[03:02] <crazed> ensemble eh? sounds pretty cool
[03:02] <crazed> i want to see it in use though
[03:15] <flaccid> tried chef or puppet before?
[03:42] <crazed> yeah i use chef mostly currently
[03:44] <crazed> chef cloudformation and cloud-init, but you need to combine three things to get the end goal, seems ensemble is trying to do all three together
[03:50] <flaccid> well cloud-init is a perfect example
[03:50] <flaccid> that relies on upstart and yet its not even a dependency of the package
[03:51] <flaccid> it won't work on most other distributions out of box. i still need to find time to port it.
[03:52] <flaccid> i guess if you are happy with being locked into ubuntu with everything, then its certainly the distro you can be happy with
[03:58] <crazed> amazon linux adopted cloud-init, i think others would as well, there's nothing else out there that provides a nice set of tools to interact with userdata
[03:59] <flaccid> yes they had to fork it and do their own version of it which really doesn't help. i need to find term to merge both so that more than 2 distros can use it.
[04:00] <flaccid> term=time
[04:00] <flaccid> its disheartening to see two distros act like proprietary companies
[04:01] <crazed> would it really be that hard to port to say centos?
[04:03] <crazed> https://bugs.launchpad.net/cloud-init/+bug/655837
[04:03] <flaccid> no i did an analysis and its not hard to port, but of course it takes a little bit of time. this doesn't address the concept of a platform independent codebase
[04:03] <crazed> oh you're subscribed to that
[04:03] <flaccid> yes thats the bug we have
[04:03] <flaccid> nobody has time
[04:04] <flaccid> good ol working backwards. amazon forked it because it depends on upstart..
[04:05] <smoser> amazon forked because it depends on apt
[04:05] <smoser> that was the biggest issue
[04:06] <smoser> the upstart isn't really all that bad.
[04:06] <smoser> and they didn't really chose to fork. they're interested in getting their changes back upstream
[04:06] <smoser> business forced them to fork (ie, they needed something *now*)
[04:06] <crazed> maybe there should be a concept similar to chef, where it abstracts the init/package systems
[04:07] <smoser> well, yes. some abstraction layer for packages would be necessary.
[04:08] <flaccid> i analyzed the code and upstart is more hard in there imho. there is no reason to depend on either.
[04:08] <flaccid> smoser: if they are interested, why have they not just done it?
[04:09] <smoser> things take time.
[04:09] <flaccid> true, but how long for this one? :)
[04:10] <flaccid> i want to do it so i can port it properly to debian and conform to deb policy, but heh i dont have time yet
[04:10] <crazed> time's always an issue
[04:10] <smoser> and priority
[04:10] <smoser> you have something that works, its hard to assign priority to doing the right thing
[04:10] <smoser> :)
[04:10] <smoser> anyway...
[04:11] <crazed> true
[04:11] <flaccid> it doesn't work on debian heh
[04:11] <smoser> i am interested in making cloud-init function properly on debian
[04:11] <flaccid> otherwise i would of included it in the AMIs i release last week
[04:11] <smoser> and would be interested in helping people make it work on fedora or the like.
[04:11] <flaccid> yep, its just time and priority for both of us right and we've talked about this before
[04:12] <flaccid> thanks again for the rightscale support though :) me and jeremydei are lovin it
[04:12] <smoser> so you're using it?
[04:12] <smoser> thats great.
[04:12] <flaccid> oh just in some natty testing stuff
[04:12] <smoser> it could be better.
[04:12] <smoser> you're kindof delegated to user-scripts with it.
[04:12] <flaccid> i have to find time to rewrite the rightlink_installer.sh for the hook
[04:12] <smoser> rather than being able to insert boothooks and such
[04:13] <flaccid> well just the one remote script to install rightlink is all we need :)
[04:23] <flaccid> smoser: tried out my images yet by any chance?
[04:24] <smoser> no.
[04:24] <flaccid> ah wells
[18:57] <koolhead17> hi all