[00:42] marcoceppi: still around? [00:43] jose: ping [00:43] thumper: pong [00:43] jose: hey, you were having problems with the local provider? [00:43] thumper: yep, that was me [00:43] jose: I have good news and bad news [00:43] what is it? [00:44] well, when lxc is installed it sets up a bridge by default [00:44] normally this is lxcbr0 [00:44] but you can change it [00:44] now you can configure the bridge for the local provider to use [00:44] jose: can you pastebin an ifconfig? [00:44] sure thing, just give me a minute [00:44] jose: the bad news, I have found this morning that the local provider is broken in 1.16 (and trunk) [00:45] and just submitted fixes a few hours ago [00:45] yay! [00:45] but when will the next release be done? [00:45] well, I'm pushing for a 1.16.1 ASAP [00:45] so, maybe tomorrow even [00:46] we have hoops to jump through [00:46] http://paste.ubuntu.com/6248564/ [00:46] I may push for a 1.17 into the juju ppa [00:46] jose: hmm... no bridge there [00:46] yeah, I was about to say that [00:47] jose: here's mine http://paste.ubuntu.com/6248565/ [00:47] jose: apt-cache policy lxc ? [00:48] jose: also, what's in /etc/lxc/default.conf ? [00:48] http://paste.ubuntu.com/6248566/ [00:50] jose: I take it you are on raring [00:50] yep [00:50] I wonder why it didn't create the bridge for you... [00:51] bug! [00:51] hallyn: any idea why lxc wouldn't create the lxcbr0 bridge when installed? [00:52] I may be being hopeful that hallyn is online [00:52] jose: however there is another bit of bad news... [00:52] thumper: which is... [00:52] jose: I found out today that lxc isn't working in precise just now [00:52] due to another bug [00:52] thumper: jose: does your kernel have bridges built-in? what does 'brctl show' say? [00:52] this is either in progress or in a review queue [00:53] wasn't that on quantal? :P [00:53] hallyn: no bridges [00:53] jose: your kernel doesnt' support them? [00:54] http://paste.ubuntu.com/6248592/ is what I get [00:54] jose: when juju starts the precise ubuntu-cloud image in lxc, there is a package configuration problem that causes most charms to fail install hooks [00:54] jose: does 'brctl addbr br0' work? [00:55] hallyn: operation not permitted [00:55] sudo !!? [00:55] ah, right! [00:55] uh, !! in the history sense, not in the i'm shouting sense :) [00:55] done [00:55] :) [00:56] thumper: it explains why it failed when I demoed it at a conference months ago :P [00:56] a month ago it worked [00:56] although that is sub optimal [00:56] then, that wasn't it [00:56] hallyn: done, btw [00:57] jose: and that worked? [00:57] http://paste.ubuntu.com/6248605/ [00:57] is anything in /var/log/upstart/lxc-net.log ? [00:58] ah... I wondered where to look for logs [00:59] http://paste.ubuntu.com/6248609/ [00:59] hmm, /me checks the dhcp server, maybe? [01:00] jose: you use 10.0.3.0/24? [01:00] no, not at all [01:00] hm [01:00] that's very weird then [01:01] so you have dnsmasq itself installed... but lxc should have made it - oh, di dyou *just* install lxc? [01:01] even my dhcp for the occasional ad-hoc network I enable is 10.42.0.0/24 [01:01] maybe /etc/init.d/dnsmasq stop; /etc/init.d/dnsmasq start would fix it [01:01] hallyn: I think yes [01:01] jose: is there a /etc/dnsmasq.d/lxc file? there should be... [01:02] my guess is dnsmasq hasn't been restarted to read that [01:02] there is [01:02] yeah, let me restart it [01:02] "have you tried turning it off and on again" would have solved this then :) [01:03] haha [01:03] (then to 'stop lxc; start lxc' to have it try to restart lxc-net) [01:03] oh wait dnsmasq was not isntalled [01:03] well then what's pinning 10.0.3.x? [01:04] should I install dnsmasq? [01:04] nope [01:04] I have bind9 installed, though [01:04] can you do 'netstat -na' and look for 10.0.3' ? [01:05] found, using port 53 [01:06] bind9? [01:06] state is blank, though] [01:06] bind9 is a dns server [01:06] no, i mean that's the one holding it [01:06] ? just to make sure all is making sense :) [01:07] then, any ideas? [01:07] what do you mean by state is blank? [01:08] i guess a 'listen-on' statement in bind9 config. [01:09] on the table, it says proto udp, recv-q 0, send-q 0, local address 10.0.3.1:53, foreign address 0.0.0.0:*, state [here, this is completely blank' [01:09] waiting for packages to install so i can test (and see the stock config) [01:10] sure [01:10] jose: a stock bind9 install followed by stock lxc install doesn't prevent lxcbr0 from being created in saucy (for me) [01:10] how custom is your config? [01:10] erm, just enabled a couple domains [01:11] let me check, though [01:11] jose: can you try adding 'listen-on { your ip addr }' in /etc/bind/named.conf.options ? [01:11] hallyn: wait, there is no problem if I turn it off as I'm not using it right now [01:12] and won't for the rest of this session [01:12] oh, you can do listen-on { ! 10.0.3.1 }; [01:12] jose: ok, in the meantime can you open a bug against lxc? we should do this automatically probably [01:12] sure thing [01:12] (though bind9 configs tend to be custom so automating it may be scary) [01:13] thx - /me heads afk [01:13] thanks to you! [01:13] hallyn: wait, turning bind9 off will solve the prob? [01:14] looks like it did [01:17] yeah, long as htat doesn't mean you can't resolve anything locally :) [01:18] it's all going good :) [01:18] thumper: then, problem solved until the fix is pushed to repos, thanks! :) [01:18] :) [01:18] if you are a fanatic and want to run from trunk... [01:18] you might get it working [01:41] thumper: I am now [01:41] marcoceppi: hey, it was just about the local provider and jose's problem [01:41] marcoceppi: but we got it mostly sorted [01:41] thumper: ah, cool [01:41] marcoceppi: however we have a few problems [01:41] marcoceppi: the local provider is broken in 1.16 [01:42] marcoceppi: we need to get 1.16.1 out to fix that, sinzui is in the loop [01:42] thumper: that's what I've heard, any idea if you guys will be fixing and addressing that in 1.16.1 ? [01:42] thumper: awesome [01:42] marcoceppi: the juju issue is fixed, but there is another lxc/precise issue [01:42] which means that while juju might "work", you can't deploy anything [01:42] awesome [01:42] as the install hooks fail [01:43] there is another group looking at that [01:43] marcoceppi: you are there next week? [01:43] thumper: I will be, should be touching down Saturday evening local time [01:43] * thumper nods [01:43] I arrive Sunday just after lunch I think [01:44] 11:15am landing actually === freeflying is now known as freeflying_away [01:44] so after noon at the hotel certainly [01:46] thumper: cool [02:39] hey guys, any idea why while trying to deploy rais it tells me ERROR no settings found for "rack"? [02:42] jose: how are you deploying it? [02:43] marcoceppi: juju deploy rails --config rails.yaml [02:43] jose: what does rails.yaml look like? [02:43] * jose pastebins [02:44] http://paste.ubuntu.com/6248944/ [02:44] jose: that's not a properly formatted --config yaml file for what you're doing [02:45] how should it be, then? [02:45] jose: https://juju.ubuntu.com/docs/charms-config.html see "Configuring a service at deployment" [02:46] ah, got it, thanks [02:47] marcoceppi: if I don't specify a value it'll use default, correct? [02:47] jose: Correct [02:47] thanks! :) [02:47] using --config is completely optional to begin with [02:47] just to confirm, are we having a charm school this week? [02:54] jose: no [02:54] then the cal was right :) [02:54] We'll have a new schedule figured out after the sprint [02:54] ok! === freeflying_away is now known as freeflying === freeflying is now known as freeflying_away === CyberJacob|Away is now known as CyberJacob === freeflying_away is now known as freeflying [13:55] marcoceppi, hey [13:55] cache-relation fails with wordpress/memcached [13:55] jcastro: this has been a known issue for months [13:55] oh [13:55] is there a bug? [13:55] why yuo no fix memcached? [13:56] because plugin author + time [13:56] oh this is the plugin one! [13:56] I remember now [13:56] https://bugs.launchpad.net/charms/+source/wordpress/+bug/1170034 [13:56] <_mup_> Bug #1170034: integration with memcached broke [13:56] ok well wordpress keeps running [13:56] so I think I'll use it as an example [13:57] http://ec2-54-221-127-143.compute-1.amazonaws.com:7777/ [13:57] this charm is in the incoming queue [13:57] it's a nodejs pastebin, it runs hastebin.com [13:58] jcastro: link? [13:58] jcastro: I'll try to fix WP charm on the plane to san fran [13:58] no promises [13:58] nah [13:58] I am doing the demo today [13:58] https://jujucharms.com/~davidolf/precise/haste-server-0/ [13:58] jcastro: k [13:59] marcoceppi, free time is for the discourses [13:59] oic ;) [13:59] jcastro: I don't see haste in http://manage.jujucharms.com/tools/review-queue [14:00] huh, the guy mailed me and asked me what to do [14:00] I'll follow up [14:00] but that was like 2 days ago, I would have assumed he would have submitted it [14:00] unless it's for the contest maybe and he's waiting? [14:00] * marcoceppi checks bug list [14:00] marcoceppi, it's pretty slick [14:00] the software itself I mean [14:01] jcastro: cool [14:01] jcastro: time to replace paste.ubuntu.com ;) [14:01] jcastro: charmers was not subscribed https://bugs.launchpad.net/charms/+bug/920797 [14:01] yeah [14:01] <_mup_> Bug #920797: Charm needed: Haste server [14:01] I'll mail him [14:01] or I can just fix it [14:02] jcastro: already did it [14:02] this is the one jcsakket was mailing us about [14:02] ones like this I mean [14:02] cool [14:02] I'm going to have it if it's a bug in charms and there's a branch attached with no resolution, add it to the queue anyway [14:03] that way when people don't follow step 23434 of our amazing launchpad submission process we still catch them [14:03] jcastro: yeah, branch attached, fix committed, bug on charms? show in the queue [14:03] yup [14:04] ok I've got wp, mediawiki, discourse, and their db's, haste [14:04] anything else I can demo for the ubuntu on air today? [14:05] jcastro: oh, for the demo - not sure [14:05] rails? [14:05] is there a way for a charm to tell the gui "don't show green, I am still doing stuff" [14:05] like the discourse charm shows up green right away [14:05] but needs like another 5 minutes to finish stuff up [14:05] jcastro: no, because there's no way to provide feedback [14:05] but if you look at the gui it looks ready to go [14:06] * marcoceppi really wants to be able to have juju describe it's state [14:06] hazmat, is that something we can talk about in SFO? [14:06] pending, installing, configuring, starting, started, relating [14:06] for when those events are firing [14:08] * marcoceppi dreams on [14:08] mysql has passed 10k downloads! [14:09] marcoceppi: I have a jujucharms.com question: why does 'openstack' never show up in the "providers" list? [14:10] could be that I'm misunderstanding what the providers list does [14:10] ehw: Possibly, could you link to a page where you're seeing "providers"? [14:10] https://jujucharms.com/precise/mysql-27/#bws-code [14:11] marcoceppi: ^^ but really any charm that I look at usually doesn't say 'openstack' [14:11] ehw: Ah, that might be a bug. Those are showing the results of our testing environment [14:11] another example, https://jujucharms.com/precise/ceph-16/ [14:12] ehw: right, we test "openstack" against HP Cloud, so it might show as that. We're still working out the kinks for testing stuff, just because it doesn't say it explictitly doesn't mean it's not supported [14:13] marcoceppi: ok, was curious about that [14:13] marcoceppi: thanks [14:14] ehw: we'll be working over the next several weeks to make those results (both testing and supported providers) better === freeflying is now known as freeflying_away === freeflying_away is now known as freeflying [16:01] hallyn: ran apport, it collected info from bind9, but doesn't seem to appear on launchpad [16:01] oh, there we go [16:05] no, not in there yet [16:14] jcastro, it sounds good to bring up, the notion of steady state is effectively a poll/time period with no active [16:16] marcoceppi, I got asked that at the ceph day last week as the ceph charm does not advertize as tested on MAAS === freeflying is now known as freeflying_away [16:17] jamespage: none do at the moment, since we don't have testing yet for MAAS, I think we should just hide those results for now, they're really misleading [16:18] marcoceppi, yeah - agreed [16:18] the assumption was that it was not tested on MAAS so did not support it [16:20] I think that's part of charmworld, sinzui are you guys working on these pages? https://jujucharms.com/precise/ceph/ [16:21] marcoceppi, juju-gui is the pages, the data is from manage.jujucharms.com (charmworld) [16:21] sinzui: thanks! gary_poster thoughts on above ^ ? [16:21] jcsackett, are we deploying charmworld? [16:22] sinzui: not at the moment, but we can. that's via RT, yes? [16:23] jcsackett, I was just inquiring as it would be factor if a jujucharm.com page had an issue [16:24] marcoceppi: ? I'm confused. We don't run tests on maas so we can't say the charm is tested on maas. What's the goal? [16:24] sinzui: ah, yeah, to my knowledge there's nothing going on with production right now. [16:24] marcoceppi: the goal is that we've had charms that we KNOW fail on ec2 before and wanted to provide info to the user [16:24] rick_h_: it's misleading, since testing isn't strong, to say that "Providers" and not list providers because there are no tests [16:24] jcsackett, if we release today, charmers get askubuntu and github integration [16:25] rick_h_: had two reports today of this causing end user confusion [16:25] just passing along the info [16:25] marcoceppi: yea, it's been around 100 times. at one point I got only failures to show, but then others wanted all listed. [16:25] marcoceppi: yea, understand. It's been a long running point of fun discussion [16:26] rick_h_: cool, here's more data for your pocket then :) [16:26] sinzui: yes, as long as we're releasing to revno 421 on charmworld and use charmworld-79 tag for the charm. [16:27] rick_h_: maybe even just saying "Tested Providers" with a caveat to mention that you can attempt to deploy charms on any provider we only have definitive results for XYZ [16:27] wait, we don't need all the way through 421. [16:27] many things have landed. :-P [16:27] sinzui: through revno 415 for github. it looks like there have several bundle related landings as well--i don't know the QA status of all of thsoe. [16:28] marcoceppi: yea, you can toss feedback out to Juju-gui-peeps if you want. I'm not sure where to take it at this point. [16:29] rick_h_: cool, I'll take this discussion to their court [16:29] jcsackett: yea, we've got proof api stuff and things in progress that are not complete for release. [16:29] jcsackett, can you do a quick review of staging. The app.log will show any misadventures with new proof. We can look at some charms to verify we see icons warning. [16:30] sinzui: popping into staging now. [16:35] sinzui: app.log shows a few tracebacks for bad YAML as IngestErrors, and some no branch, but that's all logging normal for charms with errors, iirc. [16:36] sinzui: i'm not sure i follwed the second part; what charms and what about icons warning? [16:36] jcsackett, tabs about smoke [16:36] ? [16:36] sinzui: yup. [16:36] I wrote the branch owner. He replied that he will fix them when he has time [16:37] jcsackett: sinzui so in checking no reason to not go to 421 if you want. [16:37] jcsackett, We want to confirm that charmworld is showing the proof that we include. [16:40] jcsackett, i looked at a few charms and I see sensible proof warning [16:40] * sinzui thinks the underling of them is ugly and a bug [16:41] sinzui: ok. my lunch finished cooking a moment ago. right after food, i'll start the release process. [16:42] fab [16:42] thank you rick_h_ [17:27] for those following along, marcoceppi caught up with me in #juju-gui . His point made sense to me, and he filed https://bugs.launchpad.net/juju-gui/+bug/1241075 , with my thanks. [17:27] <_mup_> Bug #1241075: Wording of "Providers" confusing on charm information page [18:14] marcoceppi: ping [18:14] rick_h_: pong [18:14] marcoceppi: got http://paste.mitechie.com/show/1048/ as the new setup for you. You can just use the root error messages. They're prefixed by the bundle name. [18:15] marcoceppi: the details are nested and they're lists now so that you can support multiple errors per service (e.g. multiple config issues) [18:15] rick_h_: it doesn't indicate severity - or are these all always error? [18:15] marcoceppi: what I'd suggest for a first cut is that, if there are errors in here, that you provide a link at the least to the api call for a "for full details go to:" until some sort of verbosity can come up [18:15] marcoceppi: right now, all the things in that blueprint we're checking are errors [18:16] we can expand, but currently it's all we have use cases for [18:16] rick_h_: cool, I'm fine with that just need to know how to prefix these messages [18:16] thanks! This looks good [18:16] https://blueprints.launchpad.net/charm-tools/+spec/charm-bundle-support says they're all critical [18:16] so error == critical and if we get to a warnings then we'll add "warning_messages", "warnings" [18:17] (note that's a typo, it should be error_messages (with an s) [18:17] rick_h_: right, missed "error" in the error_messages key [18:17] marcoceppi: cool, I've got some tests to write around this new stuff for the config checking and will get it pushed up. I'll let you know when it's on staging and you can poke at it with fury [18:17] thanks for working on this, looks good! [18:17] marcoceppi: probably sometime tomorrow [19:01] hi guys, still trying to understand some basics, with juju-local, I should be able to run precise charms on a saucy "host" right? [19:01] assuming lxc [19:07] CaptainTacoSauce: yes, despite your host being saucy, all LXC containers launched will be precise (unless you specify otherwise) [19:07] CaptainTacoSauce: however, I think there's a bug in 1.16.0 that breaks the local provider, so if you're experiencing an issue, it might be related to that [19:07] marcoceppi: ah, indeed I am, but I've been playing with it for all of 5 minutes now, thanks [19:08] CaptainTacoSauce: let me find the bug for you [19:08] CaptainTacoSauce: https://bugs.launchpad.net/juju-core/+bug/1240709 [19:08] <_mup_> Bug #1240709: local provider fails to start

[19:12] ^ I was seeing that bug when I installed on my laptop yesterday [19:54] how does the juju-gui charm get the admin-secret from enviroments.yaml ? I would like to be able to use the same for my own apps [19:59] paulczar: you have to manually enter it last I tried [20:00] when I do juju deploy juju-gui it somehow gets the admin-secret out of my juju environment [20:00] but I'm not seeing how in the charm hooks [20:00] paulczar: hum, I'll ping gary_poster but also poke at the charm if he doesn't answer [20:01] thx [20:03] paul czar, it does not--it just remembers in browser for a single url in a single browser session. However, https://launchpad.net/juju-quickstart gets out the admin secret with Python (and eventually we have talked about doing what you describe) [20:04] (that is, starting you logged in to the GUI) [20:04] paulczar, sorry ^^^ [20:32] sinzui: release has finished, heartbeat is coming up good. [20:33] I just checked the review queue [20:33] sinzui: be a bit before askubuntu items show up in it. [20:34] jcsackett, I expect new things to show by tomorrow [20:34] sinzui: that would be accurate. [20:34] i have a card for the release in assessment, but have moved releasable cards to done. [20:34] s/releasable/released/ [20:42] jcsackett, about how long until questions show up? [20:42] we talking hours or ... ? [20:42] jcsackett, thanks for working on this btw, it will be awesome! [21:23] Just upgraded from 1.14.x to 1.16.0 and now getting "ERROR cannot start bootstrap instance: index file has no data for cloud" What needs changing? [21:45] jcastro: the job runs daily, so should be up by tomorrow morning. === CyberJacob is now known as CyberJacob|Away