[00:12] bcsaller, any progress resolving that? [00:13] hazmat: I've been running into other issues on both EC2 and local trying to reproduce it [00:13] bcsaller, ? [00:13] what issues? [00:15] things like the oneiric instance isn't getting a new enough python-zookeeper so the managed classes were not there [00:16] on the local provider the UA wasn't coming up when the series was old [00:18] the json module seemed fine however which was all thats involved in that test afaics [00:26] bcsaller, juju-origin: ppa should solve the first (txzk version) [00:26] bcsaller, pls file a bug on the latter [00:33] SpamapS, trunk open? [00:33] are we there yet ;-) [01:21] hazmat: you got openstack to work at all yet ? i dont care if i got to use s3 or not, i just cant seem to get my env.y happy [01:21] its always something else once i get one fixed [01:21] jej [01:22] imbrandon, i haven't had a chance to get rolling with it. though i did just get my invite for the rackspace openstack beta squared away. i'm hoping to get back into it post my velocity talk tomorrow [01:24] rocking, if you dont mind ping me if i seem to be arond [01:25] i'll try to play at the same timeish that way we can bounce fixes.problems with less effor both ends [01:25] :) [01:26] least till successfull bootstrap or i;ll slit wrists what ever seems like it will work hahahah , totally joken before someone calls the loony bin on me [01:27] ok , gonna see if i cant migrate websitedevops.com to jekyll+bootstrap tonight ( havent started yet only been reading up on it ) hehe [01:27] afk-ish [02:15] hazmat: would you be cool if i changed the links on juju.ubuntu.com that point to the docs over to jujucharms only long enought till we get the RT ticket fixed or get the ok to charm it up [02:15] i know might mean $ so thus asking [02:16] before i did it, and i really just mean those two links in the wiki header nav [02:16] temporarly [02:17] hrm or actually we already point some pages to jujutools.github.com and multi of us herer on the core team have "owner" on that repo, i could just upload it to /docs there via git [02:18] and no cost to anyone and be on an semi official , or at least fully contoled via official ppl [02:18] * imbrandon preps that, not like it would take 30 sec to switch the link back [02:20] I'm a little confused.. I have a charm that's been deployed by someone else.. I added the environments.yaml I was given to my .juju directory, but I'm still not able to run juju status, or anything else.. Any ideas what I may be missing? [02:30] cjohnston: are you running it locally? [02:32] JoseeAntonioR: no, ec2 [02:33] cjohnston: verify all the data in the environments.yaml file is correct, then juju bootstrap, juju deploy [charmname] [02:33] its already deployed [02:33] and now I'm trying to connect to i [02:33] it [02:33] juju status should give you a public IP address for it [02:34] unless it hasn't finished downloading the master image and configuring yet [02:34] its been up for a few months... I get "juju environment not found: is the environment bootstrapped?" [02:34] I can ssh into one of the servers, I can view the site.. I just cant get to all the servers because I dont know all the ip's [02:35] hmm, maybe juju destroy-environment and juju boostrap again [02:35] noo... thats bad juju [02:35] :P [02:35] if there are things you can't lose, then don't run it [02:35] ya.. not running [02:36] you should try "juju bootstrap", if the environment is not bootstraped it will do it, but if it is already bootstrapped no changes will be applied [02:37] you won't lose anything at the end [02:38] should I have to run bootstrap in order to connect to something that was setup by someone else? [02:38] no, you should run bootstrap to bootstrap the environment [02:39] after the bootstrap, status [02:39] may work [02:43] I got asked about a fingerprint.. I guess thats a good thing [02:43] hrm.. its only showing me one agent and no services tho [02:43] cjohnston: then no charms were deployed [02:44] you're starting from scratch [02:45] then something is still broken.. I'm ssh'ed into one of the servers... I know it exists [02:47] and no juju+charms in that server, then [02:49] There should be 4 servers showing in status, a front end, a db, a memcache, and the server that the code lives on.. I'm only getting 1.. [02:50] that seems weird [02:51] tell me about it [02:57] cjohnston: I think it's only showing the server you're connected to [02:58] I sshed into the server it shows, and I don't recognize any of the processes running on the server [02:59] what the... [03:01] jml, ah, it does. [03:38] dang... cjohnston left [03:39] JoseeAntonioR: he's not using the right environment in his commands [03:39] he is in some of the same chans i am [03:39] m_3: that explains everything, I'll let him know [03:39] i.e., 'juju status -esummit' [03:40] yeah, his keys were injected as part of the 'authorized-keys:' for the environment [03:40] still online, so catch him if you want to [03:40] pm-ing... thanks [03:41] hazmat: otw... we leave an another hour or so [03:42] I'll be working from utc+10 for a couple of weeks [03:42] m_3: utc+10 !!! that's a change [03:43] m_3: nice , you have to put up with more of me then , hahaha [03:43] back [03:44] negronjl: yup... actually sitting in SFO atm :) [03:44] negronjl: taylor's got a conference in sydney [03:44] m_3: ahh ... enjoy [03:45] negronjl: yup... _crazylong_ flight first tho [03:47] how long m_3 [03:47] thats wh........ [03:48] cjohnston: til the 9th [03:48] flight.. how longs the flight [04:56] jcastro: still handing out tshirts like lollies? [05:27] heya lifeless you ever get to look into that box running the wiki/docs by chance ? [05:39] Can nodes be shut down from with juju as to power them down? Is that juju terminate? I've found my ssh key is only working on my first bootstrap node. All others did not get the ssh key and apparently did not bootstrap correctly for juju. What forces the subsequent nodes to re-bootstrap with juju and get the right ssh key? [05:40] Trying to avoid a complete re-do with destroy-environment [05:42] burnbrighter: you should be able to use your cloud providers API to get those additional nodes boot logs [05:42] this is maas [05:42] ah, not sure if it provides that [05:43] anyhow, Juju attempts to reprovision nodes that don't come up and register with the master node after a reasonable time [05:43] can I simply powerdown nodes without doing damage to the juju environment [05:43] ? [05:43] so you should be able to use MAAS to terminate and return to the pool the node. [05:43] I don't know whether the machine agent will run on reboot; if I did I could answer that question :) [05:43] I've had bad experiences with removing the node from maas without removing from juju [05:44] imbrandon: the sphinx docs q? I have the ticket to look at now, I haven't had time to look into the complexity and see about poking someone [05:44] ok, I will try just shutting it down. thanks [05:46] cool, yea me and SpamapS both receated and tested the proposed solutino of enabling -backports and upgrading the sphinx workd perfect, but yea i understand thir might be other services on the box to check [05:46] * imbrandon hopes not tho [05:46] heh [05:47] lifeless: let me know if there is anything i can do to make it easier on ya when you do get that far :if i'm around etc) [05:52] imbrandon: sure, thanks [06:38] Are there any suggestions on how to approach the ssh keyless maas nodes? Maybe diagnosing a possible problem in the provider code? [06:38] I can't get access to the nodes, so I can't see the logs === hspencer is now known as hspencer[afk] === almaisan-away is now known as al-maisan === ejat is now known as e-jat [12:49] imbrandon: wtf? Another channel? [12:49] imbrandon: #juju-dev would have been the place for them to go, IMO === e-jat is now known as ejat [13:37] aujuju: Security groups creation in EC2 [14:28] imbrandon, sure re docs, though do you know who confirmed the rt for the docs update [14:52] other than looking at $HOME/.juju/environments.yaml, is there a way to list your working/available environments? === zyga is now known as zyga-food [15:37] hazmat: ok, at this point, 544 is a failure. I'm releasing 543 as 0.5.1, and trunk is open for dev [15:37] jimbaker`: ^^ please fix natty/oneiric [15:40] wtf [15:40] why can't I call "galapagos" milestone "0.5.1" ? [15:41] ah because version was galapagos [15:41] hrm or not [15:41] weird [16:05] SpamapS, cool, fsync test stuff merged to trunk [16:06] where's the codebase that generates the review queue wiki page? === zyga-food is now known as zyga === salgado is now known as salgado-lunch === al-maisan is now known as almaisan-away [16:21] 'morning all [16:24] james_w, there's a copy in charm-tools used for the cli version [16:25] err, why did I say wiki? [16:25] bzr lp:charm-tools && less scripts/review-queue [16:25] thanks [16:25] hazmat, the web version uses that code too? [16:25] james_w, pretty much, it just takes the output and dumps it into a mongodb, toss a simple html template on it [16:26] it does the data collection as a cron job to avoid interacting with lp inline to the web request [16:26] hazmat, so the web page is live backed by mongo? [16:27] sort of. [16:27] james_w, yes nginx w caching, pyramid, and mongodb [16:27] hazmat, ok, thanks [16:27] hazmat, OOI why not just write out static html in the cron job? [16:28] james_w, i've been asking myself that same question recently about the entire jujucharms.com site.. ie render static and push to cloud front [16:28] the only dynamic bit atm is the search [16:28] yeah [16:28] which according to analytics isn't used that much [16:28] jujucharms I can understand more, as there are lots of pages [16:29] eventually static html becomes a poor tradeoff with that IME [16:29] hazmat, so the review-queue is done like that because the rest of the site is? [16:29] james_w, mostly its just ease of changing out styles and templates, and the caching makes it fairly fast regardless, even live (sans cache) its fairly fast [16:29] james_w, yup [16:29] ok, thanks [16:29] for example, status.ubuntu.com is static html [16:30] if you can do it statically it's generally better. [16:30] and takes >1hr to generate now, which means the original design decision isn't such a good idea any more [16:30] especially if it doesn't change often. [16:30] given the likely access pattern [16:30] well, part of the future goals is to add a bit more dynamicism [16:31] here's the bug list for the site.. fwiw. https://bugs.launchpad.net/charmworld [16:31] er.. feature requests [16:31] thanks [16:31] we're just wanting a review queue for a "distribution" on LP as well, but don't need a browser at this point [16:32] so mongo+nginx+pyramid seems a bit overkill [16:40] hm. haproxy charm hooks/stop stops mysql instead of haproxy. [16:44] well i guess i need to learn how to make patches and submit bugs now. [16:46] zirpu: bzr branch lp:charms/haproxy , edit, bzr commit, bzr push lp:~your-lp-username/charms/precise/haproxy/fix-stop-hook, bzr lp-propose [16:46] cool. thanks. was digging through the docs. === salgado-lunch is now known as salgado [16:56] zirpu: note that the stop hook isn't actually called anywhere yet, but it should be at some point in the near future. [17:00] ok. this was a good 1st for me. [17:01] zirpu: ^5 [17:01] i'll be going to the charm school session at velocity at 13:00 today (PDT). [17:18] zirpu: *awesome* [17:21] hi guys :-) [17:27] hah [17:27] notice [17:27] 10:21 ... join!#juju -> juju-velocitycon(~subway@67.111.115.235.ptr.us.xo.net) [17:27] subway :) [18:02] nice === zyga is now known as zyga-afk [18:41] negronjl, what's the admin web ui for hadoop? [18:41] er. where [18:41] hazmat: hadoop-master:50030 [18:41] hazmat: or .. http://:50030 [18:42] negronjl, thanks [18:42] hazmat: np [19:09] jcastro: ola [19:40] wth boto.. shipping your own certs and ignoring those in /etc/ssl/certs?! [19:40] badform [19:42] hazmat: something wicked in the fsync change me thinks [19:42] * Duration: 2 hours 50 minutes [19:42] test_resolved_stopped ... [OK] [19:42] hung there before the buildd killed it [20:00] * SpamapS bootstraps using the proposed openstack provider [20:02] agent-state: running [20:02] sahweeet [20:07] sweet! [20:07] anyone written a provider that uses vagrant? or virtualbox? [20:10] zirpu: no, just the local/lxc provider [20:11] zirpu: its certainly been wondered about a lot [20:13] Anyone know what this error is about: [20:13] ProviderInteractionError: Unexpected Error interacting with provider: 409 CONFLICT [20:13] 2012-06-27 20:19:07,255 provision:maas: juju.agents.provision INFO: Starting machine id:10 ... [20:13] 2012-06-27 20:19:07,471 provision:maas: juju.agents.provision ERROR: Cannot process machine 10 [20:13] I see lots of emails on it, but nothing that definitively says why I'm receiving this error [20:14] burnbrighter: you may have forgotten to set port in your environment.. or the constraints you have are not resolvable [20:17] Thanks - which configuration file handles this? [20:17] (this is for maas btw) [20:18] burnbrighter: yeah that only happens with MaaS .. it depends on what the problem is [20:18] burnbrighter: if its the port problem, its ~/.juju/environments.yaml [20:18] trying to run openstack and none of the nodes will start [20:18] burnbrighter: if it is a constraints problem, you can resolve it with juju set-constraints [20:19] you are right, I have no port set [20:19] I am using the example from the documentation [20:19] ah, no wait [20:19] burnbrighter: there's a fixed version of juju in precise-proposed I think [20:19] burnbrighter: it assumes 80 now ;) [20:19] yeah, that's there: [20:20] environments: [20:20] maas: [20:20] type: maas [20:20] maas-server: 'http://172.16.100.11:80/MAAS' [20:20] maas-oauth: 'XXXXXXX' [20:20] admin-secret: 'nothing' [20:20] default-series: precise [20:20] does that look right? or should I be using the latest version of juju? [20:22] burnbrighter: 0.5+bzr531 should be fine [20:24] burnbrighter: dpkg -l juju to figure out the version [20:27] so the environments.yaml looks ok? [20:27] the nodes did their bootstrapping and cloud-inits fine. i could ssh to them and everything [20:31] burnbrighter: right, so when do you get that 409 ? [20:31] just after I finished deploying all of the openstack stuff [20:31] and exposing it [20:32] you may want to run 'juju -v whatever-causes-it' and possibly read /var/log/juju/provisioning-agent.log on machine 0 to figure out why [20:32] burnbrighter: my guess is not all of your services were able to deploy because of constraint problems [20:33] but I am not a MaaS expert :p [20:33] * SpamapS heads to lunch [20:33] burnbrighter: try #maas too [20:33] thnx [20:36] hi guys from velocity live [20:36] -) [20:37] * zirpu . [20:39] svg status graph done via graphviz? [20:47] cs == charm store? [20:47] zirpu:yes and yes [21:48] Problem above identified - mass/juju REQUIRES 10 reachable nodes, I only had 6 [21:50] burnbrighter: there is a hack you can do to consolidate a few onto node 0 [22:08] burnbrighter: basically you can force mysql+rabbitmq+nova-cloud-controller+something-else onto node 0 [22:19] imbrandon: hey did you get the openstack provider to work w/ hpcloud? [22:19] signed up today.. want to see how it goes [22:31] yes [22:31] hah doh [22:31] ssh fail [22:37] * SpamapS goes on a mission to fix the mediawiki charm [23:02] SpamapS: no [23:17] SpamapS: you get it working on hp ? [23:17] imbrandon: have not tried yet [23:17] imbrandon: just using the internal canonistack so far [23:18] hrm, sneek some hp time in i;d love to have someoen else fail too so i can get mine going [23:18] but beating my head on it alone , i just gave up [23:23] jcastro: around ? [23:23] hi [23:23] jcastro: hey [23:23] jcastro: so, are you still handing out tshirts for charms ? [23:23] yeah [23:23] cool, I can has? [23:24] Yeah, mail me your shipping info [23:24] \o/ [23:24] You get a handy dandy ubuntu mug as well [23:24] https://code.launchpad.net/~lifeless/charms/precise/opentsdb/trunk btw [23:25] k, just be responsive when a review comes in [23:25] those mugs are nice, i use mine daily :) [23:25] the shirt is for when your charm is in the store itself [23:25] but it's ok, I have this amazing feeling that SpamapS will drive that queue to zero today [23:25] * jcastro snickers [23:26] i was gonna do some reviews tonight for doc's merges so i may get some charms done too [23:26] jcastro: hows velocity [23:26] pretty high end [23:27] watching how they built Pinterest [23:27] jcastro: i got yout email too, but for what i do i would have to move to ireland, no thanks [23:27] heh [23:27] duct tape and paper clips? [23:27] yea pintrest and instragram are both kick ass on the infra [23:27] * sidnei j/k [23:33] jcastro: tomorrow is my day isn't it? [23:33] jcastro: damn, I was planning on doing it tomorrow, silly me === salgado is now known as salgado-afk [23:53] i've decided i need to upgrade from my old lenovo t61 laptop to something new and better. anyone have suggestions? or laptops that you're using and like? [23:53] i've only got a measly 4gb memory. :-( [23:53] * zirpu really want's 1Tb just to feel snug. [23:54] zirpu: The system76 laptops are nice, especially the new ones, you no longer need to carry around a massive power brick [23:56] i have one on my desk at work. i stopped carrying it around. it's just a friking desktop. huge. it has 8g memory. :-) [23:59] hmm.. I wonder if charmers should have an official backports PPA