[00:01] anastasiamac, wallyworld is holding me up ;) [00:01] I will be there soon [00:01] alexisb: np - gives me more time to get ready [00:09] thumper: wow, i am never visiting NZ. bleck: http://www.buzzfeed.com/nicholaswray/nz-sux#.lweRbQN7Zl [00:18] katco: yeah, it is terrible here [00:20] fwereade, et al, the run-unit-test script will now retry tagging ec2 instances so that you will see fewer failures setting up in git-merge-juju [00:21] fwereade: the bot hates you [00:22] thumper, doesn't it just [00:22] fwereade: wanna chat? [00:22] thumper, although I love it, because one time it was doing its job and preventing me from checking in shite [00:22] thumper, just a quick one, don't want to distract myself too much ;) [00:31] thumper, you beat me too it [00:31] fwereade, I think you need to send the bot some flowers and chocolates [00:32] or maybe some scotch?? [00:32] alexisb, for sure :) [00:32] alrighty all, I am off to pick-up kiddo will be back tonight [00:59] fwereade: I see it merged [01:02] thumper, one of them :) [02:23] axw: one thing i forgot to mention in standup - tests on windows are broken because of various reasons, one of which is the loop stuff [02:23] wallyworld: ah, doh :( [02:23] I have no Windows machine ... [02:23] i thought about it ages ago and then forgot [02:23] yeah, i'll add a build directive [02:31] axw: actually, looking at the code again, i recall we abstracted out all the *nix specific stuff in tests behind stubs/mocks, i think it's just filepath stuff that breaks things [02:31] wallyworld: is there a windows build job whose output I can look at? [02:31] yeah, that's where i'm heading now, i believe there is [02:35] Bug #1428430 was filed: AllWatcher does not remove last closed port for a unit, last removed service config

[02:50] menn0: would you mind taking another look at http://reviews.vapour.ws/r/1060/ [02:52] ericsnow_: looking === ericsnow_ is now known as ericsnow [02:53] menn0: thanks! [02:56] axw: only 2 failing tests :-) trivial http://pastebin.ubuntu.com/10534173/ [02:56] wallyworld: cool [02:57] i'll fix (without windows machine but should be ok :-) [02:58] wallyworld, there is detailed instructions on the cloudbase site for testing windows workloads with juju [02:58] let me get the link [02:59] http://wiki.cloudbase.it/juju-testing [03:01] cmars, jam leads call? [03:06] ericsnow: looking good to me. have you done some manual tests against actual systems with different init systems to make sure this all works? [03:06] menn0: my local system (upstart) and a vm running systemd [03:06] ericsnow: ok cool. then ship it. [03:07] ericsnow: there's just one thing I raised (more an idea than an issue) [03:08] menn0: I had considered it (reusing identifyExecutable), but wasn't convinced it would always be the same [03:08] Bug #1428439 was filed: retry-provisioning launches instances for containers; cannot retry containers at all [03:09] ericsnow: yeah, that's why I suggested it with some uncertaintity [03:09] ericsnow: just leave it [03:09] menn0: k [03:10] menn0: thanks for the reviews [03:11] ericsnow: np [04:20] axw: wallyworld: https://github.com/go-amz/amz/pull/34 [04:23] anastasiamac: looking [04:30] now that's much better - from 58 PRs down to 38 [04:31] * thumper is done [04:31] laters [04:42] axw: wallyworld: tyvm for review :D all updated but I do not have merge rights on goamz... [04:43] np, i can do it [04:43] anastasiamac: merged [04:43] one more branch to go [04:44] axw: did you have anything to say wrt that maas doc the other day? [04:45] wallyworld: I added comments already [04:45] ah, ty. i need to learn how to use refreah i think [04:45] wallyworld: essentially, LGTM. [04:45] great [04:46] wallyworld: when u say "1 more branch", which one u have in mind? ec2 pricing? [04:46] anastasiamac: updating the current juju core branch to add instance types, depes, and pricing yeah [04:46] wallyworld: k :D === kadams54 is now known as kadams54-away [05:13] axw: wallyworld: pricing branch http://reviews.vapour.ws/r/1073/ [05:13] already looking [05:13] wallyworld: i *think* i've fixed bad formatting ... [05:13] wallyworld: tabs vs spaces, etc... [05:14] yep, looks much beter [05:16] anastasiamac: and there's no cost info for chinese regions? [05:17] wallyworld: i could not find any chines prices... maybe they r on chinese site but I do not read chines :( [05:17] wallyworld: i have looked at chines site in english but nothing... [05:18] chinese* [05:18] hmmm, i can't see how juju would work in china then as instance selection is broken without cost info from what i can see [05:18] do we have any means of testing ?.. [05:19] not easily [05:19] wallyworld: unless we put 0 for everything?... [05:19] it's mor ethan just cost, it's also knowing what instance types are available [05:21] wallyworld: would it be a separate bug tho? so this PR added frankfurt for bug 1428117 and c4+others for bug 1427840 [05:21] Bug #1428117: EC2 eu-central-1 region not in provider [05:21] Bug #1427840: ec2 provider unaware of c3 types in sa-east-1 [05:21] yeah [05:22] anastasiamac: fix axw's issue and then ship [05:25] anastasiamac: chinese instance types http://www.amazonaws.cn/en/ec2/details/ [05:25] so we can add those before shipping [05:26] set all costs to 0 or something if no cost info can be found [05:26] wallyworld: k but what shall i add for price? 0...? [05:26] wallyworld: k [05:35] anastasiamac: school pickup time back later, andrew can check the final branch before landing [05:35] wallyworld: school pickup but will try to repropose for axw :D [05:37] axw:wallyworld: btw, i'll try to do the same with USGov region as with China - just add instance types with 0 cost === Murali_ is now known as Murali [05:49] per special request by marcoceppi for 1.23 - juju action status now returns multiple matches or all if no prefix is given: OCR PTAL http://reviews.vapour.ws/r/1085/ [06:49] jw4: added some comments [06:59] wallyworld: responded to your comments on my branch, would appreciate another look when you've got time [07:26] axw: wallyworld is delayed - car broke down :( [07:26] doh [07:27] axw: yep :( dunno what's with the car but it's realllly hot here today [07:42] Bug #1425788 changed: multiple definition of http.HandlerFunc

[08:21] Bug #1292536 changed: maas provider apt-get install bridge-utils should be better

[08:21] Bug #1377262 changed: please expose network-bridge to manual provider also

[08:51] Bug #1314442 changed: MAAS non-deterministic private-address with non-ethN interfaces

[09:17] axw: sorry about delay, car broke, heavy traffic, waited ages for tow truck [09:18] wallyworld: no worries. that sucks :( [09:18] yeah, clutch cylinder i think [09:26] wallyworld: doh, I think anastasiamac may have forgotten to upload her changes before merging [09:27] other things weren't updated either [09:36] axw: yeah, i figured that after seeing another issue that was unresolved [09:49] anastasiamac: I'm definitely looking at the latest diff on RB [09:49] I'll check GitHub [09:49] wallyworld: axw: my bad - did not push :D [09:49] fixing it noww... [09:58] axw: wallyworld: probably cleaner as separate pr?.. :D http://reviews.vapour.ws/r/1087/ [09:59] anastasiamac: LGTM [09:59] axw: :P [10:00] axw: brain-numbing experience over :D tyvm for urs and wallyworld attention to detail!! [10:00] anastasiamac: :) [10:05] jam: team meeting, if you want to join [10:23] TheMue, dooferlad_, guys let's do our standup now? [10:23] dimitern: sure === dooferlad_ is now known as dooferlad [10:23] dimitern: omw [10:24] I'm in the hangout === frankban__ is now known as frankban [12:29] katco: I'd appreciate another look over my branch in your day, I've responded to your comments - thanks [14:21] axw: I assume you're off now, but thanks for the review [14:23] TheMue: thanks for your review - do you want to give it one last pass before merging? http://reviews.vapour.ws/r/1085 [14:24] jw4: sure [14:32] jw4: looks good now [14:32] TheMue, ta [14:32] jw4: yw [15:02] is the environment UUID per .jenv file or is it per bootstrap? [15:02] if i create an environment, bootstrap, destroy, bootstrap again does it have the same UUID? [15:04] natefinch: standup? [15:06] perrito666: oops, lost track of time, thanks === ChanServ changed the topic of #juju-dev to: https://juju.ubuntu.com | On-call reviewer: see calendar | Open critical bugs: 1428692 [15:39] natefinch, dimitern (and ericsnow) I reported bug 1428692 about a vivid regression [15:39] Bug #1428692: cannot boostrap vivid: Operation failed: No such file or directory

[15:40] sinzui: yep, saw that :( [15:40] sinzui: BTW, sounds like the systemd cutover is scheduled for today [15:40] ericsnow, I did a dist-upgrade on the machine and cleared containers hoping the issue was a stale machine [15:40] sinzui: so that test would be failing either way [15:41] ericsnow, the test now runs with --debug to get more info too [15:41] sinzui: unfortunately the logs for the failed job aren't too informative [15:41] sinzui: --debug will help [15:41] sinzui: so thanks [15:45] ericsnow: wallyworld natefinch http://ftp.osuosl.org/pub/fosdem//2015/devroom-go/ [15:45] perrito666: cool [15:47] word of advice, your wife will scold you if you try to watch those on the big tv during dinner [15:51] dimitern, you made bug 1428345 high. Which milestone does it go to? 1.23-beta1 or 1.24? [15:51] Bug #1428345: unit-get private-address returns ipv6 address

[15:51] Bug #1428692 was filed: cannot boostrap vivid: Operation failed: No such file or directory

[15:52] Bug #1428692 changed: cannot boostrap vivid: Operation failed: No such file or directory

[15:52] Bug #1428692 was filed: cannot boostrap vivid: Operation failed: No such file or directory

[15:52] mup, bad boy [15:52] jw4: Roses are red, violets are blue, and I don't understand what you just said. [15:52] sinzui, hey, yes I did, but it's still getting investigated - so 1.24 I guess [15:52] Bug #1428692 changed: cannot boostrap vivid: Operation failed: No such file or directory

[15:52] Bug #1428692 was filed: cannot boostrap vivid: Operation failed: No such file or directory

[15:52] thank you dimitern [15:52] Bug #1428692 changed: cannot boostrap vivid: Operation failed: No such file or directory

[15:53] Bug #1428692 was filed: cannot boostrap vivid: Operation failed: No such file or directory

[15:53] mup is heading for an excessive flood quit message [15:53] Bug #1428692 changed: cannot boostrap vivid: Operation failed: No such file or directory

[15:53] Bug #1428692 was filed: cannot boostrap vivid: Operation failed: No such file or directory

[15:53] Bug #1428692 changed: cannot boostrap vivid: Operation failed: No such file or directory

[15:54] Bug #1428692 was filed: cannot boostrap vivid: Operation failed: No such file or directory

[15:54] Bug #1428692 changed: cannot boostrap vivid: Operation failed: No such file or directory

[15:54] Bug #1428692 was filed: cannot boostrap vivid: Operation failed: No such file or directory

[15:55] Bug #1428692 changed: cannot boostrap vivid: Operation failed: No such file or directory

[15:55] Bug #1428692 was filed: cannot boostrap vivid: Operation failed: No such file or directory

[15:55] dimitern: mup is heading for a kick in the head [15:55] I am not sure what is triggering those, when I see the bug there seems to be no changes [15:58] mup, shut up [15:58] dimitern: I really wish I understood what you're trying to do. [15:59] perrito666, it's seeing somebody (itself) mentioning bug # and goes [15:59] bug 1428692 [15:59] Bug #1428692: cannot boostrap vivid: Operation failed: No such file or directory

[15:59] mup is taunting me [15:59] lol [15:59] mup, sit [15:59] ericsnow: Unknown commands are unknown. [15:59] mup, stay [15:59] ericsnow: In-com-pre-hen-si-ble-ness. [16:00] dimitern: most bots have a feat not to listen to themselves [16:02] perrito666, yeah, but this one is written in erlang ;) [16:03] dimitern: I thought this thing was written in go [16:03] or I missed a joke there [16:03] perrito666, https://launchpad.net/mup [16:04] :) no joke - it's a dead serious language lol [16:05] * TheMue reads Erlang? [16:06] didn't know that it's written in a cool language [16:06] dimitern: https://github.com/go-mup/mup [16:07] TheMue: erlang is a fine language, if you are an ericsson telephone central [16:07] mup used to be written in erlang, niemeyer ported it to go [16:07] perrito666: or a couchdb or rabbitmq or ejabberd develoepr (and many more) [16:07] no, couchdb developers did not really know erlang [16:08] :D [16:09] maybe, never looked into the couch details [16:12] TheMue: I agree there are many things written in erlang, yet its not cool in all contexts :p [16:12] perrito666: would write a mobile game app with it, yes [16:13] perrito666: but for something like juju it would be fine [16:13] perrito666, dimitern, TheMue, katco, ericsnow: Sorry for the trouble.. Launchpad seems to be giving inconsistent results, likely as a side effect of caching+balancing [16:14] It's not listening to itself, though.. note the filed/changed/filed/changed/etc [16:14] niemeyer: mup simply has been lonely and wanted to talk to us [16:14] niemeyer: you took the trouble to write the thing, no need to apologize, we are just curious on what was the cause [16:14] niemeyer, ah, I thought something more sinister was happening :) [16:14] TheMue: i like that explanation :) [16:15] I will increase the delay between updates, to give time for Launchpad to update its internal cache [16:16] I see you are hitting the same change in different caches each time, that is an odd thing for launchpad to do === urulama is now known as urulama__ [16:55] sinzui, is it possible to change http://bazaar.launchpad.net/~juju-qa/juju-ci-tools/trunk/view/head:/jujupy.py#L284 so it calls subprocess.check_output(args, env=env) instead of check_call ? [16:57] dimitern, no because we already have get_juju_output() whcih does that. what callsite needs to look at the output? [16:58] sinzui, well, looking at that vivid failure it's hard to understand whether "Operation failed: No such file or directory" comes from juju bootstrap's output or something else [17:00] sinzui, for example, it could be coming from dump_env_logs in the except BaseException as e: block below [17:01] sinzui, no, actually not that block, but the previous try that ends with a raise [17:06] dimitern, I don't think your going to see more information by switching which method we call. We are seeing the stdout and stderr in the console log already. The output captured would be the same as we are seeing now [17:08] sinzui, I guess so [17:08] dimitern, I am going to run this my hand on the machine to see if I get something more [17:09] sinzui, thanks! [17:10] sinzui, while you're there can you check if it runs out of disk space (or mongo runs out of the prealloc block)? [17:15] sinzui, it just *smells* like a mongo issue - i've just noticed juju-mongodb in vivid is /usr/lib/juju/bin/mongod --version: "db version v2.4.10\nThu Mar 5 15:00:16.736 git version: nogitversion\n" [17:16] sinzui, while in the cloud-tools, in trusty and pretty much everywhere we run and test with officially it's 2.4.9-0ubuntu3 [17:17] sinzui: FYI, this bug doesn't seem to be fixed: https://bugs.launchpad.net/juju-core/+bug/1424069 [17:17] Bug #1424069: juju resolve doesn't recognize error state

[17:24] dimitern, sorry this is taking longer. I just verified other juju do work on the machine. Mongo doesn't seem to be a factor with 1.21 and 1.22 [17:28] sinzui, is it the same version? [17:28] dimitern, yes, there is only one mongod on the host [17:29] sinzui, and where is that one ("gitnorevision"?) coming from [17:29] dimitern, I don't know. [17:30] dimitern, we are testing a package we built else where on the vivid machine. the machine was dist-upgraded to get the current packages, of which juju-db will be selected [17:32] sinzui, ok, so this is the first time we're actually testing with that version of mongo? [17:34] dimitern, I think so. I see juju-mongodb at 2.4.10-0ubuntu1 [17:38] dimitern, this is the log from the manual run https://bugs.launchpad.net/juju-core/+bug/1428692/+attachment/4335369/+files/vivid-bootstrap.log [17:38] Bug #1428692: cannot boostrap vivid: Operation failed: No such file or directory

[17:38] sinzui, ok, looking - thanks! [17:40] dimitern, though the bootstrap failed, the mongod is still up. Is there anything you want me to do to interrogate it? [17:40] sinzui, I'd appreciate you can paste cloud-init-output.log and /var/log/syslog where mongo logs a lot === kadams54 is now known as kadams54-away [17:42] sinzui, or if there's nothing mongo related in syslog, then /var/log/mongodb/mongod.log if it exists [17:44] dimitern, this is the cloud-init-output.log https://pastebin.canonical.com/127005/ [17:44] ok, thanks again [17:45] sinzui, that's how it ends? well, we'll need to look at /var/log/cloud-init.log then just in case - it should say "cloud-init ver finished" etc. [17:46] dimitern, I don't see that. remember that this is a localhost [17:46] sinzui, ah, right [17:47] sinzui, so then it's not strange [17:48] sinzui, and you can confirm the machine is not out of disk space, right? === kadams54-away is now known as kadams54 [17:49] dimitern, 26G free [17:49] sinzui, ok, good [17:52] dimitern, there are no logs for the mongo in the juju home dir or in /var/log [17:53] sinzui, and you said it's still running? [17:53] sinzui, is it listening on 37017? [17:54] dimitern, I see the pid [17:59] dimitern, I think this is just a zombie left behind from a failed destroy. Juju removed itself from the system, but left the juju-db behind [18:00] sinzui, and there's nothing you say in /var/log/syslog or /var/log/* about mongo? [18:04] dimitern, There are lines in the syslog. I am collecting them [18:06] sinzui, good [18:06]