[00:27] Success! Finally! [00:28] Now just need to figure out cinder/ceph integration and I'm golden! [00:28] kurt_, did you get your horizon + keystone issue sorted? was just gonna respond on list [00:28] yes [00:28] thanks adam [00:28] ah [00:29] I tore down everything and started at keystone - set it to precise-grizzly [00:29] explicitly set the password too [00:29] and now it works :D [00:29] kurt_, ah. i was going to suggest just upgrading keystone from essex -> folsom -> grizzly [00:30] kurt_, setting the openstack-origin on the running service will upgrade the services if the origin you're setting is > than current [00:30] kurt_, also, you should be able to set the admin-password config option, and have it set it on a running keystone [00:31] adam_g: can I change it there any time and keystone will work it's magic? [00:32] kurt_, the password? yes, you should be able to [00:32] that's awesome [00:32] * kurt_ sighs [00:32] now just cinder/ceph stuff is all I need to figure out [00:33] kurt_, note ceph isn't a requirement [00:33] kurt_, without ceph, cinder will just serve its local storage as instance volumes [00:33] well, I set my cinder block device to "None" [00:39] adam_g: will it still use local storage then? [00:40] kurt_, no, you'd need to destroy it and redeploy it with block-device set to an available device on the system [00:40] adam_g: can I do that without disrupting other parts of the deployment? [00:41] just simply re-deploy cinder? [00:41] kurt_, yeah, you'd need to re-add the correct relations to it [00:43] Ok, but I don't have to tear anything else down and start over, right? [00:44] and can you confirm - if I would have destroyed keystone, would I have had to have destroyed the other services? [00:45] I think all I did was update the distro origin string and it worked after ripping everything else out beside mysql and rabbitmq [00:48] kurt_, you should be good to destroy cinder, redeploy it and re-add it to your cloud. it gets a bit hairy if you are already using cinder and have volumes attached to instances, but if its a new cloud it should be okay [00:49] adam_g: thanks. Are there any readmes for cinder set up or is google my uncle? [00:50] kurt_, what do you mean by setup? [00:50] kurt_, deployment using juju? or just general usage? [00:50] prepping volumes, configuration, set up [00:52] I'm assuming I need to set up the cinder node in MAAS [00:52] how do I create persistent storage?? [00:52] ie. is it possible to recover volumes after a crash :D [00:54] kurt_, right, so... if you're not using ceph, you'd need to ensure there is an avaiable block device on the cinder node(s) when you are deploying it. if you are deploying with ceph, ceph essentially gives you that block device (and its HA and redundant) [00:55] actually thats wrong :) [00:55] without ceph: cinder uses local storage to create volumes and exports it from cinder nodes to instances via iSCSI [00:56] with ceph: cinder creates ceph block devices in the ceph cluster (instead of locally) and ceph exports them to instances via RBD (between ceph and nova-compute, cinder is just the mediator) [00:59] adam_g: can I add more cinder nodes which add additional storage to the system? [00:59] kurt_, yeah [01:00] kurt_, however there are no HA/redundancy guarantees. if you lose a cinder node that is exporting local storage, you lose those volumes as well [01:00] right [01:00] that's what ceph brings to the table [01:01] yup [01:01] can a combination of ceph and cinder nodes be used? [01:01] so can i start by playing around with cinder, then add ceph in after the fact? [01:02] ie. without losing my cinder volumes? [01:02] kurt_, you'd be using cinder in both cases, just different backends. cinder is growing support in Havana for multi-backends, so you could have some cinder exporting local storage, and others exporting ceph [01:02] interesting, ok [01:02] kurt_, one idea being that, as a cloud vendor, you'd charge more for the volumes that are replicated and HA :) [01:03] sure [01:03] then you'll need to start thinking about COS [01:03] :D [01:03] ya. [01:03] but atm im thinking about dinner. and need to run. :) [01:03] cya [01:03] LOL [01:03] :) [01:03] thanks adam [01:03] np === thumper-afk is now known as thumper === freeflying is now known as freeflying_away === defunctzombie is now known as defunctzombie_zz === tasdomas_afk is now known as tasdomas === defunctzombie_zz is now known as defunctzombie === defunctzombie is now known as defunctzombie_zz === stub` is now known as stub === freeflying_away is now known as freeflying === defunctzombie_zz is now known as defunctzombie === defunctzombie is now known as defunctzombie_zz [09:39] hi folks! === freeflying is now known as freeflying_away === freeflying_away is now known as freeflying [11:07] hazmat, thanks for the updates to the 0.7 branch btw - did the trick on saucy [11:08] just uploaded for re-introduction alongside juju-core [11:08] jcastro, ^^ fyi [11:27] jamespage, awesome === BradCrittenden is now known as bac === bac is now known as InigoMontoya === InigoMontoya is now known as bac [13:56] jamespage: link? [13:56] jamespage: oh you mean you reuploaded .7 [13:56] jcastro, yes [13:56] * jcastro nods === nate-finch is now known as natefinch [14:14] adam_g, https://code.launchpad.net/~james-page/charm-helpers/fixup_upstream_version/+merge/181840 [14:14] and adam_g: https://code.launchpad.net/~james-page/charm-helpers/ceph-redux/+merge/179948 [14:26] If I would like to have a wordpress and all its infrastructure on the same machine the mysql would be a subordinate? [14:39] X-warrior`: check out jcastro's post http://www.jorgecastro.org/2013/07/31/deploying-wordpress-to-the-cloud-with-juju/ [14:40] rick_h: will look it, ty === freeflying is now known as freeflying_away === natefinch is now known as natefinch-brb-af === natefinch-brb-af is now known as natefinch [15:38] marcoceppi: ping! === tasdomas is now known as tasdomas_afk [15:39] we're doing a charm school in 20 minutes on how to use the local/lxc provider! [15:40] jcastro: woot [15:45] hmm, I getting the install hook failing with both juju-gui and mysql [15:45] on the local provider [15:59] actually the charm school is on in 60 minutes [15:59] sorry for the mixup [16:11] how do I find the default port for the juju api endpoint? [16:20] Do you guys recommend using JeOS for ubuntu images within Openstack? [16:24] mthaddon: i think kapil landed a command to expose that [16:25] sidnei: he has a branch in progress for it, but it hasn't landed yet [16:25] ah, i see [16:25] https://bugs.launchpad.net/juju-core/+bug/1181382 [16:25] <_mup_> Bug #1181382: command for returning the api endpoint [16:25] but in any case, I think I've figured it out [16:25] mthaddon: it says merged? [16:26] oh, so it does - not sure why the bug is in progress still then [16:35] hey... [16:35] i'm using juju on precise [16:35] and before you tell me "dont do that", please listen further [16:35] i'm trying to test juju and precise for 12.04.3 relase (with plars) [16:36] against maas, i get [16:36] http://paste.ubuntu.com/6018407/ === TheRealMue is now known as TheMue [16:36] rvba, ? [16:54] heres a full explanation. [16:54] http://paste.ubuntu.com/6018470/ [16:54] jamespage, ^ any ideas? adam_g ? [16:54] smoser, yes - you need :80 in your url [16:54] smoser, helping plars in #ubuntu-release now [16:55] thanks [16:57] jamespage, do you happen to know about the other environments default release error ? [16:57] error: Environments configuration error: /home/ubuntu/.juju/environments.yaml: environments.quantal.default-series: expected 'precise', got 'quantal' [16:57] cause i'll just fix virtual-maas for that too. [16:57] smoser, that's cause juju 0.5 don't know nothing about raring [16:58] this does not get reported because no-one actually uses the juju in precise anymore [16:58] well, yes. [17:00] ok we're going to have a charm school on using the LXC/Local provider [17:00] starting nowish on http://ubuntuonair.com! [17:00] looking forward to it the local provider charm school jcastro and marcoceppi [17:07] no love at the url? [17:09] should be working [17:09] investigating [17:09] ppa:juju/devel or ppa:juju/stable [17:10] is it working for others? [17:12] more beer for thumper === defunctzombie_zz is now known as defunctzombie [17:15] can I tell you how excited I am about this provider? [17:17] marcoceppi: ec2-54-245-207-248.us-west-2.compute.amazonaws.com [17:18] I'm creating a charm to deploy some git site... What is the best way to git clone it? Create an ssh private key that will be on server and add this private key to git access? [17:18] X-warrior`: pass that "deployment-user" key into the charm as config [17:19] https://juju.ubuntu.com/docs/charms-config.html <-- docs for that [17:20] jcastro: m_3 looking it, ty [17:26] jinx [17:26] not seeing the juju local package in https://launchpad.net/~juju/+archive/devel [17:27] http://askubuntu.com/questions/65359/how-do-i-configure-juju-for-local-usage [17:27] sudo apt-get install juju-core lxc mongodb-server [17:28] I don't think juju-local has landed [17:28] ah ok, following along it sounded like the ppa would help me if I wasn't on saucy [17:28] it landed on saucy but not on the ppa, jcastro? [17:29] sidnei: right [17:29] it's not in the stable PPA [17:29] jcastro: or the devel [17:29] jamespage: do you know if we're putting "juju-local" in the PPA? [17:35] Is there a good post-deployment openstack-dashboard (horizon) set up/configuration guide available? [17:39] jcastro: is http://askubuntu.com/questions/65359/how-do-i-configure-juju-for-local-usage missing an expose command? [17:39] you don't need to expose locally [17:39] ah, gotcha === rogpeppe3 is now known as rogpeppe [17:49] jcastro: is there a g+ setting that turns off the "mute while I type" thing? [17:50] wooo this is sweet! juju-gui deployed in lxc deploying charmworld now. [17:50] rick_h: nice [17:54] jcastro: we should look at ingress rules in a vagrant file [17:58] I'm about to go Saucy. <--Sounds funny [17:59] marcoceppi: awesome job man! [18:03] marcoceppi: sorry I missed that part but where do I tell lxc/juju to use the squid-d-proxy? [18:03] I know the entire setup and have it, I just need to know where to point to [18:04] jcastro: you just put in your /etc/apt/apt.conf.d/ with the following [18:04] oh and that will migrate down to the containers? [18:04] http://paste.ubuntu.com/6018714/ [18:04] jcastro: yeah, Juju looks for Acquire::HTTP::Proxy line and copies it via cloud-init on deploy [18:06] don't forget to document this! [18:06] jcastro: I AM READY FOR DOCUMENTING [18:10] jcastro, it should already be in the stable PPA [18:10] jcastro, we need to sync up on packaging in devel [18:22] jcastro: crap... we forgot to talk about how well the containers survive reboot now [18:22] jamespage: doesn't appear to be in either PPA, stable or devel [18:23] jcastro, its in stable [18:23] https://launchpad.net/~juju/+archive/stable/+packages [18:23] juju-local dependency package for the Juju local provider [18:27] jamespage: oh I see, I didn't expand the expander [18:27] jamespage: my bad! [18:28] jamespage: or jam: should I go ahead and update the local instructions? [18:28] to use juju-local? [18:45] juju-core 1.13.2-4~1703~precise1 If I try and run 'juju bootstrap' against maas 1.2+bzr1373+dfsg-0ubuntu1~12.04.2 I end up with "ERROR juju supercommand.go:282 command failed: gomaasapi: got error back from server: 502 cannotconnect..." [18:46] what do I need to check? [19:20] Woohoo. Official acceptance for MS degree!!! Feels good. [19:22] Prerequisite waived!!! [19:28] weblife: congrats!! [19:28] MS in what? [19:28] Software Engineering [19:29] thank you also [19:29] cool === tasdomas_afk is now known as tasdomas [20:00] Can anyone tell me if https://bugs.launchpad.net/ubuntu/+source/quantum/+bug/1211764 affects the overall stability of the quantum-gateway in grizzly and makes it unusable? [20:00] <_mup_> Bug #1211764: Grizzly's python-quantumclient wrong dependencies [20:02] Should I not be able to run "quantum -v net-list" on the quantum node? [20:02] http://pastebin.ubuntu.com/6019060/ [20:03] * kurt_ thinks everyone has gone home for the weekend already [20:03] * kurt_ thinks he should too === tasdomas is now known as tasdomas_afk [21:05] I am just noticing that if I deploy a instance from local 'juju deploy --repository charms local:my-site my1' that trying to launch a second 'juju deploy --repository charms local:my-site my2' after I changed the code launches the same code from the 'my1' instance. Is this a bug or should I alter my second command? [21:28] weblife: AFAIK you will need to bootstrap new if you change the code. Maybe there is a short cut but I am not sure. [21:29] weblife: did you bootstrap with --upload-tools? [21:33] henninge: I could always change the name in the metadata file as a shortcut. Haven't tried upload-tools yet I will look into it. Thank you for the advice. [21:35] weblife: oh, there is also the revision thing. [21:35] weblife: you need to increase the revision number of the charm or else it will use the cached version, I believe. [21:36] henninge: That sounds better and less of a hassle. [22:20] weblife: You will need to use juju deploy -u [22:21] -u will upgrade from the repo if it's local, otherwise it'll use the cached version [22:21] weblife: upload-tools wont do what you need, that's just for versions of juju itself [22:21] and just changing the revision file for a deploy I don't think will automatically trigger it either [22:22] marcoceppi: Okay, will try next time around. Thank you! [22:22] weblife: `juju help deploy` for additional details [22:23] You can also opt to use upgrade-charm to upgrade the service running in place, depends on your workflow [22:24] marcoceppi: is there some flush that needs to be done when you change IPs in the nova-compute, quantum gateway? ie. floating IPs? [22:24] kurt_: for juju status? [22:24] in the dashboard [22:24] the floating IPs for VMs [22:25] kurt_: not that I know of. The floating ips should just work [22:25] I completely changed my topology and when I try to allocate IP, old ones still appear [22:25] maybe delete the project? [22:27] kurt_: possibly, still don't have that much openstack experience. Hoping to deploy it against maas soon and really put it through the works [22:27] yes, I've been successful getting it all the way up, but I'm seeing oddities in the interface that make me think I don't have everything right under the covers [22:28] in the dashboard rather - doing all the steps necessary to get a VM up, but things aren't working [22:58] m_3: hold off on that review of node-app. I just remembered that Chris Leas PPA installs everything to /etc/bin but the source installs to /etc/local/bin so it will have a problem when setting up the service configuration. Will fix this later this evening. bbl [22:58] I know what I need to do to correct it. Though