=== freeflyi1g is now known as freeflying === kentb-out is now known as kentb [23:02] roaksoax: did you see https://code.launchpad.net/~allenap/maas/remove-ipmi-optimisation/+merge/160146 [23:18] bigjools: yeah... did you see this: http://pastebin.ubuntu.com/5594094/ [23:18] bigjools: i just upgraded a machine from quantal t raring and seeing that error [23:18] looking [23:19] roaksoax: is that with the latest power fix? [23:19] bigjools: i believe so yes [23:19] they tested it in the lab [23:19] and it was ok [23:20] so - confused [23:20] i'm gonna revert it and see what happens [23:20] where are you running that? [23:21] bigjools: this is virsh btw, not ipmi [23:21] in a lab [23:22] bigjools: maybe this is the issue: http://bazaar.launchpad.net/~maas-maintainers/maas/trunk/revision/1456 [23:22] roaksoax: why is ipmi set as the power type for a VM? [23:22] bigjools: is not ipmi, it is virsh [23:23] roaksoax: that revno is not a problem, it is doing the right thing [23:23] so you have a failure in the virsh script [23:23] it needs to be debugged [23:23] bigjools: yeah the virsh script hasn't really changed, so it shouldn't fail [23:23] sorry - I didn't realise you were talking about virsh all of a sudden :) [23:23] hehe [23:23] roaksoax: it was probably failing before and we didn't know [23:23] yeah this is our virtual-maas [23:23] now we know [23:23] roaksoax: could just be exit status on script [23:24] bigjools: i'm thinking that it might be not having access to virsh [23:24] that's what I'm thinking [23:24] the same code works with ipmi and friends [23:24] we never tested virsh very much as it's not "metal" :) [23:24] and now you're using it a lot more [23:25] bigjools: ok apparently is virsh issue [23:25] phew [23:25] i mean not having permission for virsh [23:25] but that's strange since there's a sudoers file for it [23:26] roaksoax: we could do with more info printed in that error [23:26] indeed [23:29] bigjools: http://paste.ubuntu.com/5594116/ --> tha obviously doesn't fail - this does: http://pastebin.ubuntu.com/5594122/ [23:29] bigjools: so probably permissions [23:30] roaksoax: agreed [23:34] i'll investigate tomorrow im off now [23:35] roaksoax: cheers [23:35] bigjools: good morning [23:35] wotcha mwhudson [23:36] bigjools: i installed maas on calxeda the other day [23:36] bigjools: mostly it worked fine but i got a bit confused about a few things [23:37] understandable :) [23:37] bigjools: the biggest one was that juju defaults to asking for an amd64 node for bootstrap [23:37] and it took me a long time to figure out this was why it wasn't getting an instance [23:38] that's not very nice if it, why should it care I wonder [23:38] of it* [23:38] which juju? [23:38] py [23:38] weird then [23:38] does the Go version work on ARM? :) [23:39] i ended up adding logging statements to the maas api server to print out the arguments... [23:39] which was a bit sad-making [23:39] bigjools: i don't think so [23:39] :/ [23:40] bigjools: the other thing i wanted to bug you about was that ages ago you said that maas would soon gain the ability to splat a tarball onto disk rather than running debian-installer [23:40] bigjools: did that happen? [23:40] mwhudson: it's WIP. roaksoax and smoser are working on the installer, the MAAS changes are done though. [23:41] ok [23:41] is there a blueprint i can subscribe to or anything? [23:43] hmmm not on my list