=== niemeyer_ is now known as niemeyer [02:43] Hey there [08:11] test [14:29] Good mornings [15:11] hi all [15:11] kim0: now can say it officially :) ^o^ [15:13] koolhead17: hey :) [15:14] koolhead17: so what's up [15:14] yes i can officially work on it !! :) [15:15] koolhead17: Ensemble ? [15:15] its yeah :) [15:15] hhee cool :) [15:15] rock n roll [15:15] its in my TODO now [15:15] koolhead17: so what do you plan to do [15:16] kim0: eucalyptus support 4 ensemble [15:16] awesome [15:18] :) [15:19] hey TeTeT [15:19] hi koolhead17 [15:19] koolhead17: Wow, sweet [15:20] niemeyer: koolhead17 will probably need some assistance on pointing ensemble to use a euca cloud [15:20] TeTeT: i found one more issue in cobbler and asked a question on launchpad mentioning the dirty solution for the same :D thanks 2 you [15:21] hey niemeyer :) [15:22] koolhead17: haven't done anything but chatting, nevermind nevertheless ;) [15:22] koolhead17: if you get ensemble with UEC going, let me know, I'm interested in that as well [15:23] TeTeT: sure. am n00b for Amazon stuff so would love to see magic via euca tools on eucalyptus [15:24] TeTeT: don't worry, I'll keep pushing koolhead17 till he gets it running :P [15:25] kim0: thx ;) [15:29] kim0: That sounds good [15:29] koolhead17: Making it work well with the euca cloud may actually be your first major task :) [15:29] koolhead17: There are no big reasons why this shouldn't work, but it's a different software backing it, so minor incompatibilities are expected. [15:29] niemeyer: am still a python n00b. [15:29] koolhead17: We work mostly with the "real" EC2 [15:30] koolhead17: Me too [15:30] koolhead17 just mentioned he's refreshing info on eucalyptus first .. so it might be a bit too early [15:30] me smiles [15:30] koolhead17: There's only one way to learn anything.. [15:30] and it is? :) [15:30] koolhead17: Doing it [15:31] niemeyer: +1 [15:31] koolhead17: I'd want to refresh my python too .. I'd love to work with you on that [15:31] koolhead17: I don't really think you need much brushing up on eucalytpus .. :) [15:32] koolhead17: when you feel ready . ping me [15:33] kim0: hmm [15:33] koolhead17: If you don't mind, please pay attention to the areas of the documentation you find lacking [15:33] sure will do that niemeyer [15:33] koolhead17: Or confusing [15:33] niemeyer: I suppose we'd need to edit some ensemble configs to point it to euca ? [15:33] koolhead17: The first read is a great moment to catch stuff that makes no sense, or is incomplete [15:33] or is it just the ec2 env vars [15:33] kim0: That's right [15:34] ahasenack has done it before, and may even be able to provide some further guidelines [15:34] kim0: Yes, ec2_uri and s3_uri [15:34] ah cool [15:35] IIRC, there's a bug ahasenack mentioned related to the machine size [15:35] Since Euca has a different concept of machine size [15:35] I don't think it will work because of the txaws bug [15:36] and you have to change a line in ensemble because it expects an amazon region, and "Eucalyptus" isn't one [15:36] ahasenack: Ah, that's it [15:36] koolhead17: It might actually be less frustrating for you to start with the "real" EC2 [15:36] koolhead17: Stumbling upon those bugs from very early on may be a bit boring [15:37] niemeyer: i don`t have amazon account and i dont have Credit Card 4 there free account :( [15:37] kim0: I started with this: http://pastebin.ubuntu.com/610642/ [15:37] ahasenack: thanks man! [15:37] koolhead17: ^ [15:37] kim0: but there is a bug in txaws, let me get the number [15:38] kim0: https://bugs.launchpad.net/txaws/+bug/782375 [15:38] <_mup_> Bug #782375: URLContext() loses port and path from service endpoint < https://launchpad.net/bugs/782375 > [15:38] koolhead17: Hmm [15:39] kim0: and you need to hack the eucalyptus region in like this: http://pastebin.ubuntu.com/610643/ [15:39] ahasenack: Have you ever tried to use Ensemble through cloud deck? [15:39] niemeyer: no [15:39] koolhead17: I may find a way to fix that.. but keep going with Euca for the moment [15:39] kim0: even after I locally fixed the txaws bug, though, I got another error about wrong signature, so I stopped there. I assumed I did something wrong in my local fix [15:39] ahasenack: wow, thanks a million! [15:39] yes sure niemeyer [15:40] koolhead17: please check the awesome info ahasenack just mentioned :) [15:40] kim0: yes checkin it :) [15:40] http://aws.typepad.com/aws/2011/05/identity-and-access-management-console-support.html [15:40] Handy [15:41] obino: Hey! [15:42] hello niemeyer! [15:42] how was your trip? [15:44] obino: Long and boring, as long flights should be :-) [15:45] obino: We were just pondering about integration with Euca [15:45] obino: I guess we have a few extra steps still to make it work smoothly [15:46] tell me more :) [15:46] obino: The issues are mostly on our side [15:46] obino: We've provided pre-defined regions and machine sizes, and we'll have to relax that a bit for Euca [15:46] ok [15:46] got it [15:47] did you get access to the ECC? Are you testing against it? [15:47] obino: hey [15:47] obino: koolhead17 got it, and ahasenack tried too [15:47] hello koolhead17 [15:48] niemeyer: quick question, I have an environments.yaml with two environments, one marked as default, yet ensemble tried to bootstrap both, is that expected? Let me paste [15:49] http://pastebin.ubuntu.com/610645/ [15:49] ahasenack: No, that's a bug.. [15:49] it went through sample, and then canonicloud [15:49] ok [15:49] ahasenack: Yeah, config looks good.. it shouldn't try to do anything on both at the same time [15:50] niemeyer: koolhead17: let me know if you run into issues. I'll try to be around this channel too. [15:50] obino: Superb, thanks [15:50] obino: awesome :) [15:50] niemeyer: well, not at the same time, but in the same run/bootstrap command. One, then the other [15:50] but ok [15:50] koolhead17: There you go.. now you have another reason to try fixing support for Euca ;) [15:50] I think you got that [15:50] I will be traveling from tomorrow morning till sat night, but I will be online starting from sun (PDT) [15:50] ahasenack: Yeah [15:50] niemeyer: i would love to :D euca ++ [15:51] obino: Cool, thanks for your attention [15:51] np: I would love to see euca support too :) [15:51] when you have something running let me know! [15:51] * ahasenack -> lunch [15:52] ahasenack: Enjoy [15:52] obino: Will do [15:52] obino: We've been walking towards it [15:53] niemeyer: :) [15:55] ahasenack: Interesting.. apparently IAM in AWS doesn't actually solve the issues Cloud Deck does [15:55] bcsaller: Mornign [15:55] bcsaller: Feeling better? [15:56] niemeyer: yeah, thanks [15:56] bcsaller: Glad to hear it === niemeyer changed the topic of #ubuntu-ensemble to: http://j.mp/ensemble-dublin | http://ensemble.ubuntu.com/docs [16:02] jimbaker: The kanban is "fixed" [16:02] I'm building at people.canonical.com since it's taking a while for IS to fix the scheduling of it [16:15] obino: I just registered an account with username kim0 .. Would be great if you can approve it before flying .. thanks :) [16:16] kim0: :P [16:16] hehe :) [16:21] * niemeyer => lunch [17:00] <_mup_> ensemble/expose-provision-service-hierarchy r232 committed by jim.baker@canonical.com [17:00] <_mup_> Observer callbacks for watching service units in tests [17:07] kim0: did you apply for the ECC? I saw your account on the forum (web site) but not the ECC [17:08] obino: hmm no I didn't, where do I apply [17:08] obino: It does say, simply click on the "Apply" for account link .. but I couldn't find that [17:09] all I got was a login box [17:22] i need to take my puppy to the vet to take care of an eye infection in a few minutes [17:27] kim0: can you try again? Go to ecc.eucalyptus.com [17:28] there should be an apply link [17:31] andrew will approve your account as soon as you apply :) [17:34] kim0: heh [17:34] * koolhead17 wants to know if anyone from Berlin here? [17:38] kim0: cool. i can see the blog ^^ [17:41] * niemeyer waves [17:43] niemeyer: done with luch? [17:43] *lunch [17:43] * koolhead17 rushes 4 dinner :P [17:43] koolhead17: Yeah, good stuff ;) [17:43] koolhead17: Enjoy [17:52] obino: thanks .. just did it [17:56] kim0: you should be approved. Check the spam folder for the approval email [17:56] sadly that's where it ends up most of the time :( === deryck is now known as deryck[lunch] [18:13] jimbaker: The documentation for expose/unexpose doesn't seem to be formatted properly in terms of headers [18:13] jimbaker: It's showing twice under Drafts [18:14] Hi everybody. :) Whats the status of the LXC machine provider? [18:18] btw, formula upgrades are *awesome* :) [18:30] SpamapS: Yo [18:30] SpamapS: Sweet! [18:30] jimbaker: This is ready: https://code.launchpad.net/~jimbaker/ensemble/expose-open-ports-node/+merge/61646 [18:30] jimbaker: For being worked on [18:30] SpamapS: Have you been playing with it? [18:31] niemeyer: when applicable, I use upgrades to iterate on hook dev. [18:32] SpamapS: Nice [18:32] SpamapS: LXC has been evolving in our minds and in conversations with the LXC team [18:32] which is also good for testing the idempotency of my hook scripts [18:32] SpamapS: It's likely going to be the key feature for the milestone after the current one [18:33] SpamapS: Indeed [18:33] niemeyer: seems to me there is a bit of a challenge in providing a single s3-like place for the machine map [18:33] SpamapS: Yeah, but we should be able to workaround it somehow [18:33] SpamapS: We actually just need a place we can address for discovering the seeds [18:33] SpamapS: We might even use a local map [18:34] SpamapS: The real information is actually within the environment itself [18:37] SpamapS: Hmmm.. we actually store the formulas themselves there as well ATM [18:37] SpamapS: This is something I'd eventually like to kill [18:38] SpamapS: and have everything with the env itself [18:38] SpamapS: So it doesn't matter if we're using EC2, Rackspace, LXC, or whatever, the environment is even [18:38] I don't really know what you mean when you say "the environment" in this context. [18:42] niemeyer: also my mind keeps coming back to storage management.. that we're glossing over it a bit too much and should start to think about how to handle complex storage requirements... [18:43] SpamapS: "the environment" == "the machines which are part of a machine provider and assigned to be managed by Ensemble" [18:44] Like, there are some really good chef cookbooks for EBS+RAID .. you just throw in a call to them in your current recipe and viola.. /srv is on a RAID5 w/ 10 EBS volumes. [18:45] In order to duplicate that.. we have to monkey it into the install hook and we don't want to be making any cloud-specific calls in hooks. [18:45] SpamapS: Yeah, we'll certainly want to support more fine grained storage requirements [18:45] SpamapS: Indeed.. this is really a task for Ensemble itself.. it should work irrespective of the formula [18:45] I was just thinking that environments should be able to define storage pools. Libvirt has a lot of capability in this regard that can be studied for ideas. [18:45] SpamapS: Agreed. We discussed these ideas in the very first planning sprint of Ensemble.. we'll eventually get there === deryck[lunch] is now known as deryck [19:02] SpamapS: The next sprint will likely be more focused on the tasks for the current milestone, which is most importantly stabilizing what we've got for usage [19:02] SpamapS: After that, we should probably have another technical sprint to debate such idas [19:02] ideas [19:02] Yeah thats bolt-on refinement, we're still missing settings and containers.. two things I'm quite eager to play with. [19:04] SpamapS: Settings are mostly there [19:04] SpamapS: bcsaller should have them in review any minute now! [19:04] * niemeyer looks at bcsaller [19:04] sweeeet [19:05] back from the vet [19:16] niemeyer: another thing that I'm still not sure about. If my system detects a change that ensemble needs to know about.. is there a way to do that? [19:16] niemeyer: for instance, if my IP changes. [19:24] SpamapS: In an intereview, but will be you with you in a sec [19:37] SpamapS: So yeah, we need a way to call the ensemble tools out of band [19:37] SpamapS: out-of-band as in, not within a hook [19:37] SpamapS: There's actually no intrinsic reason why this shouldn't work.. we just need to tweak them so they work [19:43] niemeyer: basically just need to have the socket env right? [19:44] SpamapS: Yeah [19:44] SpamapS: and some internals sorted out [19:44] SpamapS: We currently have a "cache" for changes [19:44] SpamapS: So that everything is committed at the end [19:44] niemeyer: I'm struggling to find real world use cases though.. I think in most cases the only thing I'd want to be able to do would be to remove myself from the service temporarily. [19:44] SpamapS: This won't make sense for those out-of-band requests [19:45] SpamapS: Yeah, the use cases will likely be related to external events [19:45] niemeyer: most services will handle this on their own, they won't need reconfiguration tho.. so.. yeah, I don't think its all that necessary right now. [19:47] kim0: Thanks for the post! [19:47] kim0: As a minor, can you please tweak the doc link to point to the new location? [19:48] kim0: Having LXC for the 11.10 cycle is a bit optimistic as well, btw :) [19:50] * SpamapS will do anything he can do help w/ LXC ASAP :) [19:51] As fast as t1.micro's spawn ... LXC containers would spawn faster. :) [19:56] SpamapS: True. It'd certainly be cool to have it sooner. [20:46] bcsaller: Are you off today as well? [20:46] niemeyer: no, working [20:46] things are a little slow as I'm still on my old laptop [20:47] bcsaller: Cool [20:47] bcsaller: Do you have any comments on kapil's input about config-set? [20:47] niemeyer: did you ever get a chance to look at the branch in review or are you waiting for the last one [20:48] niemeyer: I made those changes, yes [20:48] but I think only in the latest branch [20:48] bcsaller: Can you please update the merge proposal with your feeling about what he said and whether you have pushed new changes or not? [20:48] ahh, ok [20:49] Otherwise it feels like a monologue [20:49] bcsaller: Thanks [21:12] Everything moved to the dublin milestone [21:21] Fading a bit here.. will step out for a while