[00:05] natefinch, you still around? === kadams54 is now known as kadams54-away [00:31] fwereade_: OMG...how big is that? [01:09] thumper: size envy? [01:10] thumper: there seem to have been an intent change for 85K lines yesterday?.. no? ;-) [01:34] anastasiamac: just wait... [01:34] patience... [01:34] * anastasiamac waits for thumper? [01:38] * thumper runs the tests one more time [01:39] 518 files changed, 9228 insertions(+), 8964 deletions(-) [01:39] $ git diff master | wc -l [01:39] 63618 [01:39] WINNING [01:39] \o/ [01:42] * anastasiamac rolls eyes [01:42] thumper: yes, urs is bigger === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [01:49] https://github.com/juju/juju/pull/1219 [01:49] * thumper wonders if review board will shit itself... [01:50] thumper: do i need to change all jc.IsNil that i have put in the last couple of days too? [01:50] anastasiamac: afraid so [01:50] anastasiamac: once this lands [01:50] anastasiamac: my branch changes those that have landed [01:51] anastasiamac: simple global replace though :) [01:51] thumper: yes, global replaces are always simple... === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [01:53] github says: Sorry, we could not display the entire diff because too many files (517) changed. [01:53] http://reviews.vapour.ws/r/528/ [01:53] haha [01:53] 26 pages of diff [01:54] global search and replace sounds trivial to me... [01:55] pitty we don't have the "trivial" review tag on this project [01:56] wallyworld: you there? [01:57] axw: there? [01:57] thumper: I am [01:57] axw: hey there [01:57] wallyworld went to get lunch [01:58] axw: trade you a bug fix (for the thing I broke on Friday) for a +1 on http://reviews.vapour.ws/r/528/ [01:58] axw: everything you need is in the first of the 26 pages of diff [01:58] very mechanical change [01:58] heh [01:58] ok [01:58] the hardest bit was adding the import to the files that missed it [02:00] thumper: done [02:01] axw: ta [02:01] * thumper goes to fix jujud [02:38] _thumper_: hi === _thumper_ is now known as thumper [02:39] oh hai wallyworld [02:39] oh, looks like i missed out on your mega diff [02:40] axw: here is the fix for bug 1395564 http://reviews.vapour.ws/r/529/ [02:40] Bug #1395564: jujud constantly spews `juju.worker runner.go:219 exited "identity-file-writer": StateServingInfo not available and we need it` [02:40] wallyworld: you did [02:40] \o/ [02:41] wallyworld: I have to land it asap as I feel it will conflict with everyone :) [02:41] indeed [02:42] wallyworld: also, I think I have found out why I get no all-machines.log [02:42] the rsyslog config is fubared [02:42] wouldn't surprise me [02:42] i get all-machines.log [02:42] wallyworld: really? [02:42] i think so [02:42] on local with master? [02:43] can I get you to check plz? [02:43] i tought so, but would need to check [02:43] sure [02:45] thumper: [02:45] $ sudo ls /var/log/juju-ian-local/ [02:45] all-machines.log ca-cert.pem logrotate.conf logrotate.run machine-0.log rsyslog-cert.pem rsyslog-key.pem [02:45] ah phooey [02:46] I wonder why I don't get it [02:46] ... [02:46] nfi [02:46] rsyslog is kinda cryptic [02:54] thumper: have a look in /etc/rsyslog.d and see if you have a config file left behind that shouldn't be there [02:55] thumper: that's what happened to me [02:55] menn0: I have an old running environment [02:55] thumper: that's the problem [02:55] which I probably don't really need... [02:55] thumper: the way we set up rsyslog doesn't work with multiple local envs [02:56] thumper: i've reported this under 1387388 [02:56] bug 1387388 even [02:56] Bug #1387388: rsyslogd configuration does not support multiple local environments

[02:57] thumper: different configs get written out for different envs but using the same rsyslog port [02:57] thumper: rsyslogd picks one [02:58] thumper: and the other config(s) don't work [02:58] ah... [02:58] yes... [02:58] that makes sense [02:58] another thing that fixing the local provider will fix [02:58] thumper: exactly [02:59] thumper: this also causes the rsyslog workers on each machine and unit to continually restart, in the env who's config didn't win [03:00] thumper: because the certs used by rsyslogd are for the other env and don't match what the rsyslog workers are trying to use [03:00] haha [03:00] oops [03:01] menn0: there is an rsyslog port that can be configured [03:01] but I think this is my problem of two different users using the came default config [03:01] thumper: cool, well that's the workaround I guess [03:04] axw: u have ben so instrumental to block/unblock cmd so far [03:04] axw: and everyone is asking u to review stuff today [03:04] heh [03:04] * axw braces himself [03:04] axw: and u r so attentive during reviews [03:04] axw: ;-) [03:05] maybe I'll just swap my OCR day... [03:05] axw: could u plz, pretty plz, have a look at http://reviews.vapour.ws/r/523/ [03:05] axw: swap is not my call but it seems fair :-p [03:06] anastasiamac surely knows how to butter people up... [03:06] heh... [03:06] * anastasiamac puzzled as to what thumper implies... [03:06] oh lord, you are so big, we're all really impressed down here I can tell you [03:06] anastasiamac: I'll take a look after I'm done with wallyworld's [03:07] axw: thank you sooooo much :-0 [03:07] anastasiamac: it is the 'u r so attentive', and 'pzr, pretty plz'... [03:07] axw: anastasiamac: i'm already reviewing [03:08] ok [03:08] wallyworld: thnx ;p [03:10] thumper: butter is an important cooking ingredient used for food [03:10] * anastasiamac would never dream of puuting food ingredients on ppl [03:19] * thumper does a little dance [03:19] mega branch landed [03:25] thumper: why do you have to wait till after upgrade to write the identity file? [03:25] thumper wins the award for the biggest diff EVER [03:27] axw: because the upgrade steps were potentially changing it [03:27] wallyworld: I'll think about one that would be bigger still!!! [03:27] ahh [03:28] thumper: it's not the size that counts, right? [03:28] wallyworld: you keep thinking that... [03:28] thumper: wouldn't it make sense just to write it in the upgrade step then? [03:28] oh... it *is* written in there... I'm lost [03:28] axw: this is run every startup [03:29] axw: on the off chance that a machine goes from not a state machine -> a state machine [03:29] axw: I didn't write that bit [03:29] axw: just had to make sure it worked [03:29] agent.WriteSystemIdentityFile is called on startup already, in StateWorker [03:30] so between that and the upgrade step which also writes the file, I don't understand the point of the additional worker... [03:55] axw: um... I thought it was removed from state worker [03:55] it should have been [03:55] IIRC [04:03] thumper: sorry, was looking at my out of date branch [04:11] thumper: reviewed [04:11] ta [04:30] * thumper goes to make dinner [04:30] later folks [04:42] RB is flaking out on me [04:42] I've hit save after editing a comment and it's stuck "Loading..." [04:42] I've tried twice now === kadams54 is now known as kadams54-away [08:19] morning [08:21] TheMue: morning :-) [08:22] anastasiamac: oh, still online? should be quite late in your TZ, or am I wrong? [08:23] TheMue: 6pm... was going to induction but m usually online later (around 9pm+) - u r rite [08:23] anastasiamac: ah, ic [08:24] * TheMue looks out of the window, 1° and a bit foggy. *brrr* [08:26] TheMue: it's 6pm, very very very hot outside. we r in the aircon set for 21C and it barely copes :-( visit Australia [08:27] anastasiamac: oh, you're inviting me? :D sure I would like to, never have been so far away from home and especially in the southern hemisphere [08:29] TheMue: the best bits r inverted seasons = Xmas on z beach :p [08:30] anastasiamac: trying to imagine this, would feel strange. :) [08:47] * fwereade_ bbiab [08:53] dimitern: I might be a couple minutes late to the interview, my son is coming home right now on the bus [08:53] jam, not to worry, I've prepared questions, etc. [08:54] jam, dimitern: I can wait until you are both ready if you like. [08:55] jamestunnicliffe, hey! welcome :) [08:55] dimitern: hi :-) [08:55] jamestunnicliffe, no, it's ok - we can start on time and jam will join us [08:56] jamestunnicliffe, we do have a standup meeting right after :) [08:56] dimitern: great. See you in a moment. [09:22] rogpeppe: ping [09:22] TheMue: pong [09:22] morning all [09:22] rogpeppe: I migrated your codec checkers to juju testing. do we have tests for the checkers? meta-testing :D [09:23] rogpeppe: otherwise could you take a look at https://github.com/juju/testing/pull/39. they are simply copied [09:24] TheMue: no, we don't have tests, but they've been used a lot [09:25] TheMue: i would suggest one change when copying them, that i've been meaning to make for a while. [09:25] rogpeppe: so kinda verification by usage :) === psivaa-holiday is now known as psivaa [09:25] rogpeppe: I'm listening [09:25] TheMue: i'd suggest that you make the first argument to the checker a string rather than a []byte [09:26] TheMue: that way if the test fails, it prints the JSON in string form rather than a load of hex byte values [09:26] rogpeppe: would make the usage more convenient, yes [09:26] TheMue: feel free to write some tests if you like [09:27] rogpeppe: ok, will change and then add tests. I'll ping you later again for a review. [09:41] dimitern, ping? [09:53] dimitern: firefox just crashed on me [09:53] give my apologies [09:54] jam, it's ok I think - my chrome almost crashed as well :) [09:54] mattyw, hey, sorry was in a meeting [10:07] jam: dimitern: TheMue: got logged out [10:07] back in a minute [10:17] dimitern: are you around most of today - or are there times you're in meetings? [10:17] dimitern: I have some little "detaily" questions I need to ask at some point [10:17] dimitern: I might collect a few of them together... [10:17] questions that is [10:20] voidspace, I have a 1:1 with alexisb in 4.5h, apart from that I'm available [10:20] dimitern: thanks - appreciated [10:20] dimitern: for AddSubnet - rather than taking 8 parameters, I've created a SubnetInfo struct to use as the arg passed in [10:21] dimitern: this mirrors what AddNetwork does [10:21] voidspace, +1 - we can probably use the same struct for updating existing subnets [10:21] dimitern: cool [10:27] dimitern: "subnets" for the collection name? [10:27] dimitern: and will we need an upgrade step to create the collection - or will it be created on first use? === ashipika is now known as ashipika_afk [10:30] voidspace, "subnets" sgtm [10:31] voidspace, no need for upgrade step - the collection will be created along with all the others when we try to insert for the first time [10:34] dimitern: cool, thanks [11:34] rebooting [11:54] wallyworld_: are we still on for storage discussion in 5 min ? [11:54] yep [12:02] jam: fwereade_: meeting? [12:30] fwereade_, when you have 2 minutes [13:36] does anyone know if debug-log is supposed to work correctly on local environments? [13:37] fwereade_, dimitern: ^ [13:39] rogpeppe, it should work if you have all-machines.log [13:39] dimitern: i'm only seeing results for machine-0 [13:40] rogpeppe, in all-machines.log ? [13:40] dimitern: yup [13:41] sw [13:41] % sudo grep -v '^machine-0:' all-machines.log | wc [13:41] 0 0 0 [13:41] % sudo wc all-machines.log [13:41] 29962 324235 4314099 all-machines.log [13:41] rogpeppe, that would be a bug then :/ [13:41] rogpeppe, hmm it looks like a bug :) [13:41] ok, i'll report it [13:42] i'm guessing not many people use debug-log in the local env [13:42] rogpeppe, I haven't even tried - sudo less ~/.juju/localenv/logs/machine-0.log is easier [13:43] dimitern: i don't have a localenv dir [13:43] dimitern: i was looking in ~/.juju/local/log/all-machines.log [13:44] dimitern: i was writing a document and wanted to give instructions that would work for both a local or a remote env [13:44] rogpeppe, localenv is the name of my local env [13:44] dimitern: ah [13:44] dimitern: i hadn't realised that dir was named after the local env [13:45] jam, you still around? [13:45] rogpeppe, yeah [13:49] fwereade_, dimitern: https://bugs.launchpad.net/juju-core/+bug/1396159 [13:49] Bug #1396159: debug-log only reports machine 0 entries in local environment [13:57] rogpeppe, thanks! [13:57] dimitern: np [13:57] dimitern: well, only minor probs :) === ChanServ changed the topic of #juju-dev to: https://juju.ubuntu.com | On-call reviewer: see calendar | Open critical bugs: None [14:18] * rogpeppe lunches [14:30] natefinch, perrito666 cloudbase call [15:22] natefinch: ping [15:28] ericsnow: howdy, I'm in moonstone [15:28] k [16:33] time-appropriate greetings everyone! :D [16:33] could I please get a quick second opinion on this PR please: https://github.com/juju/juju/pull/748 === kadams54 is now known as kadams54-away [16:44] fwereade_, how's leader el coming? [16:46] Hey guys, I'm not 100% sure of the answer to the question in #juju, is there a migration path for 0.7 => 1.X ? [16:49] marcoceppi, no [16:49] hazmat: thanks [16:50] marcoceppi, i think thats the first user request i've seen for pyjuju upgrades [16:50] to core [16:50] hazmat: the mythical beast exists! === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away [17:55] hazmat, katco can update you on what she has been doing recently for leader elections [17:56] hazmat, any reason for your question? [17:58] alexisb, desparate need to avoid hacks in charms [18:02] alexisb, just caught up date via priv message [18:41] g'night all [19:06] so i need to move MachineAgent out of main and into an actual package [19:06] any opinions on where it might best be placed? [19:07] right now it's in cmd/jujud/ which is in "main" [19:41] katco: maybe /agent? no idea if that might cause circular references, though [19:44] natefinch: root to juju/juju? (i'll take care of any dependency graphs, just looking for logical placement) [19:46] it does seem incorrect to keep it resident as a child of cmd/jujud [19:46] since that is just that -- building an executable [19:51] katco: yeah, the stuff in package main should be about *the executable* not any other logic. to me, that means stuff to parse the CLI and that's about it [19:52] natefinch: ok, so we're talking about github.com/juju/juju/agent? [19:52] katco: seems reasonable [19:52] cool... i'll ping dave tonight if i get a chance since i know he's been looking at this stuf [19:52] f [19:53] (dependencies etc) [19:53] cool. He'll probably have an idea too [19:53] i thought he might [20:02] Hey -- what is the current state of juju networking on 1.21? is there any docs about it? [20:15] dpb1: I don't think we released the network stuff yet. The 1.21 release notes don't talk about networks. [20:16] natefinch: ok great. Just noticed it leaked out to juju status command then. :/ [20:17] thumper, menn0: the func I was talking about is transformId in state/watcher.go [20:17] / transformId converts a global key for a ports document (e.g. [20:17] / "m#42#n#juju-public") into a colon-separated string with the [20:17] / machine id and network name (e.g. "42:juju-public"). [20:17] hrm, I don't see it anymore... must be smoking something. n/m natefinch [20:17] dpb1: heh, ok good. Wouldn't be the first time the release notes were lacking something and/or something slipped into the CLI before it was really ready. [20:18] why would that be needed? [20:26] waigani: it was introduced by dimiter in 14b64091 [20:26] waigani: the commit message says "openedPortsWatcher no longer exposes global keys" [20:26] waigani: I suspect is so that global keys aren't reported over the API [20:30] katco, dave is out on vacation, just fyi [20:30] back on monday [20:30] alexisb: oh doh. ok [20:30] not a big deal. ty for the info! [20:30] menn0: okay, thanks. I'll focus on working out how the localID is getting pasted into the watcher merge func [20:41] * katco pines for her two large screens @ home. [20:42] the laptop screen just isn't the same. [22:07] /query natefinch [22:07] ouch lol [22:23] * thumper sighs [22:23] * thumper stabs rsyslog in the face [22:36] Is anyone versed enough into deploying your own python soft in windows and want to chat a couple of minutes with me to help me deploy a floss project without going crazy? [22:36] I know many of us pack a windows past [22:37] s/windows/python [22:37] I tried to use windows last night [22:37] OMG what a PITA [22:37] so hard to get anything done [22:37] WTF [22:38] I only use windows to create ninja ide installer, but I am having some terribe pains packaging python for windows when my python does complex enough things [22:38] katco, perrito666: do either of you have a local environment running with more than just the bootstrap node? [22:38] thumper: nope [22:42] menn0: just proposing feature flag branch, I'll be taking the discussion to the list, so don't feel like you have to review it now [22:42] menn0: I won't land it until we've had a chat [22:43] where we is the team [22:46] thumper: how does that feature flag work out? I was just talking with wayne about those the other week so curious. Did it get handed off your way? [22:46] rick_h_: um... just writing an email to the list [22:46] rick_h_: been planning on writing it for a while [22:46] thumper: ok, I'll look forward to it [22:48] rick_h_: email sent [22:48] * thumper heads to the gym [22:48] rick_h_: aren't you supposed to be on holiday this week? [22:48] thumper: ok, so this looks less like charm flags and more feature flags [22:48] thumper: I would be if I could get my work deployed [22:48] thumper: cool, thanks for this PR as we're doing feature flags in python and looking at it in Go so this might be useful as well [22:48] rick_h_: yeah, nothing to do with charm flags [22:49] thumper: have fun at the gym, I'll point my guys at this and might want to chat with you/etc [22:49] rick_h_: we could push the generic bits into some utils package [22:49] rick_h_: rather than in juju-core [22:51] thumper: gotcha, yes, please. It's something we're eager to put into place as you can imagine with our use cases [22:52] thumper: I've sent the team an email to follow up with you and look it over for our own uses tomorrow. [22:53] thumper: no problems === fuzzy_ is now known as Fuzai [23:08] thumper: sorry for the delay. i have a manual environment running with lxc hosts on another physical machine.