[01:13] conjure-up 2 and juju 2 are still no luck [01:14] http://paste.ubuntu.com/23171693/ [06:50] Hello, I have asked question in maas devel mailing list about ability to have software raid solution with non ubuntu images, and Christian recommended to ask it here. Well, I'm looking for a way, how to tell maas to use software raid devices as TARGET, instead of simple disk. It's available with ubuntu by default, but it's disabled for other distros. builtin: [curtin, block-meta, simple] - as I understood, I should somehow pass 'custom' to that place, instea === frankban|afk is now known as frankban [07:42] roaksoax: because I have 2 nic , each one on a different subnet. The subnet used to pxe boot and the 'public' from which hosts are reachable from the rest of the network. My assumption is that I only need a default GW on the public network. === Sagar is now known as Guest75192 [08:12] Hi All, I am trying to install Ubuntu 12.04 image on a node and my MAAS version is 2.0 , but after around 40 minutes deployment fails giving timeout error message [08:12] nothing appears in the logs [08:13] did someone try Ubuntu 12.04 image deployment? [08:13] any ideas or thoughts on how I can troubleshoot? [08:14] My experiance says maas 2.0 is in very active development state, so for me this is running very unstable. [08:15] I've installed ubuntu 16.04 lts with maas 2.0 and try to deploy node with ubuntu 16.04 - same problem like you, just failed to deploy [08:16] After that i've switched to 14.04 and everyting was fine. I mean 14.04 maas node, and deploy nodes only the same version of ubuntu. [08:16] maas is 1.6 i think, default with the 14.04 lts ubuntu from repositories [10:17] Bug #1622971 opened: controllers do not load on nodes page [13:05] Bug #1623027 opened: [2.1] MAAS should not allow user to select releases that are EOL [13:11] Bug #1623027 changed: [2.1] MAAS should not allow user to select releases that are EOL [13:20] Bug #1623027 opened: [2.1] MAAS should not allow user to select releases that are EOL [13:44] roaksoax : Hi. As you said yesterday, I am trying to setup maas on KVM machines. Facing issue while setting initially maas slave nodes with PXE boot. Issue details : http://paste.openstack.org/show/573609/. I request you please provide me the solution for this. I tried to solve this. But Still I am getting the same issue. [13:51] rock, it's unclear for me -- are your KVM nodes PXE-booting successfully? [13:54] Kiko: No. While creating KVM machine[maas slave node] only I have choosen PXE boot . But It was showing "Network selection does not Support PXE". [13:55] rock, I think you need to fix that first [13:56] rock, what do you mean by "slave nodes"? [13:56] Kiko: mass nodes. [13:56] rock, VMs which you want MAAS to install Ubuntu on? [13:57] rock: that's a strange error, seems related to libvirt/kvm [13:57] rock: so your VM's are telling you that they cannot PXE boot ? [13:57] i wonder if this is with the VM/libvirt configuration itself [13:57] roaksoax: Yes. [13:58] rock: how does your libvirt network config look like ? mine looks like: http://pastebin.ubuntu.com/23173664/ [14:01] roaksoax: Mine also similar to your config. http://paste.openstack.org/show/573617/ [14:03] strange [14:05] roaksoax: Is there any other way to do. Actually I don't have enough machines to test my JUJU bundle. So I have to test that on [MAAS+OpenStack base bundle] setup. [14:07] rock: what if you try to get your config look more like mine ? [14:08] rock: honestly i'm quite surprised that libvirt would tell you that when the network seems to be configured correctly [14:08] rock: that config above I showed you is proven to work [14:14] roaksoax: Hi. Exactly I did as shown here. http://paste.openstack.org/show/573619/. [14:15] rock: do you have a screenshot of the VM config or what the VM tells you that cannot PXE boot ? [14:16] roaksoax: Yes. How can I send screeshot through chat. please tell me [14:18] rock: https://imagebin.ca/ [14:30] roaksoax: Sorry for the late. https://imagebin.ca/v/2uwUL9uVDXEY. [14:31] roaksoax: https://imagebin.ca/v/2uwUpRacX3D9 [14:36] rock: i think that should be ok because libvirt is not providing PXE, hence it is hoswing you the message [14:36] rock: what I dont understanding is the bug that libvirt gives you [14:36] rock: what if you create a new one with selecting it to be PXE [14:38] roaksoax: So Is there a bit other way of doing this?. [14:42] roaksoax: (or) What are the other possible ways to do MAAS setup on Virtual machines. If we have please provide me detailed info for that. [14:43] rock: no, you will have to create your own VM's for MAAS to manage them [14:44] need to step out for a bit [14:49] roaksoax: OK. Thank you. MAAS can manage VBox VMs? [14:50] roaksoax: (or) MAAS can Manage VMware VMs? [14:51] If it can manages VBox VMs (or) VMWare VMs , What power types it will genarally take? [14:54] roaksoax: SO How can we configure network for those setups. Is there any specific docs for that. If you have any idea please provide me the Info. === frankban is now known as frankban|afk [16:14] Bug #1623110 opened: Networks page doesn't load fully on yakkety [18:28] Bug #1518414 opened: NVMe model info/serial number is not seen [18:51] Congrats MAAS team on the 2.0 stable release! It’s been a fun process to follow along. [19:05] hi roaksoax [19:05] are you around? [19:06] http://paste.ubuntu.com/23171693/ i am getting this error on the latest conjure-up === CyberJacob is now known as Guest72045 === mwhudson is now known as Guest72628 === spammy is now known as Guest67489 [19:51] sup peeps - couple of questions about maas setup [19:51] hey baldpope [19:51] nice netsplit [19:52] in the quick-install documentation from ubuntu, the documentation implies that booting a new host will have that box show up in the list of nodes [19:52] but I've had to manually add nodes, is that normal? [19:52] sup kiko [19:52] yes, it will if auto-enlistment succeeds [19:52] so then I should assume auto-enlist is failing? [19:52] yes, it is [19:53] what would cause auto-enlist to fail? [19:53] well [19:53] the setup instructions are pretty simple - I don't think I missed anything [19:53] baldpope, does the host PXE-boot correctly from MAAS? [19:53] yes [19:53] well - best I can tell [19:53] and what happens after it boots? [19:54] ubuntu appears to be running via iscsi back to the maas host? [19:54] yep [19:54] that's very much correct [19:54] it's an ephemeral environment [19:54] read-only iscsi [19:54] okay [19:54] and then? [19:54] does not appear to be an 'and then' just sitting at a login prompt [19:54] maas still shows commissioning [19:54] hang on [19:54] commissioning? [19:55] yea [19:55] that doesn't make sense [19:55] the first time you pxe-boot a node [19:55] apologies - let me start over [19:55] :) [19:55] if I do nothing (after setting up maas) [19:55] and I boot a new host - call it blade1 [19:56] if I do nothing more than power on, it boots via pxe [19:56] goes through what appears to be a pxe boot, with an iscsi root (similar to nfs root) and then eventually will get to a traditional login prompt [19:56] if I review maas - maas never upates the list of nodes [19:57] so then if I manually add a node (same blade) - it will reboot the blade and run through the same steps - but with 'commissioning' in the status [19:57] baldpope, it is stuck at the login prompt forever? [19:57] never shut downs? [19:57] err shuts down? [19:57] okay [19:57] so yes [19:57] something in the ephemeral environment is getting stuck [19:57] well - admittedly, not watching it like a hawk - possible I missed a reboot [19:57] but yea - appers to just sit there [19:57] it won't reboot [19:58] but it will shut down if it was successful [19:58] ah [19:58] if it's not, it's not successful [19:58] can you ssh in with ubuntu:ubuntu? [19:58] no, I don't think it's powered down [19:58] cool [19:58] 1sec [19:58] roaksoax told me that's supposed to work but I've never seen it in action [19:59] cannot ssh in - publickey only auth [19:59] and stupid me I didn't define a public key at the time [19:59] i do have a public key defined now - but was not at the time of initial power on / deploy [20:03] hmmm [20:03] side note - supermicro ipmi/ikvm not working for me very well - so I lose console access as soon as the frame buffer kicks in - no more SoL [20:03] baldpope, are you talking about enlistment and commissioning [20:04] during enlistment I think we do enable passwords [20:04] well .. i would assume enlistment is a stage 1 and commissioning a stage 2?.. if that's the case then stage1 never occurs and stage 2 is reporting 'failed commissioning' after some number of minutes [20:05] that's assuming enlistment is automatic (based on pxe boot of a clean / new host) [20:09] and at the moment, it appears that the host is just sitting at a console login screen [20:12] hey kilo, is it okay to install maas 2.0, and conjure-up 2.0, juju 2.0 on the same box? [20:18] kiko, we lose you bud? [20:30] when booting up -there is a config element that shows cloud-config-url=http://blah [20:30] but the IP listed is NOT the 'internal' IP where the new hosts sit [20:30] 1sec [20:36] .. [20:51] baldpope, am otp sorry :) [20:54] s'all good [20:55] just ran dpkg-reconfigure maas-cluster-controller and changed IP to internal [20:55] booting a node - see if that fixes it [20:59] looking better - it auto enlist in nodes now [20:59] it might be obvious - but someone might wantto clarify when running the setup to mark the controller IP as the internal address [21:01] wow - it even added maas as a user in the ipmi [21:01] slick [21:03] mpw we [21:03] now we're talking! === Guest72628 is now known as mwhudson [21:04] baldpope, funny that you got caught on that one, it's such a common problem that I didn't think of asking [21:04] caught it by chance watching ilom - and found another user ran into similar problem [21:04] my mistake [21:05] ok, running commission now [21:05] btw - for reference - following http://www.ubuntu.com/download/cloud [21:10] Bug #1623192 opened: Cannot create DHCP snippet [21:16] kiko, thanks for being a sounding board - looking good [21:16] first node is showing ready now [21:16] running through the remaining 4 [21:17] the default node names are interesting [21:17] 'cooked-women' [21:17] that can't be good... [21:17] ;) [21:18] baldpope, it was a very controversial feature implementation [21:18] baldpope, and tbh some people have encountered rather offensive names [21:18] i can imagine [21:18] but.. I kind of like them :-) [21:18] i can see the humor [21:19] but as a request - should make an option in the cluster/server config for a base name [21:19] so that all nodes come prefixed? [21:19] you know you can tweak everything through the API, right? [21:20] i.e. the equivalent of a for n in get_nodes(): n.rename("us-east-"+ n.name) [21:20] was unaware [21:21] everything the UI can do (and a bit more) can be driven via the API [21:21] is that the maas tool? [21:23] the maas tool is a (somewhat clunky) API client [21:23] there are likely several bits I need to read up ... like how to configure the the nics as bonded [21:23] yep, see maas.io/docs [21:23] cool - ty [21:24] heh - impractical clocks [21:24] very... [21:26] okay, I need to split for tonight [21:26] will see you tomorrow? [21:26] * kiko waves out [21:26] later kiko [21:26] likely hit the channel up some more later/tomorrow === CyberJacob is now known as Guest45771 === menn0 is now known as menn0-exercise [21:47] ugh.. [21:47] * baldpope sigh [21:49] had to kill openstack-install - and now I cannot run it again, openstack-install -u fails... [21:49] how to manually clean up openstack-install? === menn0-exercise is now known as menn0 === mup_ is now known as mup === Guest67489 is now known as spammy [23:34] Bug #1623266 opened: [web UI] Proxy is for APT only