=== benonsoftware is now known as Guest92794 [10:58] hazmat: when you get a few minutes i [10:58] 'd like to pick your brain on what it will take to add a hpcloud provider === almaisan-away is now known as al-maisan [11:47] imbrandon, greetings [11:51] imbrandon, it wouldn't take much.. the compute api should work with juju via its ec2 compatiblity, we just need a place to store a few files per s3 compatiblity, right now compute and storage is composed as a single provider config. There are lots of options, separate out storage and compute within a provider, allow the provider to provide its own storage on top of compute, expose an s3 api to hpc's storage layer. === al-maisan is now known as almaisan-away === almaisan-away is now known as al-maisan [12:42] If I can have a number of environments in ~/.juju/environments.yaml -- how do I tell juju which one should be used e.g. when running a 'juju deploy' command [12:42] ? [12:46] ah, http://askubuntu.com/questions/65364/how-can-i-configure-multiple-deployment-environments-for-juju [13:08] I went through the mysql/wordpress example and have a question: even after a 'juju unexpose wordpress' command the wordpress site continues to be visible and usable in the browser -- is that how it should be? [13:08] If yes, what is the purpose of the 'unexpose' command? [13:15] al-maisan, -e env name [13:15] al-maisan, expose/unexpose only work for ec2 [13:15] al-maisan, we need to start using machine level firewalls instead of provider capabilities to implement that functionality [13:16] currently it relies on provider capabilities and is limited to ec2 [13:16] ah, I see, thanks! [13:17] al-maisan, re environment selection, you can also define a default: env_name in environments.yaml.. and you can also specify it via JUJU_ENV env var [13:19] That's very nice, thanks. I would imagine the question about environments would be a FAQ :-) [13:21] The FAQs also state "If you need a custom AMI you can specify it in the environments.yaml file" .. is this documented somewhere or is there an example of this? [13:21] al-maisan, its there is juju doc i suppose, not in FAQ [13:21] * al-maisan looks [13:29] al-maisan, we kinda of want to avoid people using those.. we have it right now mostly to support private clouds, where we don't have good introspection into the available images [13:29] al-maisan, right now a charm determines the distro series etc, when people use ami-id, it basically breaks that notion.. [13:30] hmm .. I see .. good to know [13:30] al-maisan, there's a few new features that are underdocumented, after we get our next upload done, a focus for the team will be on improving docs [13:30] cool [13:30] specifically i'm thinking of the constraints functionality (select a machine based on capabilities) and the subordinate charms [13:30] as far as poorly documented features [13:31] contributions welcome! bzr branch lp:juju/docs :-) [13:31] re. custom AMIs .. there is the case to be made for systems like ours (a lot of additional software needs to be downloaded and installed on top of the plain ubuntu server AMI) [13:32] hazmat: re. documentation .. thanks for the pointer .. will add stuff as I discover it :-) [13:33] hazmat: going back to the custom AMI: in such cases it would be advantageous to use custom AMIs in order to reduce time/cost of deployment IMHO [13:52] al-maisan, shaving a minute seems like a petty cost [13:52] vs the cost of maintenance [13:52] or functional mismatch against the series that a charm was written for [13:53] That's a fair point. [13:53] al-maisan, we default to using the daily cloud build of any given series on ec2 [13:53] I see. [13:53] ie. its already up to date [13:53] gotcha === koolhead17 is now known as koolhead17|afk [14:36] hmm .. it's been a while .. how does one push a branch that should be stacked on lp:juju/docs ? [14:42] OK .. got it the "Merge into:" target needs to be lp:juju/docs [14:45] hazmat: we do not use the daily cloud build! [14:45] hazmat: we use the released build [14:45] SpamapS, oh.. [14:46] so fo precise we'll use beta2 [14:46] for even [14:46] SpamapS, you ever debug cloud-init? [14:46] hazmat: would be a nice thing to be able to influence the query process with some parameters like 'daily' [14:46] SpamapS, i've got the lxc cloud images working, but cloud-init refuses to run the cloud-config parts [14:46] hazmat: I have in the past.. [14:46] SpamapS, that capability is present, but perhaps we're not using it [14:47] SpamapS, its kinda of a mess in there, i've traced through the code, but debugging the absence of working isn't proving fruitful [14:47] re ^ cloudinit [14:48] Its picking up the right user-data just not doing anything with it [14:48] hazmat: we use the same mechanism in orchestra... [14:49] and I presume maas [15:23] SpamapS, indeed, maas has its own datasource [15:23] i guess it will wait till post weekend and rendevous with smoser === al-maisan is now known as almaisan-away [15:38] hazmat: but what about the orchestra data source? [15:41] SpamapS, dunno i assume it used the nocloud source [15:43] juju stuffed it into KSMETA as USER_DATA_BASE64 [16:27] SpamapS, what do you think of a jitsu cli command to front instead of wrapping [16:27] juju [16:47] * hazmat wanders of into django land === koolhead17|afk is now known as koolhead17 [16:56] hi all [18:15] does juju support running more than 1 service per vm instance? [18:25] zirpu, currently no [19:17] My google fu is failing me. Can someone point me to a good reference for the complete beginner looking to make a charm? [19:36] lazyPower: https://juju.ubuntu.com/CharmSchool that and just poking at the examples and current charms is how I got started [19:36] https://juju.ubuntu.com/Charms [19:45] lazyPower: Actually, this is a pretty good starting point: https://juju.ubuntu.com/docs/write-charm.html [19:45] Although a bit outdated in some pars === almaisan-away is now known as al-maisan [20:06] hazmat: thanks for the review, shall I merge the changes to trunk? [20:10] al-maisan, sure.. give me a moment, i need to adjust the permissions [20:11] shazzner: the python looks good for parse_upstream [20:13] al-maisan, you should be able to merge it now [20:14] * al-maisan tries [20:15] hazmat: it worked, thanks again! [20:16] al-maisan, thank you :-) [20:16] :-) [20:16] ttyl === al-maisan is now known as almaisan-away [20:34] marcoceppi: Thanks! exactly what I was looking for. === Guest92794 is now known as benonsoftware