" the test wrong [12:24] but made me wonder [12:53] fwereade: http://reviews.vapour.ws/r/917/diff/# is just my "use nil instead of make(chan)" proposal [12:53] jam, LGTM [13:38] fwereade: I made some progress on updating the tests to use ContentAsserterC and NotifyAsserterC, I'll finish that up tonight and propose it. [14:22] voidspace, you around? [14:31] voidspace: thanks [14:44] wwitzel3: want to do our 1:1? [14:46] natefinch: sure [14:47] natefinch: just refilling my coffee [14:47] wwitzel3: a worthy cause for delay :) [15:02] our ReviewBoard site will be having some downtime today [15:02] (for real this time) [15:02] ericsnow: :) [15:02] I'll send out a message about 10 minutes before taking it down [15:02] and right after it comes up [15:05] ericsnow, perrito666: you guys coming to the standup? [15:05] my cal says its in 1h but sure [15:05] natefinch: yeah, in 55 minutes :) [15:05] ericsnow, perrito666: I just moved it to now :) [15:06] ericsnow, perrito666: that way it's always at the same time.... the meetings that had interfered with it in the past are gone, so figured I'd fix it [15:06] ....somewhat retroactively === benji_ is now known as benji [15:09] how are you, devs? [15:14] voidspace: I proposed a fix for the ports issue here: https://github.com/juju/juju/pull/1588 [15:31] alexisb: brt, still in standup [15:31] nw [15:32] no rush [15:35] ericsnow: didn't you fix 1420426? [15:35] perrito666: yep [15:35] * perrito666 gets denied by juju bot because of it [15:35] natefinch: you ever heard of https://github.com/moovweb/gvm? [15:36] perrito666: Tim landed it last night [15:36] i think that would be cool to have for managing juju versions [15:36] a similar tool [15:36] make it easier to test against development releases [15:36] stokachu: yeah, I know of it. For go, I think it's a terrible idea. :) For Juju, well, it might be a better idea. [15:38] cool, just a thought i had :) [15:39] stokachu: ideally we'd make it easier to run different juju clients side by side [15:39] frankban: ok, thanks [15:39] natefinch: yea that would be perfect, im trying to work out the best way to automate testing our installer against all available juju versions [15:40] frankban: reading through it [15:40] stokachu: in theory, just switching $JUJU_HOME per version of juju can let you run them side by side [15:40] voidspace: thanks [15:41] natefinch: good point, i could probably script something up for that [15:41] we use a custom home anyway for juju [15:44] voidspace: dow you know why that MP is not automatically hooked to review board? [15:51] no [15:51] frankban: is your membership of the juju team public? [15:51] voidspace: I guess so, and my last proposal was correctly hooked [15:51] frankban: I think that's a requirement [15:51] ah [15:51] ericsnow: ping [15:51] voidspace: hey [15:51] voidspace: OTP [15:52] ericsnow: this PR hasn't landed on reviewboard [15:52] ericsnow: https://github.com/juju/juju/pull/1588/files [15:52] ericsnow: ok [15:52] voidspace: I'll take a look in a bit [15:52] ericsnow: thanks [15:52] voidspace: it's probably a unicode issue (Python 2.7) :P [15:53] ericsnow: hah :-) [15:53] voidspace: I'm actually not kidding! (we have unicode somewhere in our mongo package that RB fails on) [15:54] :-/ [16:03] frankban: so lines 156 on in the megawatcher [16:03] frankban: you moved the code assigning PortRanges and Ports [16:04] frankban: to "preserve the current status and ports." [16:04] frankban: why are we preserving the old ones instead of using the new ones? [16:06] voidspace: on call, back in a minute, but the answer is that we only need to calculate ports on new entities. If an entity is already there, then its ports are updated by the bakingOpenPorts [16:06] frankban: ah, ok - cool, thanks [16:06] frankban: that answers my next question too :-) [16:13] frankban: so your PR LGTM, but I'm not familiar with the intricacies of all the entities involved [16:14] frankban: in essence is the megawatcher now watching another collection? [16:14] voidspace: yes [16:14] frankban: cool [16:14] frankban: I'm going to give it an LGTM [16:14] voidspace: it now watches the openports collection [16:14] voidspace: ty! [16:57] guys, where can I start to study something about using Juju with Windows Machines? [17:11] Muntaner: hi there. I think I may be able to help [17:11] Muntaner: on which provider would you like to use Windows? [17:11] MaaS, OpenStack, etc? [17:12] gsamfira: I don't have the details under my hands right now - bosses will tell me them soon [17:12] gsamfira: o/ [17:12] gsamfira: hi [17:12] and I wanted to start studying the situatio [17:12] n [17:12] in general [17:13] Muntaner: okay then. So in general there are a few requirements to get windows up and running. And it depends on the provider you are planning to use [17:13] for an Openstack private cloud, you will need 2 things. The windows agent tools, and a windows image preinstalled with cloudbase-init [17:14] I am not sure if the windows agent tools are part of the official simplestreams that juju offers for 1.21 [17:14] but if it is not, I can help you build them [17:14] Muntaner: a windows image can be downloaded freely from: http://www.cloudbase.it/ws2012r2/ [17:15] its an evaluation image valid for 180 days (I think) [17:15] I can write up a wiki page on setting up your OpenStack private cloud (Incidentally I need to set one up anyway....I can just document the steps) [17:15] gsamfira: thanks! [17:16] My pleasure [17:16] If you plan to use MaaS, you'll need a windows image for that as well. There is a image builder in the works at the moment that will be released as part of MaaS [17:18] Muntaner: in any case, I'll be online a lot in the next couple of weeks. Ping me anytime on IRC :) [17:18] voidspace: hi :D [17:18] gsamfira: great! thanks :D [18:43] g'night all [18:43] EOD [19:35] in 10 minutes I'm going to be taking reviewboard down for a little while (to move to a new host) [19:37] * perrito666 panics and runs in circles [19:45] reviewboard going down now [19:49] If you want to know why GCE isnt finished being reviewed, you can blame ericsnow ;) [19:50] haha [19:56] morning folks [19:57] fwereade: if you are around this evening, it would be great to have a catch up call [20:03] reviewboard is back up (on the new host) [20:03] let me know if anything is amiss [20:07] ericsnow: no new features :p [20:07] perrito666: wasn't an upgrade ;P [20:07] i know [20:07] I just fancy upgrades [20:07] :p [20:08] ericsnow: I'm getting a 500 [20:08] ericsnow: http://reviews.vapour.ws/r/771/ [20:08] natefinch: impossible [20:08] natefinch: ;) [20:10] natefinch: ah, I guess the charm doesn't install git for you === kadams54 is now known as kadams54-away [20:10] natefinch: problem solved [20:11] ericsnow: I could do with a different theme (?) [20:11] perrito666: go for it [20:11] ericsnow: oh :( I thought these where canned things [20:11] perrito666: perhaps (I don't know) [20:19] hey, next tue i am ocr but its a holiday here, anyone wants to trade? [20:21] perrito666: i can trade, i'm the next day (wed.) [20:21] sweet, tx :D [20:22] perrito666: happy holiday :) [20:22] tx, its carnival :D [20:22] fun times [20:22] though since cellphones we can no longer have as much fun :p [20:22] lol because of cameras everywhere? [20:22] and well there is the fact that being a grown up I no longer have one of those wather proof wallets [20:23] katco: because of cellphones not being nice with wather [20:23] ah [20:23] when I was a kid you would walk pass a neigborhood with kids an they would splash you throwing these small balloons filled with wather === kadams54-away is now known as kadams54 [20:25] yeah that would not be too fun with an expensive computer in your pocket [20:29] be back online ~1 hour [21:09] CI is blocked... still :( [21:18] again [21:18] not still [21:45] menn0, you available for a call at the top of the hour? [21:46] alexisb: yep [21:46] thumper, ^^? [21:46] sure [21:46] call about? [21:46] this re the issue menn0 worked last week [21:46] ok [21:47] and menn0 we are not volunteering you for anything but we need to determine next step for the customer === Guest82007_ is now known as Viperz28 === kadams54 is now known as kadams54-away [22:49] wwitzel3, ping [22:49] alexisb: pong [22:50] wwitzel3, nevermind, I see #juju [23:08] any know issues with upgrading local charms in 1.22-beta3-trusty-amd64? [23:08] * whit is experiencing a hang on aws [23:14] whit, nonw [23:14] none [23:14] whit, these are the known issues https://launchpad.net/juju-core/+milestone/1.22-beta4 [23:14] ok, I'll report it [23:29] sinzui: FYI, I don't think bugs #1403243 and #1391276 apply to 1.22 and 1.23. [23:29] Bug #1403243: juju-backup restarts rabbitmq, reconnect is unstable

[23:29] Bug #1391276: a terminated juju backup can leaves server in a dirty state

[23:30] sinzui: also, why is #1399310 marked as high priority? [23:30] Bug #1399310: The code in apiserver/backup.go belongs in apiserver/backups.

[23:31]