[00:00] NTP time says pencils UP [00:00] negronjl: https://code.launchpad.net/~komputes/charms/precise/remote-assistance/trunk [00:00] * SpamapS urges all to bzr commit / push [00:00] OK OK [00:01] pushed at 5pm on the dot [00:01] :D [00:01] nice [00:01] how to i translate this into postgresql add unit [00:01] http://paste.ubuntu.com/978954/ [00:01] SpamapS: It's UDS .... keep working :) [00:01] negronjl: :) [00:02] oo late entries include zentyal [00:02] :D I didnt finish Openphoto or Reddit stack charms but I did get Locker and Nyancat Termianl pushed :) [00:02] SpamapS: how to insert / create db after lunching the db instance [00:02] bkerensa: sweet [00:02] bkerensa: c00l [00:03] ejat: which db? [00:03] stressful [00:03] postgresql [00:03] +1 marcoceppi [00:03] negronjl: https://bugs.launchpad.net/charms/+bug/997386 [00:03] <_mup_> Bug #997386: remote-assistance charm created < https://launchpad.net/bugs/997386 > [00:03] ejat: the pgsql interface creates a database for you [00:03] ejat: and gives you back hostname/database name/password/user to use to access it [00:03] * ejat can we request to extend for another day ... thats make marcoceppi release his stress [00:03] hah, no more stress [00:03] SpamapS: when lp complains about a package not there ( ie: /charms/package_name/trunk ) how do you work around that ? [00:04] all the stress is on us judges now [00:04] SpamapS: if there is createdb.sql for example ? [00:04] negronjl: /charms/series/packagename/trunk [00:04] SpamapS: error: packagename is not a valid package name [00:04] ejat: ??? how can we know what the schema should be [00:04] SpamapS: got it [00:04] SpamapS: thx [00:05] SpamapS: its bundle with the apps .. for sample data [00:05] and schema [00:05] ejat: so do it the way upstream says to do it? [00:05] SpamapS: https://bugs.launchpad.net/charms/+bug/997339 [00:05] <_mup_> Bug #997339: Charm needed : Openbravo < https://launchpad.net/bugs/997339 > [00:06] now .. the charm will install its package .. bundle all together [00:06] * ejat just thinking how to make its scale by separating the db to another unit [00:06] ejat: just check to see if the tables are already created [00:07] for example .. i juju deploy openbravo3 [00:07] then i juju deploy postgresql [00:07] if tables_do_not_exist; then create_tables ; else juju-log "tables already exist" ; fi [00:07] juju add-relation openbravo3 postgresql [00:08] put in in add-relation ? [00:08] ejat: you can't really create the database until you have a relationship with the db server, can you? [00:08] ejat: so yes, you would create the database during the db-relation-changed hook [00:09] * SpamapS goes afk for a few [00:12] SpamapS: dont know ... i guess cant create db until there is relationship rite ? [00:29] marcoceppi: 2012-05-09 17:28:31,415 unit:gitolite/0: hook.output ERROR: invalid user: `gitolite:gitolite' [00:30] marcoceppi: oops, disregard.. different gitolite ;) [00:30] stupid revision file! [01:43] Hey Marco, these charms can't be accessed from precise, but the are listed in the charm store [01:44] keystone, nova-cloud-controller, nova-volume, glance, openstack-dashboard [02:00] royfeldman: they're a little bit wonky.. [02:00] royfeldman: we have a TODO to fix those [02:01] royfeldman: basically, use cs:~charmers/precise/keystone [02:08] OK [02:09] I will give it a shot now, right now I am using the charms in a local repo [02:10] Yup, that works. Thanks for the tip. [02:12] Anybody know about the charm github mirror? [02:13] Is say is a mirror updated nightly, but it is missing many of the charms in the store. [04:21] Hi! [04:22] When using juju add-relation [04:22] is there a way to be sure that one part of the relation will be done before the other part [04:22] by instance [04:22] juju add-relation foo-master foo-slave:slave [04:23] I want to be sure that the master-relation-joined to be executed before the slave-relation-joined [04:23] how can I do that ? [07:21] <_mup_> juju/scale-test r539 committed by kapil.thangavelu@canonical.com [07:21] <_mup_> switch topology from yaml to json [07:27] huats, you don't [07:28] huats, just do it in rel-changed, and if the master hasn't set some value, then exit in the slave hook side === almaisan-away is now known as al-maisan [09:12] i'm getting an unbound variable error when trying to use $JUJU_SERVICE environment variable in my relation-joined hook... am i misreading this? === al-maisan is now known as almaisan-away [10:22] morning :) ... So I've been using zc.buildout to do isolated Django website deployment. Looking at Juju, it seems that this can greatly improve my deployment process. I have a couple of Questions tho. Is there a way to test the charm locally, or do I need a EC account? [10:56] andrewebdev: http://askubuntu.com/questions/65359/how-do-i-configure-juju-for-local-usage [10:56] yes :) [11:09] thanks nathwill! that's quite comprehensive [11:09] thank the guy who wrote the answer ;) i just bookmarked it [12:30] So I followed the instructions here: http://askubuntu.com/questions/65359/how-do-i-configure-juju-for-local-usage and everything worked so far. The status command does not show the same however. [12:31] I dont have the same machine ips, in fact I just have [] [12:31] and state is just: pending [14:41] are there any docs on the apt proxy that is set up automatically for you by juju when using lxc containers? === al-maisan is now known as almaisan-away [14:59] Hi, just playing with juju for the first time. Try to experiment on a local (LXC) setup. All worked fine until I rebooted the host machine. Now, it appears that zookeeper has not been restarted. So everything I try times out. Any clue how I can restart zookeeper? [15:09] hippysurfer: juju destroy-environment; juju bootstrap [15:09] hippysurfer: don't know if there is a way to keep it up between boots, but that will fix it. [15:11] thank you, that has at least got me back to a sane state. I little harsh to have to destroy the whole environment, but I suppose the nodes are meant to be desposable. [15:13] hippysurfer: well, that is how I was looking at it. but I'm new as well to juju, maybe someone more experienced will have some magic to keep it useable between boots. :) [15:40] jcastro, I gots me a question about juju- are the dependencies included? [15:54] Who won the charm competition? I didn't get to finish mine in time for submission but figured it was sub-par and needed more work anyways :) [16:09] Or anyone else that may know about dependencies [16:10] * MarkDude is trying to see about getting juju in Fedora reops [16:10] repos that is [16:13] is there a trick to getting the lxc unit tests to run? [16:14] I'm running: TEST_SUDO=1 ./test juju/lib/lxc/tests/ [16:14] the work, but the first one fails (test_container_clone). then it seems to hang after that [16:29] jcastro, http://askubuntu.com/questions/134237/error-charm-not-found-in-local-repository [16:51] Still getting this issue whenever I reboot and try to deploy a local charm: "2012-05-10 10:50:54,730 ERROR Unable to create file storage for environment" [16:52] then I do juju destroy-environment - which asks for my sudo password and then posts the same error [16:52] Then - if I reboot and juju bootstrap - it "just works" TM === vednis is now known as mars [17:20] SpamapS, nice README for the Puppet charm. Very readable and concise. [18:18] hazmat: thansk thatsolved my isssue === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away [19:29] has anyone had success deploying a large local charm? Something like 100mb? It doesn't seem to transmit to ec2 properly [22:32] <_mup_> juju/scale-test r540 committed by kapil.thangavelu@canonical.com [22:32] <_mup_> extra sanity check for provisioning agent [22:35] what does MaaS Enroll mean ? [22:50] zykes-, its how a machine registers with maas [22:57] Question - im trying to follow this link to install openstack: http://tinyurl.com/bqfpocl ... however im getting the following errors: http://pastie.org/private/r4cwz6kem3bxmuvuqwhfq [22:58] and here's my .cfg file [22:58] http://pastie.org/3892150 [23:00] hazmat: but why does it power down the machine after it's first been booted up + [23:01] zykes-, because after it tells maas to manage it, maas turns off the machine till something tells maas it wants to use the machine === fenris_ is now known as Guest98532 === Guest98532 is now known as ejat [23:03] ah ok [23:03] how can I setup interfaces using the MaaS ? [23:03] like bonding and so on [23:24] hi, I'm usinc lxc and tried to deploy a standard mysql charm from the storm on precise, but juju status shows it's pending and I don't know why. The logs inside the container hint at a network problem perhaps: [23:24] http://pastebin.ubuntu.com/980796/ [23:24] Failure: txzookeeper.client.ConnectionTimeoutException: could not connect before timeout [23:25] any hints as to where debug further? [23:36] hm, I only have one cgroup fs mounted, which is odd, a colleague has many (cpu, cpuacct, etc) mounted and it's working for him [23:41] ok, found it [23:41] ufw [23:43] Question - im new to juju. I've deployed lodgeit with "juju deploy lodgeit", juju status shows me a private ip. However, browsing to it gives an error. Has anyone gotten this working ?? if so what could i be missing ? [23:46] ahasenack, yeah.. ufw breaks lxc local provider typically [23:46] interesting that it interfered with cgroups [23:47] zykes-, i don't think bonding is supported yet by maas [23:47] pdtpatrick, what error? [23:47] pdtpatrick, this in local provider or ec2? .. on ec2 you also need to expose the service, via juju expose service_name [23:47] hazmat: I now don't think cgroups had anything to do with it [23:47] hazmat, this is a local service. It is giving me a 500 error so a server error. [23:48] hazmat: still something odd going on here in the sense that I don't have cgroup mounted after boot, but I can workaround that [23:51] hazmat, http://pastie.org/3892348 [23:53] pdtpatrick, that looks sane.. so you can't access it the page in a browser, or there is an error from the lodgeit app? [23:53] pdtpatrick, that sounds like a problem in the charm [23:54] hazmat, i have debug on another screen and im not seeing anything that says why it is failing. [23:55] ahasenack, do you have the cgroup-lite package? [23:55] ahasenack, thats the one that has the upstart jobs to auto mount the cgroups [23:55] hazmat: yep, that's the initscript I have to call manually after booting to get cgroup mounted, I haven't figured out why yet [23:56] it's upstart, and seem to trigger on /sys being mounted [23:56] seems [23:56] anyway, it's working now, it's just this odd thing [23:56] which I can workaround [23:56] ahasenack, cool though unfortunate [23:56] pdtpatrick, if you could pastebin the unit log i might be able to help more [23:57] * hazmat runs to the next uds session [23:58] hazmat, http://pastie.org/3892379