| grapz | Hi. I'm giving it a shot trying to fix a juju bug, but I was wondering how I can run it from the bzr checkout I have? If i do a bin/juju in my checked out folder, it seems to be loading the juju files that comes with the official package, not the ones I've changed. | 01:58 |
|---|---|---|
| hspencer | hey guys | 06:08 |
| hspencer | quick question | 06:08 |
| hspencer | why was the decision made to go with txAWS instead of boto? | 06:08 |
| hspencer | just curioius? | 06:08 |
| hspencer | curious i mean | 06:08 |
| SpamapS | hspencer: because txaws makes it easier to take advantage of twisted for doing other things while waiting for AWS | 08:23 |
| SpamapS | like responding to zookeeper watches | 08:23 |
| fwereade_ | grapz, if you're around, "export PYTHONPATH=$(pwd)" and then you can bin/juju ... to your heart's content | 09:57 |
| === fenris is now known as Guest25296 | ||
| === Guest25296 is now known as ejat | ||
| avl_ | hello | 12:22 |
| === avl_ is now known as ^LoGiN^ | ||
| ^LoGiN^ | does anybody familiar with juju/orchestra integration ? | 12:22 |
| ^LoGiN^ | i am getting following issue durung juju bootstrapping within orchestra environment | 12:23 |
| ^LoGiN^ | juju bootstrap | 12:24 |
| ^LoGiN^ | 2012-01-29 14:23:56,159 INFO Bootstrapping environment 'orchestra' (type: orchestra)... | 12:24 |
| ^LoGiN^ | Could not find any Cobbler systems marked as available and configured for network boot. | 12:24 |
| ^LoGiN^ | 2012-01-29 14:23:56,184 ERROR Could not find any Cobbler systems marked as available and configured for network boot. | 12:24 |
| fwereade_ | ^LoGiN^, can you tell me a bit more about your environment? | 12:28 |
| ^LoGiN^ | i have 2 servers | 12:29 |
| ^LoGiN^ | 1 orchestra server+juju | 12:29 |
| ^LoGiN^ | second one is orchestra installed node | 12:29 |
| ^LoGiN^ | all of them are running in Virtualbox VMS | 12:29 |
| fwereade_ | ^LoGiN^, ok, what systems do you have configured in cobbler on the orchestra server? | 12:30 |
| ^LoGiN^ | orchestra server works like a charm but when i try to run juju on it it doesn' bootstrap | 12:30 |
| fwereade_ | if you go to /cobbler_web on the orchestra server and log in there, there should be a sidebar link to"Systems" | 12:31 |
| ^LoGiN^ | there is one system configured | 12:31 |
| ^LoGiN^ | Profiles | 12:31 |
| ^LoGiN^ | Systems | 12:31 |
| ^LoGiN^ | Repos | 12:31 |
| ^LoGiN^ | Images | 12:31 |
| ^LoGiN^ | Kickstart Templates | 12:31 |
| ^LoGiN^ | Snippets | 12:31 |
| ^LoGiN^ | Management Classes | 12:31 |
| fwereade_ | ^LoGiN^, cool: would you check the management classes on that system? | 12:31 |
| ^LoGiN^ | Resources | 12:32 |
| ^LoGiN^ | Packages | 12:32 |
| ^LoGiN^ | Files | 12:32 |
| ^LoGiN^ | Actions | 12:32 |
| ^LoGiN^ | Import DVD | 12:32 |
| ^LoGiN^ | Sync ☼ | 12:32 |
| ^LoGiN^ | Reposync ☼ | 12:32 |
| ^LoGiN^ | Hardlink ☼ | 12:32 |
| ^LoGiN^ | Build ISO ☼ | 12:32 |
| ^LoGiN^ | Cobbler | 12:32 |
| ^LoGiN^ | Settings | 12:32 |
| ^LoGiN^ | Check | 12:32 |
| ^LoGiN^ | Events | 12:32 |
| ^LoGiN^ | Online Documentation | 12:32 |
| ^LoGiN^ | Online Help Chat | 12:32 |
| ^LoGiN^ | Systems | 12:32 |
| ^LoGiN^ | Create new system | 12:32 |
| ^LoGiN^ | Items/page: ⇐ ⇒ | 12:32 |
| ^LoGiN^ | Name ↓ Profile Netboot_Enabled Actions | 12:32 |
| ^LoGiN^ | oneiric01.openstack.lan oneiric-x86_64-juju | 12:32 |
| ^LoGiN^ | how can i do it | 12:32 |
| ^LoGiN^ | in config ? | 12:32 |
| fwereade_ | ok, I think there's an "edit" link next tothe system in the systems view | 12:33 |
| ^LoGiN^ | orchestra-juju-available is selected for this system | 12:33 |
| fwereade_ | ^LoGiN^, cool, there should also be a netboot enabled checkbox somewhere | 12:34 |
| fwereade_ | ^LoGiN^, is that ticked? | 12:34 |
| ^LoGiN^ | yes | 12:34 |
| ^LoGiN^ | it is checked and i cannot change it | 12:35 |
| fwereade_ | ^LoGiN^, hmm, ok | 12:35 |
| * fwereade_ thinks | 12:35 | |
| fwereade_ | ^LoGiN^, would you do a "juju -v bootstrap" and pastebin me the output? | 12:35 |
| ^LoGiN^ | holdon | 12:35 |
| ^LoGiN^ | i think i found it | 12:35 |
| fwereade_ | ^LoGiN^, cool | 12:36 |
| ^LoGiN^ | when i have started the system it was enabled | 12:36 |
| ^LoGiN^ | after deployment it was disabled somehow not by me | 12:36 |
| ^LoGiN^ | works now | 12:36 |
| ^LoGiN^ | thanks a lot | 12:36 |
| fwereade_ | ^LoGiN^, yeah, that's intentional -- once you've deployed a bunch of stuff to that system we expect you'd be pissed off if it reinstalled itself from scratch just because you rebooted it | 12:37 |
| ^LoGiN^ | i see | 12:37 |
| fwereade_ | ^LoGiN^, if you terminate-machine that system, or if you destroy-environment, then it should be set back to the state it was in before yu started | 12:37 |
| ^LoGiN^ | how can i access to the newly installed system ? | 12:38 |
| ^LoGiN^ | i have generated key but it asks me a password which i don't know | 12:38 |
| fwereade_ | ^LoGiN^, once it's gone through the whole process (which does take a while, it installs it from scratch) -- ie when juju status actually shows it as running -- you should be able to "juju ssh 0" (to machine 0, substitute other numbers as required) | 12:39 |
| ^LoGiN^ | i see | 12:40 |
| fwereade_ | ^LoGiN^, btw, I'd suggest having 4 VMs to play around with really: one orchestra server, one for juju to bootstrap on, and 2 more so you can deploy 2 services and relate them | 12:40 |
| ^LoGiN^ | do i need to run coation ?mmandprompt for massive deployments or there will be some web interface integr | 12:40 |
| fwereade_ | ^LoGiN^, if you want to ssh manually it should be fine: juju will have installed your public key, and you can "ssh ubuntu@..." | 12:41 |
| ^LoGiN^ | do i need to run commandprompt for massive deployments or there will be some web interface integration? | 12:41 |
| ^LoGiN^ | i want to deploy openstack cluster on it | 12:41 |
| fwereade_ | ^LoGiN^, we'd definitely like a web interface but for now it's just the command line really | 12:41 |
| fwereade_ | ^LoGiN^, I've been focused on the framework much more than the charms but as I understand it the openstack charms are in a good state | 12:42 |
| fwereade_ | ^LoGiN^, I should point out that it's readily scriptable | 12:42 |
| fwereade_ | ^LoGiN^, and I'm 99% sure that you can "juju add-unit foo -n 20" for 20 new units of foo | 12:43 |
| fwereade_ | ^LoGiN^, let me double-check that | 12:43 |
| fwereade_ | ^LoGiN^, yep, -n or --num-units | 12:43 |
| ^LoGiN^ | yes i am trying to play with that to deploy in production | 12:44 |
| ^LoGiN^ | juju bootstrap | 12:44 |
| ^LoGiN^ | 2012-01-29 14:35:40,680 INFO Bootstrapping environment 'orchestra' (type: orchestra)... | 12:44 |
| ^LoGiN^ | 2012-01-29 14:35:40,920 INFO 'bootstrap' command finished successfully | 12:44 |
| fwereade_ | ^LoGiN^, sweet :D | 12:44 |
| fwereade_ | ^LoGiN^, btw, are you familiar with the local provider? I understand not using it if you're trying to do the closest-possible thing to a real orchestra deployment, but when you're casually experimenting I think it's the most convenient way to go | 12:46 |
| fwereade_ | ^LoGiN^, it brings up new services in LXC containers on your local machine, so you shouldn't have to configure VMs etc | 12:46 |
| ^LoGiN^ | actually i am trying to deploy smth realistic | 12:46 |
| ^LoGiN^ | to make a copy on real systems | 12:46 |
| fwereade_ | ^LoGiN^, sounds sensible :) | 12:46 |
| ^LoGiN^ | actually i have demand from customers to deploy mixed environment(openvz,kvm,lxc) | 12:47 |
| ^LoGiN^ | can you tell me something about orchestra dns ? | 12:48 |
| fwereade_ | ^LoGiN^, interesting, I know cobbler can work with kvm machines, that's what I developed against | 12:48 |
| ^LoGiN^ | how does it work | 12:49 |
| ^LoGiN^ | where machine names are added | 12:49 |
| fwereade_ | ^LoGiN^, ...heh, probably not, I'm afraid I'm really not a networking guy :( | 12:49 |
| ^LoGiN^ | ok | 12:49 |
| fwereade_ | ^LoGiN^, sorry about that | 12:50 |
| ^LoGiN^ | np, | 12:50 |
| fwereade_ | ^LoGiN^, #ubuntu-server may be your best bet there | 12:50 |
| ^LoGiN^ | anyway thanks for steering me out of the problem | 12:50 |
| fwereade_ | ^LoGiN^, a pleasure, glad to help :) | 12:50 |
| ^LoGiN^ | by the way do you know what can be the issue ? | 12:54 |
| ^LoGiN^ | stic | 12:54 |
| ^LoGiN^ | <^LoGiN^> to make a copy on real systems | 12:54 |
| ^LoGiN^ | <fwereade_> ^LoGiN^, sounds sensible :) | 12:54 |
| ^LoGiN^ | <^LoGiN^> actually i have demand from customers to deploy mixed environment(openvz,kvm,lxc) | 12:54 |
| ^LoGiN^ | * rog вийшов (Quit: Konversation terminated!) | 12:54 |
| ^LoGiN^ | juju status | 12:55 |
| ^LoGiN^ | 2012-01-29 14:54:28,813 INFO Connecting to environment. | 12:55 |
| ^LoGiN^ | 2012-01-29 14:54:34,274 ERROR Invalid host for SSH forwarding: ssh: Could not resolve hostname oneiric01.openstack.lan: Name or service not known | 12:55 |
| ^LoGiN^ | Cannot connect to machine MTMyNzgzMzQwNS45Nzg3NDg1MS42MjQ1NTY (perhaps still initializing): Invalid host for SSH forwarding: ssh: Could not resolve hostname oneiric01.openstack.lan: Name or service not known | 12:55 |
| ^LoGiN^ | 2012-01-29 14:54:34,283 ERROR Cannot connect to machine MTMyNzgzMzQwNS45Nzg3NDg1MS42MjQ1NTY (perhaps still initializing): Invalid host for SSH forwarding: ssh: Could not resolve hostname oneiric01.openstack.lan: Name or service not known | 12:55 |
| fwereade_ | ^LoGiN^, hm, assuming you know that machine is installed and running... sounds like network configuration :( | 12:56 |
| fwereade_ | ^LoGiN^, re: mixed environment, I'm not sure how well cobbler will play with containers as opposed to VMs -- I do know that some people are using the local provider on single servers to deploy in LXC containers, but I'm not sure we really *recommend* that | 12:58 |
| ^LoGiN^ | i mean we would have 2 or 3 types of nodes | 12:58 |
| ^LoGiN^ | but management will be centralised by cobbler+juju | 12:59 |
| fwereade_ | ^LoGiN^, understood -- and we've talked about similar use cases, but I'm afraid that it's not planned for anytime very soon | 13:01 |
| fwereade_ | ^LoGiN^, can I ask you to enter a bug on launchpad describing your use case? it's the best way to make sure it gets proper visibility | 13:02 |
| ^LoGiN^ | anyway it would be very interesting for optimal resource distribution | 13:02 |
| fwereade_ | ^LoGiN^, definitely agreed | 13:02 |
| fwereade_ | ^LoGiN^, one thing that's definitely on our minds (although certainly not in time for 12.04 sadly) is machine sharing by using containers | 13:03 |
| fwereade_ | ^LoGiN^, I'm not au fait with the details but I understand that the networking is the really tricky bit there | 13:03 |
| ^LoGiN^ | very useful thing for extreme fast service rollouts) | 13:04 |
| fwereade_ | ^LoGiN^, we will have "subordinate charms" by 12.04, which are "helper" services (sharing machines *without* segregation), but I don't think that really addresses your use case | 13:04 |
| fwereade_ | ^LoGiN^, definitely | 13:05 |
| ^LoGiN^ | first of all i need to understand how could it be used i real application world | 13:06 |
| ^LoGiN^ | for now it is still blurry | 13:06 |
| fwereade_ | ^LoGiN^, I'm not sure how best to help there -- I rather like the "it's like apt, but for whole servers" formulation, but I'm not sure that answers your question | 13:15 |
| grapz | fwereade_, thanks :) | 13:58 |
| SpamapS | ssh -A clint-MacBookPro.local | 16:00 |
| SpamapS | DOHHHH | 16:00 |
| fwereade_ | grapz, btw, just shout if you need pointers about the code | 16:31 |
| hspencer | SpamapS, thanks for the response | 17:08 |
| hspencer | appreciate it | 17:08 |
| === flaviamissi__ is now known as flaviamissi | ||
| === flaviamissi_ is now known as flaviamissi | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!