File size: 7,658 Bytes
4aa5fce |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 |
=== Spads_ is now known as Spads [00:43] <redir> so we now require 3.5 GB ram for a controller? [01:00] <babbageclunk> hey redir, I get this error making a kvm on a trusty node: [01:01] <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 [01:02] <babbageclunk> 1.2.2-0ubuntu13.1.17 [01:03] <redir> hmmm [01:04] <babbageclunk> redir: xml file: http://paste.ubuntu.com/23670985/ [01:05] <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 [01:06] <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 [01:08] <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 [01:10] <babbageclunk> vmx turned on [01:10] <babbageclunk> I guess not suprising, since the other machine works. [01:11] <redir> is trusty the base os or on one of the vms in vmaas? [01:11] <redir> and what kind of cpu? [01:12] <babbageclunk> on the vm in vmaas [01:12] <babbageclunk> cpu - not sure [01:12] <babbageclunk> what's the best way to find out? [01:13] <redir> first look at virsh capabilities [01:13] <redir> $ virsh capabilities [01:14] <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 [01:15] <redir> but prolly the one that is failing [01:16] <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/ [01:17] <babbageclunk> This from the host machine: http://paste.ubuntu.com/23671048/ [01:18] <redir> ok [01:18] <babbageclunk> Hmm - that's weird - the failing one is v different. Maybe the host-passthrough didn't stick. [01:19] <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 [01:23] <babbageclunk> redir: reprovisioning the machine now. [01:32] <redir> sorry neighbors stopped by [01:32] <redir> let me look at others made no sense [01:33] <redir> I think that error is related to not having qemu-kvm installed or loaded [01:34] <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 [01:36] <babbageclunk> on the reprovisioned node virsh capabilities now says Haswell-noTSX [01:36] <babbageclunk> oh no, hang on - forgot the --series xenial. [01:37] <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 [01:38] <redir> I suppose I'll have to hit up frobware and jam when they are around [01:41] <babbageclunk> bums [01:41] <babbageclunk> sorry, I mean I forgot --series trusty. I think my brain's already gone on holiday. [01:42] <babbageclunk> rerereprovisioning now [01:46] <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> :| === deanman_ is now known as deanman [11:49] <deanman> Hey, anyone seen this before `message: 'hook failed: "cni-relation-changed" for flannel:cni'` on canonical-kube LXD deployment ? [12:23] <voidspace> jam: ping [12:28] <perrito666> voidspace: today is saturday for jam so you lost him until next year I presume [12:29] <voidspace> perrito666: yeah, I remembered after I pinged. Thanks. [12:40] <voidspace> perrito666: but hey, we have each other, so all is not lost ;-) [13:25] <jam> voidspace: perrito666: I don't pretend to be around, but I guess I can't hide either. [13:26] <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 [13:27] <voidspace> jam: anyway, happy holidays and see you on the other side [13:36] <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. [13:38] <perrito666> I see but my question was in the cultural side [13:39] <perrito666> Is it some form of holiday where you are? [13:43] * perrito666 is under the impression most cultures celebrate something around this time of the year [14:29] <perrito666> redir: your patch violates the 500 lines or less by about 30000 lines :p === deanman is now known as deanman_ [16:35] <redir> perrito666: in reality it is about 2000 lines maybe even less [16:36] <redir> perrito666: and the first two commits have been reviewed once. [16:44] <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 [16:45] <redir> perrito the whole thing is good:) [16:45] <redir> it needs it since it is over 500 lines [17:19] <redir> jam frobware voidspace just a bug in your ear https://github.com/juju/juju/pull/6748#issuecomment-268935865 [17:20] <redir> maybe it will register something [17:24] <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 [17:25] <redir> perrito666: aw shucks:) [17:26] <redir> and thanks for taking on the big PR, I appreciate it [17:27] <perrito666> np, as I said, interesting read [19:29] <voidspace> redir: I've only just twigged that you're reed... !! [19:30] <redir> reed-er [19:30] <redir> reed was taken... [19:32] <voidspace> redir: heh [19:32] <voidspace> redir: anyway o/ happy christmas :-) [19:33] <redir> you too voidspace and HNY! [19:37] <perrito666> I thought I was the only one left [19:41] <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 [19:42] <voidspace> perrito666: but I might not make it [19:55] <perrito666> k I am heading out early, happy holidays to all === petevg is now known as petevg_happyholi === petevg_happyholi is now known as petevg_holidays [20:21] <redir> later perrito666 Feliz Ano Nuevo [20:22] * redir steps AFK for a bit. |