[00:05] woah.. that's awesome.. [00:05] job requirement -> An infrastructure management nut: Chef or Juju preferred. [00:05] hazmat: nice [00:08] we've arrived [00:55] http://careers.stackoverflow.com/jobs/15084/devops-infrastructure-engineer-surveymonkey has the detailed requirement. very cool indeed [01:01] and apparently one other: http://www.dovajobs.com/job-details.html?jobid=4405192&jobtitle=devops+engineer+-+mountain+view%2c+ca&joblocation=mountain+view&jobcountryid=222 [01:02] ironically, juju.com is a job search engine; doubly ironically, one of the local python users works for them (remotely) [01:08] jimbaker: haha [01:08] jimbaker: I used to work for a job board company. Its a weird space.. so dominated by monster and linkedin at this point. :p [01:31] * hazmat picks up a book on ITIL [04:35] <_mup_> juju/ssh-known_hosts r441 committed by jim.baker@canonical.com [04:35] <_mup_> Properly handles key injection [10:22] niemeyer, sorry, I never followed up on the start hook guarantee discussion [10:22] niemeyer, you had more thoughts? [10:23] fwereade: I have some, and it'd be nice to catch up about the subject at some point [10:24] fwereade: Right now is not a great time for me, but maybe later today? [10:24] niemeyer, sure, just ping me when it works for you [10:29] fwereade: Sounds great, thanks! [10:43] niemeyer, offhand: when will we find out dates of UDS in April? [10:43] niemeyer, (or who should I be asking who might know ;)) [10:44] fwereade: Marianna might be the best bet [10:44] fwereade: But the page is generally updated as soon as she knows it, IIRC [10:45] niemeyer, cool, thanks [11:00] "We are excited to announce our new South America (Sao Paulo) Region." [11:00] Brilliant! [11:01] niemeyer, finally fast AWS for you? :) [11:07] therve: Finally :) [11:59] * hazmat yawnsw [11:59] another day, another ec2 region [12:00] hmm.. the sao paulo region appears to be the most expensive of the bunch [12:59] hi all. After deploying a charm, all was ok. But after rebooting the nodes, juju status say that 'state: down' for all services, and 'state: not-started ' for bootstrap. But all services are working fine ... Does anybody know how to say juju that everything is ok ? [14:15] why 'juju status' bring my node to state 'down' after a reboot ? Everything seems to work well ! [14:16] oarcher_: Which provider are you using? Local? [14:17] yes, i'm using local [14:18] oarcher_: Ok.. it still doesn't work over reboots [14:21] Is it a bug report that i can follow ? [14:44] oarcher_, its a known problem with work addressing already in progress [15:23] hey guys o/ I'm writing a charm for a service that generates a certificate on the server side, I'd like to know if there's any best practices in order to transfer the certificate to the client charm (needed for bilateral auth) [15:23] any charm I can have a look that would give me an example of that? [15:31] lynx man, bet… installed your port install of juju yesterday and got it working. However, juju did not work for local provider because of apt issues [15:32] mchenetz: yeah I couldn't include apt in my build since OSX was lacking it :) [15:32] hehe [15:32] mchenetz: happy to hear all the rest worked though, I need to find a way around that [15:32] lynxman: i am working on a solution with virtual box as a provider [15:33] mchenetz: oooh schweet [15:33] i will probably create a brew repo after that [15:33] mchenetz: let me know when you do so I can give it a ride :) [15:33] most defintiely [15:36] lynxman, via relation data [15:36] hazmat: any example I can have a look at? :) [15:37] lynxman, no.. but you'd just encode the cert you want to transfer and set it on the relation... via relation-set [15:37] hazmat: cool! [15:37] lynxman, one qualification to doing that its available to all related units of the remote service [15:37] hazmat: yeah it should be intended to work like that in this case so it's more than fine [15:37] er.. all units of the related service [15:37] cool [15:46] is anyone else having problems with launchpad? [15:46] i'm consistently seeing this message: http://paste.ubuntu.com/771240/ [15:47] hazmat, fwereade: ^ [15:48] rog, I've seen it once or twice in the past, it usually goes away again quite quickly [15:48] fwereade: i hope so - i've mucked up lp:juju/go and i need to restore it before anyone notices :-) [15:48] rog, and it's happening for me too ATM [15:49] good. i thought it might be something i'd done locally [15:49] fwereade: thanks [15:49] rog, don't worry, probably nobody else could see it... until you confessed :p [15:49] fwereade: they might have noticed that it had lost the last 2 revisions... [15:56] "Warning: criss-cross merge encountered. See bzr help criss-cross." [15:56] yay [15:59] rog, that's *usually* less painful than it sounds, were there any actual conflicts? [16:14] fwereade: actually, no. i reckon they shouldn't write that message if there are no conflicts... [16:15] rog, I dunno, I think it's quite handy to be notified when you're merging weirdly, but the message could probably use some work [16:27] fwereade: o/ ahoy there! [16:27] rog: ^^ [16:27] SpamapS: hi [16:29] heya SpamapS [16:55] nijaba: hey, I need to use the ch_peer_* stuff in a charm I'm working on ... are you going to have time to update with the stuff I mentioned in my review? [16:56] nijaba, btw that stuff is awesome (ch peer*) [16:56] glad its there [16:56] Yeah, its going to make peer relations much easier to read. :) [16:56] and write [16:57] hazmat: hopefully ch_peer_leader will be able to just call 'relation-leader' at some point. ;) [16:59] anyone seen this error while running unit tests.. OpenSSL.SSL.Error: [('system library', 'fopen', 'No such file or directory'), ('BIO routines', 'FILE_CTRL', 'system lib'), ('SSL routines', 'SSL_CTX_use_certificate_file', 'system lib')] [16:59] hazmat: where would we even be using OpenSSL.SSL ? [17:03] SpamapS, it gets invoked as part of the digest auth impl cert checking by twisted [17:06] what cert checking do we do? :p [17:06] meeting? [17:06] rog, invites out [17:07] SpamapS, niemeyer, m_3, jcastro, jimbaker, bcsaller, TheMue, fwereade.. its that time of the week [17:07] hazmat, indeed [17:08] hazmat: i see no invite [17:08] Yep, no invite yet [17:09] hmm [17:09] hazmat: joining shortly [17:09] TheMue, rog resent to you individually [17:09] i tried using a circle this time around, but it seems a little flakey [17:09] hazmat: got it now [17:10] * SpamapS adds sa-east-1 to the regions allowed [17:11] be on late... might miss this one [17:13] Nothing here either.. and I [17:13] have a meeting in ~17 mins [17:13] Ah, got the invite now.. joining [17:14] Interesting that it didn't send a notification [17:14] * robbiew can't attend...but also knows he's the least important in the "room"...and has NO problem with that reality :P [17:16] It's really bad today for some reason [17:16] I can't hear a single thing :( [17:16] Now it says "The server connection has been lost" [17:23] hazmat: Am I the only one having issues? [17:27] niemeyer, yes [17:27] niemeyer, let me try inviting again.. not sure what else to do [17:28] hazmat: The invite isn't the issue.. I can see it [17:28] hazmat: It's just so laggy that I can only hear noise [17:33] niemeyer, i'm taking notes i'll send out [17:35] niemeyer, sometimes killing all flash on the system i've found helpful [17:39] * SpamapS runs test suite for 4th time a/ sa-east-1 added [17:45] Doesn't really work.. and Chrome seems to be leaving processes behind when I close it [17:45] Will try reinstalling the beta [17:52] Downgrading didn't help either [17:52] I'm giving up this time around.. [18:29] niemeyer, i had that problem once i had to shutdown both browsers, and hand kill the processes, to clear out all the flash players [18:30] er. chrome /firefox [18:30] twas strange [18:30] m_3, is that github mirror stuff automated? [18:31] hazmat: Yeah :/ [18:33] hazmat: it's partially so atm... should be fully automated by tonight [18:34] hazmat: and I'd rather trigger from pushes, but it's cron atm [18:35] putting tests and notifications in place b/c git-bzr-ng is... um... problematic at times :) [19:28] m_3: ping...you see my email re:objectives? :) [19:28] hm, test suite failing on trunk [19:29] http://paste.ubuntu.com/771484/ [19:29] hazmat: any ideas? ^ [19:30] SpamapS, you need to clean your checkout of pyc files [19:31] SpamapS, so in particular, juju.providers.ec2.tests.test_connect is gone [19:31] AHH [19:31] hmm [19:31] indeed [19:32] so, I think we need to fix bug #893176 [19:32] <_mup_> Bug #893176: do not limit ec2 region to static list of regions < https://launchpad.net/bugs/893176 > [19:32] seems like Amazon brings on a new region every 6 - 8 weeks [19:33] What about just changing it to a free-form string [19:34] btw do you guys have a good way to speed up the test suite? ;) [19:34] SpamapS, not fast enough for you? [19:35] takes 10 minutes [19:35] maybe we could use zk's chroot functionality [19:35] so we could run chunks in parallel don't know [19:36] 10 min is still quite managable.. just wondering if there are any tricks. :) [19:36] SpamapS, the other question is, are you running w/ a ssd? [19:37] because my time is much better. too slow, i'd like to run it on every change. but fast enough that i'm not certain what it is right now :) [19:37] jimbaker`: well the .py and .pyc's are all in cache, and the ZK is on tmpfs.. so it shouldn't matter really [19:38] hm actually /tmp is just / ... so not tmpfs. Hrm. [19:39] SpamapS, it takes under 5 with an ssd [19:40] SpamapS, you can do partial runs by passing in either a package or filename to ./test [19:40] robbiew: yup, sure did [19:40] * m_3 reading them now [19:40] m_3: lol...okay [19:41] hazmat: at this point, I always run the whole thing. :) [19:41] SpamapS, i agree with that sentiment [19:41] hazmat: I do partial runs while developing, but always the full run before commit. :) [19:41] and now, the full run, in a clean oneiric chroot [19:41] ok, i just reran test on my laptop - 2m50s [19:41] nice [19:42] jimbaker`, out of curiosity which ssd do you have [19:42] robbiew: sounds perfect... thanks! [19:43] m_3: awesome...3 down...too many more to go! :P [19:43] hazmat, it's the vertex whatever you recommended. i have two, one in my desktop, one in my laptop [19:43] * m_3 has objectives now! [19:43] hrm, why would SSD matter that much? Just for ZK? [19:43] thanks for the recommendation! [19:43] we should clean up the status tests, their dogs, they setup entire universes to test a little piece, that will shave at least a 1.5 m off [19:43] SpamapS, yeah.. primarily.. you could also try a tmpfs mount [19:43] SpamapS, everything in ZK is logged persistently [19:43] m_3: If one of them isn't "make clint a sandwich" ... you may want to revise.. ;) [19:43] SpamapS: ha [19:44] * m_3 revises objectives [19:44] hazmat: I always thought /tmp was tmpfs.. but it isn't.. remounting it soon :) [19:45] SpamapS, 38s in.. http://www.youtube.com/watch?v=H7PJ1oeEyGg [19:45] actually 35s in [19:46] his whole speil is get an ssd [19:46] gotta run to airport, bbiab [20:03] so.. weird.. when hazmat pushed to juju trunk the other day, _mup_ noted it. But I just pushed, and no note from _mup_ [20:04] is _mup_ only watching by *user* ? [20:04] SpamapS, it's a client side bzr config afaict [20:04] as in, they're somehow pinging mup? [20:05] yeah, it's a postcommit hook [20:05] that seems rather silly [20:06] I suppose its more efficient than the bot polling all the branches.. but really.. hrm. [20:22] SpamapS: there was talk of moving mup over to juju-dev... perhaps that's halfway done? [20:25] I think the thought was to just have it *also* in juju-dev [20:52] SpamapS, its a per user client side thing [20:52] SpamapS, i think a bot polling would be more useful [20:52] that sounds like a nice funtime project [21:01] hazmat: another thing to consider is to just have WTF do it (jenkins has an awesome irc bot.. ) === whit_ is now known as whit [21:38] Woohay.. new lbox built [21:43] what's an lbox? lxc? [21:44] zirpu: Nah, just a tool some of us use for development [21:44] zirpu: It's not so exciting really.. it's exciting because I can now do something else ;)