=== tasdomas_afk is now known as tasdomas [06:06] greetings all [06:06] fwereade: when you come online, I'm available for a chat [06:06] wallyworld_: hey, just the guy I wanted to say hi to [06:11] wallyworld_: ? [06:14] morning all [06:21] hi dimitern [06:21] jam: hi, feeling better? [06:22] dimitern: yeah, fever is down, though still not 100%, but probably 80 or so :) [06:25] jam: good to hear it's improving [06:26] dimitern: yeah, my temperature spiked to almost 40, with accompanied whole-bod [06:26] body ache [06:27] and headache that was pretty bad [06:27] but only lasted the afternoon and a bit in the morning. [07:02] fwereade: ping [07:02] dimitern, jam, hi both [07:03] TheMue, hi also [07:03] fwereade: hey, take a look at this? https://codereview.appspot.com/9811044/ [07:04] fwereade: (should've changed the description, but will do before I submit) [07:11] dimitern, reviewed, looks sane, enough little quibbles that it's not quite there yet [07:14] fwereade, dimitern, jam: morning [07:22] fwereade: updated [07:25] dimitern, LGTM [07:26] fwereade: cheers! [07:26] TheMue: morning btw [07:26] anyone else wants a small, easy review? https://codereview.appspot.com/9811044/ [07:28] jam, TheMue: ^^ [07:37] bbiab [07:42] jam: hi, sorry i missed you, school pickup [07:42] jam: feeling better? [07:43] wallyworld_: hey [07:43] G'DAY [07:43] wallyworld_: about 80%, the fever, headache, and whole body ache are gone, but still a little upset stomach. [07:43] oops caps, sorry [07:43] at least it's recovering [07:43] wallyworld_: quick review? https://codereview.appspot.com/9811044/ [07:43] sure [07:43] dimitern: review done [07:44] wallyworld_: yeah, I can at least think and work, I just have to stay near a bathroom :) [07:44] TheMue: tyvm [07:56] fwereade: so now on to the 3 loops === wedgwood_away is now known as wedgwood [08:07] * TheMue just returned to the keyboard after hunting the escaped rabbits of our daughter ;) those guys found out how they can open the door. don't tell me anything about dumb rabbits anymore. :D [08:32] jam, is mgz around today? [08:37] fwereade: he is off the whole week, and part of next [08:38] I think he is back next Tues [10:03] Hi guys, when we were testing juju-core with MAAS we where using: "juju bootstrap --upload-tools --fake-series=precise,raring,quantal". [10:03] I saw somewhere that --upload-tools is a dev tool only there when building juju-core from source. [10:03] What should I be using instead now? (using juju-core from raring) [10:04] rvba: if I understand you correctly, you want to have the tools available on your MaaS cluster for multiple series [10:04] Yes [10:04] The MAAS cluster is raring and I want to deploy precise nodes. [10:04] we have a (fairly crufty but serviceable) command "juju sync-tools". Which pulls the tools from the public ec2 bucket, and puts them into your local environment. [10:05] you have to source EC2 credentials so it can read the public bucket. [10:05] (which is part of the 'crufty' nature of the command) [10:08] jam: thanks, I'll try that… [10:08] rvba: feel free to poke here for feedback on getting it working. It was designed for some internal use cases (sharing with HP and Canonistack), so it still needs some love for widespread adoption. [10:38] * fwereade lunch === gary_poster|away is now known as gary_poster [12:11] fwereade: ping [12:11] jam, pong [12:11] fwereade: you wanted to chat with me about some API changes. I was wondering if you wanted to have a hangout with me and ian [12:11] jam, how's it going? [12:11] sure, give me 2 mins [12:12] jam, you in a hangout already? [12:12] fwereade: not yet, but I can set one up [12:13] jam, thanks [12:14] well... I used to know where the start a hangout link was, give me a sec [12:17] fwereade: wallyworld_: https://plus.google.com/hangouts/_/6c0cd174f8cbe910b0ceaa885a6c216331e77f07 [13:21] mramm: heya, whens the next release planned? [13:22] 08:18 marcoceppi: mramm: Any idea when the next release of juju-core will be landing? [13:22] 08:19 mramm: davecheney was working on it this weekend [13:22] 08:19 mramm: I expect it out any day now, last I heard it was delayed by some build contention on the build servers [13:22] ah [13:22] seems like this must have come up in a meeting ;) [13:23] no I just wanna do cool stuff [13:23] and need a release [13:23] :) [13:23] cool [13:23] if it's builder contention I can help him [13:23] I will touch base with dave tonight [13:23] ok [13:23] but I take it it'll be a 24 hour turn around? [13:23] I just really want to be able to use the majority of charms that have relation-list -r :) [13:24] wait [13:24] is that in the PPA? [13:24] I'm not sure if that made it in before the last PPA release [13:25] 4:20 and 5:50 delays in the PPAs, distro is smaller. [13:25] interesting [13:25] mramm: I'm going to ask the team if we can get a bump in priority [13:25] ok [13:25] if a builder takes 6 hours to build it means cheney gets one shot a day to get it working [13:26] if not, we wait 24 more hours. [13:26] and I'll talk to dave about the release tonight (when he is in) [13:30] * jcastro nods [13:31] I'll send him a mail wrt. builders, see if there's anything I can do [13:45] When downloading the mongo binary, where should I put it? [13:49] Also, in juju-core's readme it refers to running `go install -v launchpad.net/juju-core/...` which doesn't actually work [13:51] "launchpad.net/juju-core/..." matched no packages [14:12] jcastro: doesn't matter if the fix is in the ppa, it needs to be in the tools tarballs too, unless you plan to use --upload-tools [14:12] fwereade: got a second to help me resolve building from source? [14:12] marcoceppi: I can help, I did it a few times [14:12] marcoceppi, sorry, meeting [14:12] ahasenack: ah bummer. :-/ [14:13] ahasenack: cool, I've followed the readme up until go install -v launchpad.net/juju-core/... which just gives me a matched no packages error [14:13] jcastro: https://wiki.canonical.com/InformationInfrastructure/IS/CanonicalOpenstack/CanonistackWithJujuCore and I have a g+ post about it too [14:13] marcoceppi: did you do "go get -v launchpad.net/juju-core/...", with the "..." at the end? [14:14] ahasenack: yup, let me try again [14:14] ahasenack: http://paste.ubuntu.com/5710325/ [14:15] marcoceppi: start with go get [14:15] ahasenack: fudge, I copied the wrong command [14:15] thanks [14:16] marcoceppi: you will also need to use boostrap with --upload-tools later, or else it will download the old version of the tools into your units, and not the trunk one you have now [14:16] ahasenack: ack, I'll hopefully remember to do that :) [14:17] marcoceppi: https://plus.google.com/114091308548248656535/posts/cLgeVv7xBcp hopefully helps [14:17] marcoceppi: with canonistack and other private clouds, there is a new complicator, I'm working on a post about that [14:17] it doesn't support default-image-id anymore, so there are a few other commands to run [14:17] ahasenack: so the readme mentions having to grab mongodb from source, etc. Do I just ignore that? [14:18] marcoceppi: yes, I didn't have to do that [14:18] go get probably takes care of all that? [14:18] ahasenack: cool, thanks [14:18] dpb1: I don't know [14:18] dpb1: maybe mongodb doesn't need to be updated that frequently and whatever juju is fetching is good enough [14:21] ya [14:36] is there a way to use juju image-metadata to specify default images for multiple regions? [14:43] ahasenack: is what you were talking about with "private" clouds and default-image-id related to this: error: cannot start bootstrap instance: no "precise" images in az-3.region-a.geo-1 with arches [amd64] [14:43] marcoceppi: hm, *maybe*, which cloud is that? hp? [14:44] ahasenack: yeah, hp [14:44] marcoceppi: do you have a public-bucket-url for that hp cloud in your environments.yaml file? [14:44] ahasenack: yes, https://region-a.geo-1.objects.hpcloudsvc.com/v1/60502529753910 [14:44] marcoceppi: also, did you bootstrap with -v? Was there any mention of an url with juju-dist? [14:45] marcoceppi: ok, take a look: [14:45] marcoceppi: https://region-a.geo-1.objects.hpcloudsvc.com/v1/60502529753910/juju-dist/ [14:45] (I'm debugging with you now, I don't know yet what the problem is) [14:45] marcoceppi: https://region-a.geo-1.objects.hpcloudsvc.com/v1/60502529753910/juju-dist/streams/v1/imagemetadata.json and [14:45] ahasenack: (I'm all for debugging thanks for taking the time) [14:45] marcoceppi: https://region-a.geo-1.objects.hpcloudsvc.com/v1/60502529753910/juju-dist/streams/v1/index.json [14:45] are of interest to us [14:45] that uses the new "simplestreams" from smoser to specify the image id juju will use [14:46] ahasenack: so -v didn't give any mention of a url with juju-dist [14:46] marcoceppi: if I understand it correctly, it has an image id definition of 81078 for region az-1.region-a.geo-1 [14:46] just some warnings about depreciated configs [14:47] marcoceppi: ok, the way I understood it now is that you can only use region az-1.region-a.geo-1 [14:47] Ah, so I'm using az-3, let me try az-1 [14:47] marcoceppi: well, for some definition of "only", it's how it's setup. To use another one there are other steps to take [14:47] ahasenack: path of least resistance is the one for me atm [14:48] marcoceppi: ok, so switch regions and try again [14:48] ack, activating/switching [14:49] ahasenack: that resolved that, thanks [14:49] marcoceppi: nice! [14:59] ahasenack: so, how much harder is it to get this working with other az's? [14:59] marcoceppi: I'm still writing it up, only have a few notes so far. This is what I have for canonistack so far: https://pastebin.canonical.com/91689/ [15:00] marcoceppi: you have to run the new image-metadata command, then create a juju-dist bucket and upload the files to it [15:00] and adjust acls so it can be read anonymously [15:00] ahasenack: doesn't sound too bad [15:00] marcoceppi: I don't know yet how to make it work for multiple regions, as that image-metadata command seems to only create the files for one at a time [15:00] marcoceppi: I think for other regions you have to just edit the files manually [15:00] I mean, to support multiple regions in one single such file [15:01] ahasenack: I only have need to support on region that isn't supported, so that's out of scope for me anyways [15:01] one* [15:01] marcoceppi: hah, how murphy of you, needing the region that is not supported :) [15:02] hah, yeah. I've really lucked out. Figured the highest number region would be the best to use! [15:17] ahasenack, marcoceppi: fwiw, environments are single-region, so that point is hopefully moot [15:18] fwereade: given the location of the index.js and imagemetadata.js files, and their names [15:18] fwereade: I mean, the region is not in the path, or the filename [15:18] fwereade: so it has to be in the content [15:18] fwereade: so you have to be able to specify multiple regions inside these files, no? [15:18] ahasenack, if you're sharing a public-bucket across regions, I agree, you'll want to have metadata for all regions in play [15:19] fwereade: in canonistack at least hte public bucket is the same, I don't see a region differentiation on it [15:20] fwereade: I don't know about hp [15:20] fwereade: but the file in their public bucket only had one region, and juju failed when marcoceppi tried to bootstrap in another [15:20] ahasenack, I believe that we will be publishing simplestreams data to HP anyway [15:20] fwereade: so juju wasn't able to find the default image id for the other region [15:20] ahasenack, ha, ok [15:21] fwereade: so, the end goal, just to understand, is that those json files describe the default image id for all regions? [15:21] (amongst other things I suppose, since simplestreams are generic) [15:22] or, well, use different public buckets per environment [15:23] ahasenack, the goal is that the simplestreams files describe the images that are available, so that juju can pick a/the right one for what you need [15:24] fwereade: ok, this can be solved then by either adding all of them to the file, or using different files in different buckets [15:24] the difference being one single public-bucket-url for all regions, or one per region [15:25] ahasenack, in general we will be looking in cloud-images.ubuntu.com/releases [15:25] ahasenack, AIUI this will hold data for a whole bunch of public clouds [15:25] fwereade: that's for public clouds [15:25] ok [15:26] but for my own private cloud, I need to manage those simplestream files [15:26] ahasenack, right; and in canonistack I believe we are publishing simplestreams data and advertising it via openstack already for lcy01, just not for lcy02 [15:27] ahasenack, for your own private cloud, we need tooling, I agree [15:27] fwereade: so canonistack seems to be using something different. I managed to bootstrap on lcy01, but there is no juju-dist/streams/v1/index.json file for example [15:28] ahasenack, it's published... somehow, in openstack, that I am trying to remember and look up [15:28] so it's not using simplestreams as far as I understand, but something different === mthaddon` is now known as mthaddon [17:16] Are gocheck tests run in the order they're defined? [17:17] I'm getting "cannot add service \"riak\": service already exists" on the first test, so I'm curious./ [17:20] nvm -gocheck.vv shows the order. === wedgwood is now known as wedgwood_away [17:48] so, quitting, cu tomorrow [17:50] hm, I wish bootstrap -v showed the buckets it is querying [19:07] Hm. Test fails with "service \"riak\" not found", so I try deploying it at the beginning of the test, which gives me "cannot add service \"riak\": service already exists". [21:10] Is there a way to enable verbose request logging with juju-core ? including HTTP headers, etc. [21:10] I was hoping to not use wireshark, but that would work if necessary... [21:13] morning folks [21:16] hmmm [21:16] oh hey thumper :D ( David @ S76 ) [21:16] hi FunnyLookinHat === tasdomas is now known as tasdomas_afk [21:17] thumper, You aware of any means to do verbose logging of all requests made when bootstrapping ? [21:17] ( or in general ) [21:17] umm... [21:17] I think -v is about as verbose as it gets (right now) [21:18] the whole logging, and output stuff is something that I have an interest in fixing [21:19] Ah ok - trying to see if we can close the gap on rackspace support... narrowed it down to security groups ( which they are implementing ) and a strange 411 error when trying to authenticate [21:21] FunnyLookinHat, --debug is more verbose than -v, but I can't guarantee it'll give you what you're after; logging is somewhat spotty [21:21] Ah ok - thanks fwereade [21:22] hi fwereade [21:22] thumper, heyhey [21:22] fwereade: did you want a catch up chat? [21:22] sorry about yesterday [21:22] snow day here [21:22] very chaotic [21:22] thumper, sgtm, although probably not a very long one [21:22] thumper, no worries [21:22] sounds fine [21:22] I have a chat with mramm at 10 [21:22] (in about 35min) [21:22] different times I can handle, but different seasons? madness [21:23] haha [21:23] * thumper writes a todo list [21:23] Another snow day? [21:23] ;) [21:23] mramm: nah, sunny and high of 11°C [21:23] I can be flexible with that meeting time if it helps [21:23] well, should be sunny later [21:23] ahh [21:23] pretty grey right now [21:23] got it [21:24] snow is all melting [21:24] much to the kids disappointment [21:24] still have two at home sick [21:24] mramm, you shouldn't need to be too flexible, I'll want to go to bed around then ;) [21:24] but they aren't the two that fight so much [23:50] * thumper heads to the gym [23:50] wallyworld_: I'll chat with you when I'm back [23:50] ok