=== defunctzombie is now known as defunctzombie_zz | ||
=== defunctzombie_zz is now known as defunctzombie | ||
=== defunctzombie is now known as defunctzombie_zz | ||
=== wedgwood_away is now known as wedgwood | ||
jcastro | jamespage: man, you guys totally crushed the incoming charm queue | 12:21 |
---|---|---|
jcastro | http://jujucharms.com/review-queue | 12:21 |
sinzui | hey charmer's is it possible to restart the agent on a unit. I have a stopped agent, but I can see the unit's service appears to be fine | 13:14 |
jcastro | hey FunnyLookinHat | 15:33 |
jcastro | did you ever find out if it worked on Rackspace? | 15:33 |
FunnyLookinHat | jcastro, we're close | 15:33 |
FunnyLookinHat | Two issues: | 15:34 |
FunnyLookinHat | 1) RS currently is returning a 411 error when trying to auth with Juju Core - the strange this is that the request isn't being sent with a content-length header... but they really shouldn't be rejecting the request for that I don't think | 15:34 |
FunnyLookinHat | 2) RS doesn't _currently_ support security groups - nova / quantum - but they're "working on it" - in reality, a better scenario might be a noop flag to not use security. | 15:35 |
jcastro | also did you notice a new juju-core release announcement on the list? | 15:35 |
FunnyLookinHat | so I'm currently forking out code to remove all Open/Close port stuff to see if it will "work" that way - and if so I might submit a patch for the noop, | 15:35 |
FunnyLookinHat | Ah no - which list? | 15:35 |
FunnyLookinHat | I don't think I'm on it. | 15:35 |
jcastro | https://lists.ubuntu.com/mailman/listinfo/juju | 15:36 |
FunnyLookinHat | Thanks | 15:36 |
SpamapS | FunnyLookinHat: how is rackspace working w/o cloud-init ? | 15:40 |
FunnyLookinHat | SpamapS, cloud-init ? I don't know... I'm only going off of what hazmat has suggested and my own conversations with #rackspace | 15:41 |
FunnyLookinHat | Realistically - I've been stuck at this 411 error for three days now, so I'm going completely off of conversations and not experience | 15:42 |
=== defunctzombie_zz is now known as defunctzombie | ||
hazmat | FunnyLookinHat, SpamapS, they have config drive support for userdata afaicr | 16:24 |
hazmat | smoser was working with them on that for cloudinit | 16:28 |
SpamapS | hazmat: and cloud-init in their ubuntu images? | 16:43 |
hazmat | SpamapS, indeed | 16:49 |
SpamapS | hazmat: sweet | 16:50 |
=== wedgwood is now known as wedgwood_away | ||
=== scuttlemonkey_ is now known as scuttlemonkey | ||
=== defunctzombie is now known as defunctzombie_zz | ||
kmyst | i'm getting a ConnectionRefusedError trying to issue a juju bootstrap against a node I enlisted and commissioned...the enviroments.yaml file i used was the one from the maas documents. what (if anything) is wrong? | 20:58 |
marcoceppi | kmyst: can you run juju bootstrap -v and include the output in a pastebin? | 21:01 |
kmyst | marcoceppi: sure thing | 21:02 |
kmyst | marcoceppi: http://pastebin.com/uzsKubYy | 21:03 |
marcoceppi | kmyst: What's maas-server set to? Do you have access to that address? | 21:06 |
kmyst | marcoceppi: what do you mean set to? and yes i'm logged in to the machine | 21:07 |
marcoceppi | kmyst: in your environments.yaml, what's it set to | 21:08 |
kmyst | marcoceppi: oh! http://192.168.99.1/MAAS i've also tried http://192.168.99.1:80 and http://192.168.99.1:5240 and possibly tacked /MAAS onto the end of those last two | 21:09 |
marcoceppi | try http://192.168.99.1:80/MAAS | 21:09 |
kmyst | marcoceppi: http://pastebin.com/tmT90vD0 | 21:11 |
marcoceppi | kmyst: alright! You've connected. Now, it's saying you don't have any nodes "commisioned" (I think that's the term, it's been a long time since I've used maas and juju) | 21:12 |
kmyst | marcoceppi: uh well it was in commissioning and then it went to ready then i told it to start and now it's allocated and installed and running :) | 21:14 |
marcoceppi | kmyst: so, you'll need more nodes commissioned/ready in order to deploy things | 21:15 |
kmyst | marcoceppi: so one isn't enough for testing in my lab? | 21:17 |
marcoceppi | kmyst: no, so you'll need multiple ones. The idea is MAAS provides the ability to treat bare metal like cloud servers | 21:17 |
kmyst | marcoceppi: i just never got juju to do anything since i got the node up and running/allocated/installed | 21:17 |
marcoceppi | So when you bootstrap and deploy you allocate physical maas nodes to juju for it to install a charm/service on each | 21:18 |
kmyst | marcoceppi: right i know that much and was testing to try and spin up one cloud server | 21:18 |
kmyst | so even with me doing the above i can't juju against the one node? | 21:19 |
marcoceppi | kmyst: you can bootstrap against one node, but you'll need another node to deploy anything | 21:21 |
marcoceppi | bootstrap is just what's used for the orchestration, you must bootstrap (and use one node) before you can start deploying things | 21:22 |
kmyst | marcoceppi: gotcha so if i'm reading this output correctly the bootstrap failed because no matching node is available? | 21:25 |
marcoceppi | kmyst: correct. So did you provision the maas node, then install juju on that and are trying to bootstrap from within the "only" maas machine? | 21:25 |
kmyst | basically | 21:26 |
marcoceppi | kmyst: so, you don't need to actually do that :) | 21:26 |
kmyst | missed that in the documents :) | 21:26 |
marcoceppi | You run juju from your local machine, as a client, when you bootstrap it it'll then use that "only" node as the bootstrap node. The more nodes you add the more things you can deploy with | 21:27 |
kmyst | i feel a facepalm coming on | 21:27 |
kmyst | i got that one node declared then when it was in commissioning i tried juju, nothing, so i powered it up and it did its thing and went to ready so i tried juju again with no results so i then just started it from the nodes page and it went to allocated | 21:30 |
kmyst | the whole while i was thinking juju wasn't supposed to work at this point yet :) | 21:30 |
marcoceppi | kmyst: so once it's commisisioned, and you power it on, and it does its thing, you just need to power it off at that point | 21:30 |
marcoceppi | then it's all ready to roll and you can juju against it :) | 21:30 |
kmyst | that's kinda what i thought but when i got that connection refused error i just kept pressing on and started it manually and watched it install the OS | 21:31 |
kmyst | so i guess my one node is goofed up eh? :) | 21:34 |
hazmat | SpamapS, FunnyLookinHat no cloud-init in default rackspace images | 21:47 |
SpamapS | doh | 21:56 |
FunnyLookinHat | hazmat, couldn't you just build a custom image w/ it? | 22:31 |
hazmat | FunnyLookinHat, definitely | 22:33 |
hazmat | FunnyLookinHat, but you'll get a deprecation warning from juju-core re default-image-id | 22:33 |
FunnyLookinHat | eh - I'll kindly thank juju-core for the warning and move on about my day | 22:33 |
FunnyLookinHat | :D | 22:33 |
FunnyLookinHat | hazmat, building an open stack cluster in the office tomorrow w/ a few 1Us we have lying around... any good tutorials on getting cloud-ready Ubuntu images to be used in it? | 22:34 |
hazmat | FunnyLookinHat, the images are @ hosted cloud-images.ubuntu.com, the tutorials for openstack + maas + juju also cover it (https://help.ubuntu.com/community/UbuntuCloudInfrastructure) | 22:36 |
FunnyLookinHat | hazmat, ah slick - bookmarked. Thanks ! | 22:36 |
hazmat | FunnyLookinHat, the ha setup info is here.. its a bit more involved https://wiki.ubuntu.com/ServerTeam/OpenStackHA | 22:37 |
FunnyLookinHat | oh wow - minimum recommended servers to just install for testing is 6 w/ juju-jitsu ? What about with juju-core ? | 22:37 |
hazmat | no different | 22:37 |
FunnyLookinHat | I feel like I should be able to get something running with 3... :D | 22:38 |
hazmat | except jitsu deploy-to is built in | 22:38 |
FunnyLookinHat | I'm aware that 6 is recommended for production though | 22:38 |
hazmat | via juju deploy --force-machine | 22:38 |
FunnyLookinHat | ah | 22:38 |
=== thumper is now known as thumper-afk |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!