[13:26] <pmatulis_> i have an alien cluster showing in the output of euca-describe-availability-zones, how to dump it?
[13:55] <superxgl> hi all , as in http://support.rightscale.com/12Guides/RightScale_Methodologies/Monitoring_System/Writing_custom_collectd_plugins/Custom_Collectd_Plug-ins_for_Linux
[13:56] <superxgl> it uses instance-id as the hostname, i don't know if it is can directly apply to eucalyptus ?
[14:00] <superxgl> is the instance-id binded with it's ip address ?
[15:04] <TeTeT> smoser: hi, do you have a few minutes to discuss the ebs root, lucid and Launchpad things?
[15:38] <smoser> TeTeT, sure. i'm just back now.
[15:39] <TeTeT> smoser: great. I've started writing up the exercise, will take the remainer of today to be finished and checking it tomorrow
[15:40] <smoser> there are lots of things we can do to make it more polished.
[15:40] <TeTeT> smoser: on the launchpad project, I'll happily get the infrastructure going. Should there be a special team driving it? Like server team or a newly founded one? Do you have any preference on project/team name?
[15:41] <TeTeT> smoser: I was thinking of calling it 'uec-ebs-root' or 'uec-persistency'
[15:41] <smoser> hm.. i kind of like both of those. i like staying away from "ebs-root" as its not really "ebs root".
[15:42] <TeTeT> smoser: ok, so let's call it 'uec-persistency', that would allow the project to foster other relevant pieces, if need be
[15:42] <smoser> sure...
[15:43] <smoser> ignore the ...
[15:43] <smoser> i dont knowwhy i typed that
[15:43] <TeTeT> smoser: I'll get it started asap and create a likely named team, making you, tom and me admins, so we can add people at will
[15:44] <smoser> great.
[15:44] <smoser> thanks, TeTeT
[15:46] <kim0> what's that uec-persistency thing ?
[15:47] <smoser> TeTeT, in the end, i really think we should have downloadables of a "loader" kernel and ramdisk .
[15:47] <smoser> then, the only thing hte user really needs to do is create the ebs volume.
[15:47] <smoser> kim0, https://code.launchpad.net/~smoser/+junk/kexec-loader
[15:48] <smoser> the readme there is probably not all that readable
[15:49] <TeTeT> kim0: I'm writing a UEC class exercise for it, if you want to proof and test it, let me know
[15:49] <smoser> but th eidea is that we use kexec as a loader. we supply a kernel and ramdisk.  that allows both ebs-root-like function *and* reboot into a new kernel on 10.04 +
[15:49] <smoser> reboot into new kernel function exists in 10.10 already
[15:49] <kim0> wow :)
[15:50] <kim0> sounds yummy
[15:50] <TeTeT> smoser: do you think it can be automated that much? Guess a script can do it
[15:51] <smoser> well, we just supply the kernel/ramdisk. there is only a need for 1.
[15:51] <smoser> well, 2.
[15:51] <kim0> we don't have kexec reboots on metal yet right
[15:51] <smoser> 1 kernel, 1 ramdisk for ebs-root-like, 1 ramdisk for non-ebs-root.
[15:52] <smoser> then, all we need is somethign that takes a ami and puts it onto an ebs volume (this is trivial for ubuntu images, and simple for anything that has grub2)
[15:52] <smoser> kim0, no ?
[15:52] <kim0> smoser: you tell me
[15:53] <kim0> just never heard of Ubuntu having it
[15:53] <kim0> if we do, why isn't it on by default
[15:53] <kim0> as in "reboot" does it ?
[15:53] <smoser> you've never done 'apt-get install kexec-tools'
[15:53] <TeTeT> TREllis: ^^^
[15:53] <smoser> yes, it "just works".
[15:53] <smoser> but this is different, actually, kim0. it doesn't use that.
[15:54] <kim0> smoser: you're saying, apt-get install kexec-tools && reboot
[15:54] <kim0> will kexec ?
[15:54] <smoser> the reason its different is that eucalyptus continually loads the initial kernel and ramdisk from outside the image.
[15:54] <smoser> every time.
[15:54] <smoser> and that kernel/ramdisk is just configured such that it mounts root, finds kernel and ramdisk, and kexec loads that.
[15:54] <smoser> kim0, you have to turn it on, (/etc/default/kexec) but other than that, yes.
[15:54] <kim0> awesome .. can't wait to reboot :)
[15:55] <kim0> is that eucalyptus's way of pvgrub
[15:55] <smoser> well, no, this is not from eucalyptus
[15:55] <smoser> and this is separate from what we *have* in 10.10
[15:55] <smoser> 10.10 solutoin is actually cleaner.
[15:56] <smoser> we load any multiboot image that the user provides
[16:29] <kim0> smoser: do you agree with the high level structure of the updated page
[16:29] <smoser> i think so , yeah
[16:29] <kim0> feel free to revert it :)
[16:29] <kim0> hehe
[17:52] <kim0> anyone tested KVM with spice on Ubuntu, does it work
[19:58] <jo-erlend> is there any plans to get SPICE-support in Ubuntu?
[21:52] <Kyle__> I'm trying to setup a small cluster, one head, three nodes, and keep having problems trying to get walrus going.
[21:53] <Kyle__> WHen I go to configuration, it simply says "failed to contact server".  According to /etc/init.d/eucalyptus-walrus and /etc/init.d/eucalyptus-walrus-publication, everything is running.
[22:15] <Kyle__> Anyone?  Simple UEC install (chose defaults) on closed network, walrus not showing up, can't download images.
[22:15] <Kyle__> (successfully download images)