[00:01] redir, yes empty string [00:08] part 1 https://github.com/go-amz/amz/pull/72 PTAL [00:20] anyone have a minute to look at something with me? [00:24] * thumper heading up to the cafe [00:30] * redir heads to GOSF to look at some talks. Most specifically one about https://github.com/ekanite/ekanite because... What if we just used a raft log to persist state? Where'd the logs go? [00:32] also there's this: https://github.com/reedobrien/juju/tree/lp/1612645%2B1634289_new-aws-regions [00:33] but it fails because: cannot write file "provider-state" to control bucket: The specified bucket does not exist [00:33] so I must have missed something [00:33] in adding the regions, or something else is wrong [00:33] later === hml_ is now known as hml [02:19] anyone have hany thoughts about why the control bucjet doesn't exist? [02:20] context? [02:21] menn0: have you had a chance to take another look at that commit with my latest changes to fix import ordering etc? [02:22] wallyworld: added new ohio region to amz and 1.25 provider and tried to deploy [02:22] got the error 6 lines up [02:22] 6 lines up from where? [02:22] wondering if I missed something obvious, [02:22] wallyworld: sorry, I haven't looked yet, otp but will look after [02:22] this channel [02:22] menn0: np, ty [02:23] redir: not 6 lines for me, i have other notifications turned on, more like 20 and it's scroll off my screen :-) [02:23] wallyworld: this channel at 17:33:07 [02:23] redir: did your environment.yaml have a control bucket? it shouldn't [02:23] * redir looks [02:23] also not 17:33 more me? [02:24] :-) [02:25] hehe [02:26] 33:07 after the hours two hours ago [02:26] no control-bucket in environments.yaml [02:28] redir: can you provide debug logs? aws is expected to create a controller bucker using a self generated bucket name from memory [02:29] wallyworld: I'll work on it tomorrow [02:29] it's been 2 years since i saw that code, so will have to go digging [02:29] it only fails in US East new region. The AP South one I added works [02:29] hmmm, interesting [02:29] * redir guesses it assumes the wrong things for useast [02:29] maybe that new rwgion doesn't have s3 support yet [02:29] or the s3 endpoints are wrong [02:30] doesn't sounds like a juju issue [02:30] if it's just the one region. more likely a go-amz issue [02:30] it says it does, but not at s3.a....com it needs to be at s3.us-east-2....com [02:31] ok, ok, so you'll need to add update go-amz to ensure it has the right endpoints [02:31] wallyworld: right. My guess too, I'll dig in to goamz tomorrow [02:31] ok [02:32] thanks wallyworld much appreciated. Wanted to make sure it wasn't something simple [02:32] redir: tis ok, i didn't tell you much except propose a guess or two :-) [02:34] wallyworld: sometimes it's nice when the rubberduck squeaks:) [02:34] depends on how hard you squeeze it :-) [02:34] normally I quack [02:35] :) [02:57] wallyworld: ship it [02:58] menn0: tyvvvvvvm for looking [02:58] wallyworld: opening the diff nearly killed my browser [02:58] menn0: another reson why github reviews suck [02:58] with rb, you get a paginated list of files [02:59] i don't get how people navigate gh reviews [02:59] without pulling their hair out [03:04] wallyworld: I just had a quick play with the gerrit instance for the chromium project. it's pretty nice, if intimidating at first. [03:04] wallyworld: knobs and buttons for everything [03:04] wallyworld: but fast and has good keyboard shortcuts for navigation. [03:04] menn0: would love to seriously consider that [03:05] but people need to get over the NIH syndrome [03:05] which I think will be impossible for our group :-( [03:07] wallyworld: I don't know about that. I can see why some people are attracted to GH reviews. [03:08] i am having trouble. they are really hard to navigate (all one big page), and so much wasted space [03:08] and no way to manage comments [03:09] the main argument seems to be "well it might get better sometime" [03:09] but it's crap *now* [03:11] w00t [03:11] wow [03:11] that looks really out of place now [03:11] my view had this as the last comment: [03:11] yay vsphere provider working correctly [03:11] then I scrolled down [03:11] heh [03:36] amz bits ready for review: PTAL https://github.com/go-amz/amz/pull/72 [03:40] wallyworld, if you are still around can you provide a review for redir ^^ [03:41] can do in a minute [03:41] wallyworld, thank you! [03:41] ekanite looks pretty nice. A syslog server that accepts msgs over UDP/TCP [+TLS] parses RFC5424 headers, does full text indexing, and manages retention [03:42] redir: did you sort the s3 issue? [03:42] also sorts by log timestamp rather than the order received [03:42] wallyworld: yup [03:42] greta [03:42] great [03:42] and will have a juju 1.25 PR when the AMZ bits are merged [03:42] wallyworld: ^^ [03:43] oh and the ekanite binary is 12MB [03:44] redir: did you researc that us-east-2 requires the region bools to be true ie bucket lower case etc? [03:44] wallyworld: that fixed the issue [03:44] and US East is the only region that doesn't [03:44] in amz [03:45] us-east-1 [03:45] interesting [03:46] redir: lgtm [03:46] See https://github.com/reedobrien/amz/blob/2eab8e64ed1675ae196f785bbc2d293e45aaf7a1/aws/aws.go [03:46] everything else is not false.... so looks like they were lax on the original region [03:46] and changed hteir minds for everything else [03:47] wallyworld: my guess it that since bucket can be used in the website url they wanted to make it consisten with case insensitivity... [03:48] after the original s3 in 2006 was let out into the wild [03:48] sounds plausible yeah [03:48] can we merge amz? [03:48] does all the $$merge$$ stuff work there? [03:49] should do [03:54] k. tx [08:21] dooferlad, frobware, macgreagoir: I'd appreciate a review on https://github.com/juju/juju/pull/6481 fixing bug 1616048 ta! [08:21] Bug #1616048: Create interface for ofono [08:22] oops - bug 1616098 [08:22] Bug #1616098: Juju 2.0 uses random IP for 'PUBLIC-ADDRESS' with MAAS 2.0 <4010> [08:22] dimitern: will do - otp [08:22] frobware: no rush [08:22] ;) [10:39] * dimitern steps out for ~1h [12:28] frobware: ping [12:31] macgreagoir: hey [12:31] macgreagoir: got some time for a review? :) [12:33] dimitern: I think ur PR do not have a check run coz they r awesome :) m sure that's what ballons coded in the script... "if dimitern > don't check" :) [12:33] anastasiamac: :D [12:33] anastasiamac: that's a nicer way of looking at it, sure ;) [12:34] anastasiamac: I'd appreciate a review as well hehe [12:38] dimitern: i'd love to review but m furiously working to remove myself from keyboard: it's 11pm \o/ if it can wait til my morning -m happy to do it then :D [12:38] anastasiamac: uuh you should definitely get some rest! [12:39] dimitern: \o/ will soon :D thank you for care and concern :) [12:40] anastasiamac: any time ;) [12:45] rick_h_: ping [12:46] dimitern: pong [12:46] rick_h_: I thought about picking up bug 1580501 tagged 1.25 .. it seems it's only related to 2.0 though [12:46] Bug #1580501: cloudimg-base-url parameters not in Juju2 anymore <4010> [12:46] dimitern: please stick with cards on the board [12:47] rick_h_: that's the first one on the board :) [12:47] rick_h_: top left [12:48] oh wtf, the title of the bug and the card got out of wack [12:48] dimitern: ugh, ok this is a can of worms because this falls under work planned for the lxd image cache stuff, and yes is 2.0 only [12:48] dimitern: so going to yank that one off the board [12:49] rick_h_: +1 I'll look into bug 1589680 then [12:49] dimitern: one sec, let me see if I did bad copy/paste voodoo [12:49] Bug #1589680: Upgrading to cloud-archive:mitaka breaks lxc creation

[12:49] dimitern: https://bugs.launchpad.net/juju-core/+bug/1560487 [12:49] Bug #1560487: local provider fails to create lxc container from template

[12:49] dimitern: I pasted the wrong bug number to the card, my bad [12:50] dimitern: apologies for the confusion [12:50] rick_h_: right :) np - do you want me to look into that last one instead? [12:50] dimitern: either one is ok thank you [12:51] rick_h_: ok - 1589680 it is then [12:52] * dimitern dusts off his 1.25 branch .. it's been a while [12:52] hm, we're not doing staging and such like with 1.25 I presume? [12:52] mgz: no, we're not [12:53] it's free for all :D [12:55] frobware: ping, can I grab a sec when you're available? [13:04] dimitern: You nean 6481? Sorry, I started and got distracted by my own :-) [13:10] dimitern: I'd expect that we can just upgrade the lxc in order to find a path forward [13:10] dimitern: so if we deploy something on trusty/default lxc and it's pre-2.0, then manually add the PPA/upgrade lxc on there...how would Juju know to use/follow newer lxc patterns [13:11] macgreagoir: np, if you can have a look at it - great! [13:12] rick_h_: I'd expect setting openstack-origin to cloud-archive:mitaka will trigger adding the repo and the lxc upgrade.. [13:12] will know soon anyway.. [13:13] dimitern: right, but my point is you can do this with the ubuntu charm [13:13] dimitern: no need for anything openstack [13:13] dimitern: fewest moving parts/complications the better [13:13] rick_h_: yeah, I guess so.. double checking c-a:m's lxc version.. [13:13] dimitern: just check the version in trusty vs the lxc ppa for trusty [13:14] dimitern: if you can get a pre-2.0 in default trusty and then get post 2.0 via the ppa you can "do the upgrade" and chase down what we need to help update to make it work [13:16] yeah: lxc 2.0.5-0ubuntu1~ubuntu16.04.1~cloud0 is in cloud-archive:mitaka [13:16] and lxc is not there in c-a:icehouse [13:17] dimitern : afaicr the error was when going from c-a:trusty-liberty -> c-a:trusty-mitaka [13:18] And orig. install was plain trusty, sans c-a [13:18] chrome0: yeah, I'm trying exactly that now :) [13:18] +1 [13:19] rick_h_: post 2.0 lxc won't work with 1.25 AIUI [13:19] dimitern: can you hop in the standup real quick? [13:19] owm [13:23] rick_h_: yep, back [13:24] frobware: meet you in the standup room please? [13:24] rick_h_: omw [13:28] is there a good way to find out when a relation is completely removed? [13:28] i find it really hard to test adding and removing relations [13:32] simply looking at the departed/broken hooks doesnt seem to be enough to determine the actual status of the relation in juju [13:34] natefinch: ping for chat [13:36] rick_h_: oops coming [14:00] frobware: dimitern natefinch ping for standup [14:41] dooferlad: yo, free for bothering? [14:45] rick_h_: I think we have a nice conclusion for the openstack bug [14:45] rick_h_: https://bugs.launchpad.net/juju/+bug/1621590/comments/18 [14:45] Bug #1621590: openstack provider ignores a properly created bootstrap machine [14:47] dimitern: now looking at your PR :) [14:50] mgz: yt? [14:50] frobware: sweet! :) thanks! [14:51] mgz: can you $$merge$$ this https://github.com/go-amz/amz/pull/72 or add me to the right group, or tell me who to ask to get added to the right group? Please:) [14:51] dooferlad: ping [14:52] redir: sure, I'll take a look [14:52] rick_h_: hi [14:52] mgz: sorry, still talking to Mick [14:52] dooferlad: heads up, assigned another bug your way. It's another config param getting ignored like your current one. [14:52] rick_h_: ok, thanks [14:52] dooferlad: the possible idea is that it's a more generic problem and that two birds/one stone and such [14:52] dooferlad: and it was brought up on the cross team call by the stakeholders there that it should be replicatable with canonistack [14:53] dooferlad: so if getting OS goign is a burden please dump that and run with an already running OS [14:53] redir: heh, that's a group I'm not actually owner on, but I can trigger the merge for you [14:53] dooferlad: bah, LP won't let me change the assigner atm, timing out. https://bugs.launchpad.net/juju/+bug/1614239 [14:53] Bug #1614239: bootstrap-timeout ignored in --config [14:59] redir: merged as 7754380 [15:06] dimitern: what were the rules on 500 lines of diff? You're dangerously close. :) [15:07] frobware: I really tried to minimize the changes :/ [15:08] frobware: the only "luxuries" are a few added tests, but the refactoring of the common code / tests *did* reduce the original diff by a 100 lines :) [15:10] dimitern: have time to HO? [15:11] frobware: yeah [15:12] frobware: I'm in 'core' [15:13] dimitern: omw [15:15] dooferlad: so... free now? [15:20] mgz: thanks [15:20] mgz you know who the owner is? [15:22] redir: canonical-is and niemeyer [15:22] I don't know who has access to the canonical-is role account [15:26] mgz: sorry - have a headache so bad I want to be sick. Can we talk tomorrow? Perhaps invite me to a meeting so we have a slot booked? [15:26] dooferlad: no worries, I'll set something up for tomorrow [15:26] mgz: thanks - much appreciated [15:26] * dooferlad goes to hide in a dark room [15:33] another PR ready for review; https://github.com/juju/juju/pull/6483 PTAL [15:33] bbiab after dentist [15:37] hello, anyone got some time guiding me on how to test adding and removing relations? [15:38] i was using the relation function amulet to test wether or not the relation was sucessfully removed, and that worked, but now it doesnt anymore [15:38] would really appreciate some help! [15:42] dimitern: frobware: hey, is this robust? i.e. will it do the right thing in complicated network setups? https://github.com/juju/juju/blob/staging/environs/manual/addresses.go#L17 [15:46] abentley: was that update for Ohio because of the amz release that happened? [15:46] this morning [15:47] redir: No, it was because AWS announced the region on the 17. [15:48] i see, thanks abentley [16:14] anyone, PTAL https://github.com/juju/juju/pull/6483 [16:37] axw, katco, https://github.com/juju/juju/pull/6414 [18:11] ls [18:11] whoops [18:45] ahh the glorious feeling when you've gotten down in the weeds typing code and you finally get it back to the state where it'll gofmt. [18:46] looooool [19:02] * redir lunches [19:10] made a new struct that helps you ask the user repeated questions.... called it Pollster. Figured it was appropriate given current events. [19:12] As long as it produces output that isn't actually relevant to anyone, natefinch [19:13] and diferent results fo rthe same input by different users [19:14] heh [19:52] rick_h_: http://pastebin.ubuntu.com/23355482/ [19:53] natefinch: coolio [19:54] rick_h_: so, one reason to use add-cloud for ec2 etc ..... it's the only way to save a cloud-specific config [19:54] rick_h_: instead of haviong to do bootstrap --config=myconfig.yaml [19:54] natefinch: yea, but I think we need a different approach for config as a whole [19:54] natefinch: because config is across all providers and that's more of an "edit cloud" than an add-cloud and such [19:54] natefinch: so I've punted that for now because adding it here is a hack imo [19:55] rick_h_: that's fine. I can pare the list of clouds back to just the custom ones [19:55] natefinch: though I wonder if we can bootstrap by adding an empty config section whenever we write it out [19:55] natefinch: so that it's ready to be edited ootb [19:57] well, you don't really need a config section [19:57] natefinch: I mean that you can specify a config key to the cloud definition and then add overrides/etc to that file [20:02] rick_h_: one problem I'm not quite sure how to solve from a UX standpoint is how to query for "choose M of N", like for authentication types, you can choose 1-3 of 3 choices. [20:13] rick_h_: I guess just asking for them comma separated will work for our needs [20:14] natefinch: yea, I think pick a format and to validate it [20:31] easy review https://github.com/juju/juju/pull/6483 PTAL [20:32] redir: you haven't got anyone to bite yet? I'll take it. [20:35] redir: you took the pricing stuff from the webpage, or andrews's new generation thing? [20:36] mgz the web page:/ Didn't know about a generator [20:36] redir: it's what I did last time, I think that's reasonable for 1.25 [20:36] we know it's not always up to date but should be good enough [20:36] redir: lgtmed [20:36] mgz: tx [20:37] ...why did I give andrew an extra s... [20:38] channelling Gollem? [20:39] preciousss anddrewsss [20:39] ^ nice [20:40] rick_h_: forgot about a school meeting I have in the morning at 9am. Probably will miss standup, though I might get lucky and it won't take forever. Doing well with the add-cloud stuff, though. [20:43] natefinch: k [23:13] wallyworld: is there no new price list info for us-east-2? [23:14] axw: not yet, that bit was not updated in juju [23:14] so far, just the region and endpoints have been updated so it works [23:15] wallyworld: yeah, it's just that it only works because of a fall-back. it could be inaccurate [23:15] anyways, nice that it works :) [23:15] yeah [23:16] wallyworld: hm, there is info for ohio in the latest index [23:16] I'll update [23:16] ta [23:16] ... right after I remember to go to standup === hml_ is now known as hml