=== dendrobates is now known as dendro-afk [02:25] heh i didnt know about this channel [08:20] Hi, I created an AMI using ec2-bundle-vol with -s 1500. Now my instance has only 1.5G of disk space. How do I avoid this? [08:22] Is that maybe realted to the fact that I used a micro instance for that? === dendro-afk is now known as dendrobates === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates [22:08] kirkland: who should I nag when I have questions about UEC/eucalyptus design whenever I start looking into providing HA? [22:08] RoAkSoAx: Daviey, mostly [22:08] RoAkSoAx: this is the best channel for that [22:09] RoAkSoAx: ttx, me, smoser, mathiaz can all help, to some extent [22:09] RoAkSoAx: but Daviey owns UEC's maintenance at this point [22:11] kirkland: ok cool. Right now I need to know more about design of the controllers itself, such as where are databases stored, what's the data that needs to be replicated, where it is located [22:11] all that sort of stuff [22:14] kirkland: though If you can point me out to documentation that specifies those things, that's also helpful :) [22:26] RoAkSoAx: hmm, those are hard questions. [22:26] RoAkSoAx: most of the operational stuff is at: [22:26] http://help.ubuntu.com/community/UEC [22:26] RoAkSoAx: i wrote much of that myself [22:26] RoAkSoAx: for the internals of Eucalyptus, you can try #eucalyptus [22:27] RoAkSoAx: though that channel isn't terribly accurate [22:29] kirkland: I've already been through much of the documentation http://help.ubuntu.com/community/UEC. Indeed they are hard questions. However, to be able to provide HA all those sort of stuff are considered, because that means that we will have to have two severs with the same data at all times [22:29] so if one fails, the otherone will continue to work as if it was the same node as before [22:29] and so on [22:31] kirkland: for example, what would someone need to do to replace a CLC? COpy all the data from the CLC to another one, as well as keys? [23:11] RoAkSoAx: interesting questions; I think most of the critical data is in /var/lib/eucalyptus [23:12] kirkland: what I'm just gonna do is replicate all the data under /var/lib/eucalyptus and start testing [23:12] to see what happens [23:13] to see if the second node loads all that data and acts as if it was the original node [23:13] RoAkSoAx: i bet that gets you 90% of the way there [23:14] kirkland: yeah probably, though there are also things to consider such as VirtualIP's both for the public and private network [23:14] and if so, how will that affect key's and certs creations [23:14] and stuff [23:15] no idea [23:16] kirkland: are the CC/SC/Walrus/NC aware of the CLC by IP? [23:17] RoAkSoAx: yes [23:22] kirkland: so that'd mean that each CLC node will have its own IP, but they will share a Virtual IP, so that every other application in the network will recognize it with that VIP