=== almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away === TheMue_ is now known as TheMue === almaisan-away is now known as al-maisan [09:21] Hola juju crowd ! i do not understand whay a service public-address actually show a private instance address http://pastebin.com/j8YpJ6DH [09:21] any idea about this ? [09:25] grumble, machine froze just after i posted here. [09:26] if anyone reply, i did not see it. (and now my cloud is dead too :-) ) [10:07] morning all === al-maisan is now known as almaisan-away [10:41] well, back after the freeze, i can reproduce the same "problem" agan (service public-adress seems to be actually the private one) [10:41] http://pastebin.com/2i3jDM0m [10:42] anyone understanding why mysql public adress is associated with a name i cannot resolve ? And is there a way to workaround this so i can use juju debug-hooks [10:42] (and try to understand what mysql did not install correctly) === koolhead11 is now known as koolhead17 === almaisan-away is now known as al-maisan [12:54] hi fwereade_ -- yesterday you asked me to run 'juju -v deploy' and send you the traceback. i've done it but no traceback is generated. [12:54] Morning all [12:59] heya bac, niemeyer [13:00] bac, so you just get the error and no traceback? that's ...surprising :( [13:00] bac, and I've been unable to repro at all [13:00] bac, was your deployment initially set up with an earlier version of juju? [13:02] bac, because that *should* stop you from using constraints at all (and I can't *see* any holes in the logic there...) but what you've reported sounds like it may be consistent with that [13:04] bac, btw, I was wondering about the script you pastebinned... it seemed as though you were manually specifying cpu/mem from the m2.4xlarge specification; is there a particular reason you weren't just setting `instance-type=m2.4xlarge`? [13:30] fwereade_, i've dug into it a little more and found the problem was on our side. for testing purposes we've been using a wrapper that invokes /usr/bin/juju and it was not passing the params properly. so sorry to have drug you into this mess. [13:37] bac, np at all, glad it's sorted out :) [13:40] gents, I'm unaccountably exhausted, taking a short break === al-maisan is now known as almaisan-away [14:10] fwereade_: Don't kill yourself main.. we need your brain [14:11] s/main/man [14:22] any idea what the following error means: [14:22] juju.errors.ProviderInteractionError: Unexpected Error interacting with provider: 409 CONFLICT [14:24] hmm, ok, foud it [14:28] <_mup_> Bug #980855 was filed: Unhelpful error message when no nodes are available with maas provider < https://launchpad.net/bugs/980855 > [15:48] jcastro: so I guess I should ship these juju shirts somewhere today [15:50] LOL [15:50] :P robbiew are you the shipping guy [15:53] bkerensa: man...I'm the *everything* guy it often seems for our events [15:53] lol [15:53] robbiew: where is my juju travel mug :P [15:53] ^ [15:53] lol [15:54] heh [15:55] robbiew: :P I was serious... I didnt get my shirt or cup :P is it coming on the Tibet Express? [15:55] takes me a month when I order stuff from Canonical Store :P [15:55] bkerensa: that's *all* jcastro [15:56] ;p [16:03] bkerensa: yeah it's slow because I am shipping them all at once [16:04] bkerensa: I'm shipping them today it took a while to get addresses, etc. [16:04] jcastro: no worries I was just trying to give robbiew a hard time [16:04] oh, just ask him how openstack is going, that outta do it [16:04] LOL [16:04] [16:04] jcastro: is he on the same team as adam_g? [16:06] bkerensa: he's the server team manager [16:06] ahh [16:06] :D === jkyle_ is now known as jkyl === jkyl is now known as jkyle [16:59] flacoste, is that 409 Conflict specific to that error? [17:10] hazmat: i think that's the API response we gave whenever we aren't able to satisfy a constraints [17:10] so no nodes available [17:10] or the requested name is not available [17:29] heya shazzner [17:29] did any of your charms make it in the store yet? Last I checked you had a few pending [17:34] hey [17:35] jcastro: not yet, I think marco mentioned promulgating them but I hadn't heard anything yet [17:35] k [17:36] the gitolite and kusabax charms [17:36] thanks man :) [17:37] m_3: here's that charm that is in github: https://bugs.launchpad.net/charms/+bug/977552 [17:37] <_mup_> Bug #977552: Charm Needed: Terraria server < https://launchpad.net/bugs/977552 > === shazzner is now known as shazzner-brb [17:39] https://bugs.launchpad.net/charms/+bug/966484 [17:39] <_mup_> Bug #966484: New Charm: Kusaba X < https://launchpad.net/bugs/966484 > [17:39] https://bugs.launchpad.net/charms/+bug/964936 [17:39] <_mup_> Bug #964936: Drupal Charm: superchared Drupal charm with nginx,, apc, php-fpm, all setup to scale to the moon and be Best Practices. < https://launchpad.net/bugs/964936 > [17:39] https://bugs.launchpad.net/charms/+bug/906176 [17:39] <_mup_> Bug #906176: Gitolite charm < https://launchpad.net/bugs/906176 > [17:39] SpamapS: m_3: marcoceppi: negronjl: all of those are in the queue ^ [17:39] who wants to promulgate? [17:42] oh look, its friday friday not thursday its friday [17:43] all except one are in late stages of review, easy pickins' Friday? [17:44] SpamapS: hmm, you and m_3 are doing precise store aren't you? [17:44] should we just punt these to next week? [17:45] right now I'm doing "landing subordinates in the PPA and eventually precise" [17:45] the last few commits of juju added a dep on a newer txzookeeper [18:01] SpamapS: ok so these can wait, o biggie [18:01] negronjl: if you're looking for something to do, page up. :) [18:01] jcastro: I'm already looking at all three of them ... just give me a few [18:02] jcastro: It appears that I am having AWS issues :/ [18:03] <3 thanks dude [18:21] adam_g: you got a minute or two? [18:21] bkerensa: im here [18:22] adam_g: I have a question about a former employer of yours who is making a offering :P [18:22] can I PM? [18:22] uh ya :) [18:59] jcastro: It appears that I can't get any instances out of AWS from juju ( juju deploy doesn't work ) so, the reviews will have to wait until I get that sorted out. [18:59] jcastro: this was working a couple of days ago .... did I miss anything ? [18:59] bah if only we could get it on hp cloud [19:00] negronjl: I don't think so, precise? [19:00] are you on distro or ppa? [19:00] jcastro: precise ... ppa [19:00] jcastro: actually ... launching from precise/ppa into oneiric/ppa [19:01] could be juju problems [19:01] SpamapS: late changes breaking things ? [19:02] negronjl: quite possible [19:02] the PPA is about to break stuff again.. subordinates and some txzookeeper changes landing (all good stuff.. ) [19:02] SpamapS: thought so ... I'll just wait [19:02] jcastro: Reviews are on hold until juju issues are sorted out :/ [19:03] no worries, thanks for chipping in though! [19:03] SpamapS: nooooo [19:04] marcoceppi: time to lock down to a branch :) [19:04] SpamapS: yeah, going to have to [19:04] marcoceppi: tho distro won't change after this next upload [19:05] SpamapS: when's the next distro upload? [19:05] marcoceppi: as soon as the PPA builds and I test it... so probably monday [19:05] that will be the last precise upload :) [19:05] cool, we'll just switch to distro then [19:08] *awesome* thunderstorm in west LA right now [19:08] like.. BOOM BOOM BOOM .. we never get these [19:14] <_mup_> Bug #981086 was filed: typo in the subordinate-services.rst documentation file < https://launchpad.net/bugs/981086 > [19:22] SpamapS: sure its not gunshots, or aliens in the battle for LA ? [19:22] hehe [19:22] imbrandon: no I am used to hearing those things every few days [19:22] :) [19:22] Hey so people [19:22] it landed quietly [19:22] but subordinates is *out* [19:22] oh [19:22] precise and oneiric have it in the PPA [19:22] always on the friday [19:23] heh [19:23] Well it landed yesterday.. [19:23] but the PPA has been broken for a few days [19:23] "ok done with this time to relax, oh wait new feature I've wanted for six months just landed!" [19:23] thats cool though, cuz i just paid my aws bill so i got a nice clean slate [19:23] I am glad it is here though, awesome job guys. [19:23] for the weekend [19:23] one disclaimer.. I have no idea how to use it [19:24] SpamapS: hahah nice, i'm sure we'll all figure it out over the weekend [19:24] bcsaller: ^^ whats a good starting point for using subordinates? [19:24] me and jcastro and marcoceppi are all half sick but i'm sure as hell still gonna play [19:24] I think this is a game changer [19:25] Its going to make things like nodejs and symfony REALLY useful [19:25] SpamapS: start with the jeos build :) [19:25] because you can make your app a subordinate to them [19:25] imbrandon: meh.. I think thats a recipe for disaster eventually :) [19:25] likely [19:25] but many jeos ones will be ok [19:25] one for each bigger group [19:25] i think [19:26] SpamapS: its pretty simple (hopefully) the spec shows it and there is an example charm 'recorder' that is subordinate. It will only see the entity its subordinate to for things like relation list [19:26] I think we just need a --placement machine:4 flag to allow overriding the "unassigned" placement mode [19:26] SpamapS: and status works as expected (even dot) so its easy to see whats going on [19:27] bcsaller: so.. you're telling me.. this is actually up to date: https://juju.ubuntu.com/docs/subordinate-services.html .. .I don't believe you! [19:27] ;) [19:27] I do actually, thats awesome, thank you! [19:28] spoiled with docs as soon as a feature lands [19:28] back when ...... get off my lawn^Winternets [19:28] SpamapS: yeah, it should be :) [19:30] marcoceppi: btw i want to rename and clean that plugin up a little before its used anywhere but omg, just fyi , it will be good for the general wp one and maybe even public release but its a bit silly right now :) [19:30] SpamapS: I'd love to try that out [19:31] marcoceppi: the OMG case would be to have the wordpress charm with all the generic stuff, then 'omg-wp' just have the themes and stuff. [19:31] anyway, I need lunch.. then I'll start playing with it [19:31] kk [19:33] jcastro: http://www.youtube.com/watch?v=kfVsfOSbJY0 [19:35] * bcsaller grumbles at his wireless [19:36] SpamapS: Right, and then be a subordinate of an nginx-proxy charm, right? [19:38] * marcoceppi is excited [19:38] yea i'm wonder if nginx and nginx-proxy should be seperate [19:38] guess so [19:46] whoa [19:46] umm [19:46] hazmat: does awesome mean no need for providers [19:47] imbrandon, well it lessons the need for a native ostack provider at the cost of running a proxy [19:47] nice [19:48] i've just seen the anouncement not read it all the way [19:48] are we talking a set of programs proxy or a whole machine instance [19:48] imbrandon, its basically a proxy that turns ec2 api calls into openstack native calls [19:49] and the proxy would have to be setup to be accessible from both the client and the environment (the provisioning agent needs to be able to use it as well) [19:49] ahh [19:49] so a bit more to it [19:49] but still [19:49] i still think a native openstack provider has some value [19:50] yea, a lot less headache i think [19:51] i got it kinda workin btw but i'm not happy at all with my python skills [20:26] How do I reslove "Relation-errors" in juju? [20:26] juju resolved doesn't seem to work [20:28] marcoceppi: it should [20:28] http://paste.ubuntu.com/928475/ [20:29] SpamapS It failed for the first unit-add, so I upgraded charm with the fix, removed the extra unit, re-added it. It succeeded on the new unit but the old unit still has a relation error [20:29] It's also not executing subsequent additional peer relations [20:30] Scratch that last line [20:31] I'm not sure if that's 100% true or not [20:32] marcoceppi: there was at one time a problem with upgrade-charm when the state was error [20:33] marcoceppi, juju resolved -h [20:33] hazmat: thanks [20:33] marcoceppi, you have to pass the relation name, else it just tries to resolve the unit alone [20:34] SpamapS, that's still the default, but its possible to pass a --force flag now [20:34] sweet [20:39] the upgrade state check only applies to the unit's agent-state not its relations, --force will bypass that check, but also won't execute the upgrade-charm hook (else the latter would have to cope with arbitrary conditions). [20:42] Okay, so I resolved the relation, but when the hook re-runs it's not getting the relation data anymore [20:53] marcoceppi, what do you mean? [20:53] marcoceppi, it can't access data of itself or a related unit? [20:53] hazmat: when the relation re-ran it didn't seem to actually receive the data from the peer relation [20:54] marcoceppi, can you pastebin the relevant parts of the debug or unit log? [20:54] well, it looks like the relation re-runs when you execute juju resolved, when it did it produced the files it was supposed, only it was missing all the data exposed in the peer releation [20:55] hazmat: shoot, I just destroyed the environment. Not feeling 100% today and about to head home [20:55] marcoceppi, bummer. hope you feel better, feel free to ping on it whenever [20:55] ^me [20:55] hazmat: I have this from the charm.log on the unit that failed: [20:56] actually [20:56] wait, no it chocked on a different hook [20:56] nevermind [21:02] SpamapS: your on mac hardware right ? [21:03] ever seen this problem ? http://f.cl.ly/items/2N2k141d3G3Z193I2K0T/IMG_0084.JPG [21:03] yes [21:03] btw the normal image seems to work, the made for mac one wouldent even boot [21:04] imbrandon: no, but I'm on a MBP 5,1 w/ nvidia [21:04] or MBA 4,1, which is intel [21:04] hrm , this has ati [21:04] yeah I got no ATI [21:04] but the green monitor is actually usb [21:04] I don't think I'll ever buy anything but intel graphics ever again [21:04] its a dvi --> usb [21:04] so tired of proprietary drivers [21:04] hehe :) [21:05] ugh cmd t not new tab [21:05] gonna have to retrain my hands [21:06] yea the two hooked to the ati card are fine, the one hooked to usb is green [21:06] :( [21:06] anyhow install finished, let me reboot and i'll see if i cant get it fixed up, least got a native 12.04 install now [21:07] so far thats the only issue, abeit a big one [21:08] odd just had an ec2 instance stuck in pending for 30m. [21:08] * hazmat watches the cloud burp [21:09] ah.. ec2 status page notes increased errors and latencies [21:13] hazmat: -> us-west-2 [21:13] hazmat: I've been using it a lot lately. [21:13] SpamapS, yup just switched out to it [21:13] i wish ec2-describe-regions actually had the location name as well, would make it a bit easier [21:13] hazmat: I've also noticed that us-east-1 is throwing more and more instances into random AZ's where it used to mostly put them all in a [21:15] this destroy env waiting this is totally bogus [21:15] on ec2, it never works [21:15] we should zero out the groups perms lazily [21:15] instead [21:24] totally needs a '-y' or '-f' too imo... I've got to `yes | juju destroy-environment -e...` [21:28] m_3: IMO I think if you supply a -e with destroy-environment it shouldn't prompt you yes or no, only when you run it with the default environment [21:30] marcoceppi: I'm fine with a '-y' yes, I really wanna [21:30] but something automatable [21:41] * SpamapS updates juju packaging branch to require txzookeeper >= 0.9.5~ [22:03] dmesg [22:03] ugh [22:07] vi /usr/share/X11/xorg.conf.d/52-displaylink.conf [22:08] jesus this hotkeys are gonna kill me [22:11] nice, I just upgraded my machine 0 from 519 -> 529 w/o incident [22:19] heh i get the 3rd display working and the other 2 stop [22:19] ok time to call in the big dawgs and find someone in #ubuntu-x [22:26] and yup, my luck, no one arround on a friday at 5 [22:26] :) [22:40] anybody else playing with subordinate charms yet? [22:43] SpamapS: not yet.. how's it going? [22:44] interestingly [22:46] * m_3 is just left hanging... [22:46] well its not clear to me how arbitrary things can be related to other arbitrary things, as I thought they would [22:47] but perhaps I should RTFM [22:49] I get 'No matching endpoints' [22:49] which contradicts the docs [22:51] ah ok [22:51] the docs aren't clear that you have to use interface: juju-info for that [22:54] doh, s3 fail [22:55] east-coast primetime... yay [23:02] SpamapS: I plan on playing with them in a few hours [23:03] SpamapS: can't wait to experiment [23:08] don't try to do munin [23:08] nearly done with that [23:08] :) [23:10] jesus frakin christ, i feel like its 1999 and i'm having to edit an xorg file and cant get the damn thing right , and to top it off recovery console dident work because after it loads the console font it scrambles the screen [23:10] fml [23:11] imbrandon: I know the feeling. proprietary drivers are the suck [23:11] there is a reason i use a desktop that "just works" its so my bloodbreasure isnt this high [23:11] SpamapS: i have all free drivers [23:11] SpamapS: displaylink drivers are open too, as well as the ati ones i have loaded, they just dont work togather [23:12] ahh [23:12] so its just X fail [23:12] ugh i'm just gonna use 2 screens for now [23:12] yea x fail [23:13] i have like 3 of these adapters, they are run of the mill $30 at amazon, i have like 2 laying extra on my desk, maybe i can bribe some x hacker to make it "just work": with one [23:13] i'll blog later about it [23:14] pisses me off though, simple shit like this should work, i know its a common setup [23:14] i see it daily [23:15] its almost as bad as the big open whitespace on g+ now lol [23:15] anyone else notice that ? [23:41] http://ec2-23-22-28-42.compute-1.amazonaws.com/munin/ [23:41] woot [23:41] nice [23:44] lp:~clint-fewbar/charms/oneiric/munin-node/trunk [23:46] feels a bit weird.. the service name on the other end is munin-node, not the real service name.. [23:46] so interfaces will probably have to push stuff like that through relation channels [23:51] SpamapS: cool man [23:54] ok, so , it seems to work. time for copious amounts of alcohol and sushi [23:55] lol [23:56] I have to say.. I looked at doing munin-node with existing chef/puppet solutions for it. [23:56] they were all twice as complex [23:56] the chef cookbook is basically the same.. [23:57] but has all the extra logic for installing on every platform under the sun [23:57] ok... gone [23:58] thats not nessesarily bad if done right, juju will have to grow multi platform wings someday to survive imho [23:59] should be abstratced at the juju engine api though, not at the charm/cookbook