=== Spads_ is now known as Spads | ||
redir | so we now require 3.5 GB ram for a controller? | 00:43 |
---|---|---|
babbageclunk | hey redir, I get this error making a kvm on a trusty node: | 01:00 |
redir | :( | 01:01 |
babbageclunk | 2016-12-23 00:52:59 DEBUG juju.container.kvm run.go:21 output: error: Failed to define domain from /var/lib/juju/containers/juju-862251-1-kvm-3/juju-862251-1-kvm-3.xml | 01:01 |
babbageclunk | error: unknown OS type hvm | 01:01 |
babbageclunk | What should I send you? | 01:01 |
redir | what version of libvirt? | 01:01 |
babbageclunk | probs that xml file. | 01:01 |
redir | libvirt-bin | 01:01 |
redir | sure | 01:01 |
babbageclunk | um, on the trusty machine? | 01:01 |
redir | yes | 01:01 |
babbageclunk | 1.2.2-0ubuntu13.1.17 | 01:02 |
redir | hmmm | 01:03 |
babbageclunk | redir: xml file: http://paste.ubuntu.com/23670985/ | 01:04 |
redir | So I think that happens if kvm kernel mods aren't loaded o | 01:05 |
redir | which is installed by qemu-kvm I believe | 01:05 |
babbageclunk | I can check whether that's installed on the machine | 01:05 |
redir | k | 01:05 |
babbageclunk | redir: 2.0.0+dfsg-2ubuntu1.30 | 01:06 |
redir | also if it is kvm in kvm check if host-passthrough is really set | 01:06 |
babbageclunk | I checked that | 01:06 |
babbageclunk | it is | 01:06 |
redir | what's lsmod | grep kvm say? | 01:08 |
babbageclunk | ha, I was just copying that from a serverfault article! | 01:08 |
babbageclunk | kvm_intel 143187 0 | 01:08 |
babbageclunk | kvm 455843 1 kvm_intel | 01:08 |
redir | google rules | 01:08 |
babbageclunk | vmx turned on | 01:10 |
babbageclunk | I guess not suprising, since the other machine works. | 01:10 |
redir | is trusty the base os or on one of the vms in vmaas? | 01:11 |
redir | and what kind of cpu? | 01:11 |
babbageclunk | on the vm in vmaas | 01:12 |
babbageclunk | cpu - not sure | 01:12 |
babbageclunk | what's the best way to find out? | 01:12 |
redir | first look at virsh capabilities | 01:13 |
redir | $ virsh capabilities | 01:13 |
redir | can you send or pastebin that? | 01:14 |
babbageclunk | It says SandyBridge - sure, pastebinning. | 01:14 |
babbageclunk | oh - do you want that run on the vmaas node, or the host? | 01:14 |
redir | hmmm | 01:14 |
redir | maybe both | 01:14 |
redir | but prolly the one that is failing | 01:15 |
babbageclunk | Ok, this is from the vmaas node that's failing: http://paste.ubuntu.com/23671042/ | 01:16 |
babbageclunk | redir: this is from the node that works: http://paste.ubuntu.com/23671045/ | 01:16 |
babbageclunk | This from the host machine: http://paste.ubuntu.com/23671048/ | 01:17 |
redir | ok | 01:18 |
babbageclunk | Hmm - that's weird - the failing one is v different. Maybe the host-passthrough didn't stick. | 01:18 |
babbageclunk | I'll shut the machine down and set it more firmly | 01:19 |
redir | can you log into the failing node and try manually installing qemu-kvm and let me look at the others | 01:19 |
babbageclunk | redir: reprovisioning the machine now. | 01:23 |
redir | sorry neighbors stopped by | 01:32 |
redir | let me look at others made no sense | 01:32 |
redir | I think that error is related to not having qemu-kvm installed or loaded | 01:33 |
redir | which would make sense if pass-through didn't work or apt-get failed to install qemu-kvm for some reason | 01:34 |
redir | I dont' know what happens if the package manager in juju fails to install a package | 01:34 |
babbageclunk | on the reprovisioned node virsh capabilities now says Haswell-noTSX | 01:36 |
babbageclunk | oh no, hang on - forgot the --series xenial. | 01:36 |
redir | whelp merging in the stuff from develop seems to break things anyhow | 01:37 |
redir | I'm was a bit behind | 01:37 |
redir | had to resolve a couple conflicts but it seems that there's some issue with bridge and spaces | 01:37 |
redir | I suppose I'll have to hit up frobware and jam when they are around | 01:38 |
babbageclunk | bums | 01:41 |
babbageclunk | sorry, I mean I forgot --series trusty. I think my brain's already gone on holiday. | 01:41 |
babbageclunk | rerereprovisioning now | 01:42 |
redir | yeah, they are failing with juju.provisioner failed to prepare container "1/kvm/0" network config: unable to find host bridge for spaces [] for container "1/kvm/0" | 01:46 |
redir | so it only worked until I merged in develop | 01:46 |
redir | :) | 01:46 |
redir | :| | 01:46 |
=== deanman_ is now known as deanman | ||
deanman | Hey, anyone seen this before `message: 'hook failed: "cni-relation-changed" for flannel:cni'` on canonical-kube LXD deployment ? | 11:49 |
voidspace | jam: ping | 12:23 |
perrito666 | voidspace: today is saturday for jam so you lost him until next year I presume | 12:28 |
voidspace | perrito666: yeah, I remembered after I pinged. Thanks. | 12:29 |
voidspace | perrito666: but hey, we have each other, so all is not lost ;-) | 12:40 |
jam | voidspace: perrito666: I don't pretend to be around, but I guess I can't hide either. | 13:25 |
voidspace | jam: none of us can hide, not for long anyway... | 13:26 |
voidspace | jam: :-) | 13:26 |
voidspace | jam: anyway, I "made a decision" and am just seeing if it works - you can comment on the review when it gets there | 13:26 |
voidspace | jam: anyway, happy holidays and see you on the other side | 13:27 |
perrito666 | Jam is this an actual holiday for you are just dragged by the shutdown? | 13:36 |
jam | voidspace: thanks, you too | 13:36 |
jam | perrito666: We have some stuff we're trying to get done for 2.1, so I'm half-around trying to shepherd it through. | 13:36 |
perrito666 | I see but my question was in the cultural side | 13:38 |
perrito666 | Is it some form of holiday where you are? | 13:39 |
* perrito666 is under the impression most cultures celebrate something around this time of the year | 13:43 | |
perrito666 | redir: your patch violates the 500 lines or less by about 30000 lines :p | 14:29 |
=== deanman is now known as deanman_ | ||
redir | perrito666: in reality it is about 2000 lines maybe even less | 16:35 |
redir | perrito666: and the first two commits have been reviewed once. | 16:36 |
perrito666 | I am reviewing the whole thing redir | 16:44 |
perrito666 | so far like a lot what I see | 16:44 |
perrito666 | point me to the relevant commit if you want | 16:44 |
perrito666 | brb lunch | 16:44 |
redir | perrito the whole thing is good:) | 16:45 |
redir | it needs it since it is over 500 lines | 16:45 |
redir | jam frobware voidspace just a bug in your ear https://github.com/juju/juju/pull/6748#issuecomment-268935865 | 17:19 |
redir | maybe it will register something | 17:20 |
perrito666 | redir: I sent several changes but apart from those LGTM | 17:24 |
perrito666 | also, your code is usually so clean that it makes me all warm and fuzzy inside | 17:24 |
redir | perrito666: aw shucks:) | 17:25 |
redir | and thanks for taking on the big PR, I appreciate it | 17:26 |
perrito666 | np, as I said, interesting read | 17:27 |
voidspace | redir: I've only just twigged that you're reed... !! | 19:29 |
redir | reed-er | 19:30 |
redir | reed was taken... | 19:30 |
voidspace | redir: heh | 19:32 |
voidspace | redir: anyway o/ happy christmas :-) | 19:32 |
redir | you too voidspace and HNY! | 19:33 |
perrito666 | I thought I was the only one left | 19:37 |
voidspace | perrito666: I think I've verified my fix works, but I haven't written a test and I need to scour the code to check for unintentional semantic changes | 19:41 |
voidspace | perrito666: it would be nice to get a PR tonight, it's not 8pm yet | 19:41 |
voidspace | perrito666: but I might not make it | 19:42 |
perrito666 | k I am heading out early, happy holidays to all | 19:55 |
=== petevg is now known as petevg_happyholi | ||
=== petevg_happyholi is now known as petevg_holidays | ||
redir | later perrito666 Feliz Ano Nuevo | 20:21 |
* redir steps AFK for a bit. | 20:22 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!