[02:45] ok this inconsistency is driving me nuts, install maas using apt install maas, then log in using the CLI and create a dynamic reserved range and try the associated enabling of DHCP which fails inconsistently with the rack controller not being registered to the region controller on the same box! [02:46] is it just my day for weird issues :-/ [07:52] Bug #1668880 opened: maas partitions create size cannot be percentage === frankban|afk is now known as frankban === frankban is now known as frankban|afk [11:59] hi, how come maas can't find the virsh power type when i just installed libvirt-bin ? [12:02] Bug #1668957 opened: [2.1.3] maas API 2.0 - link_subnet does not work when linking subnet on a different fabric [12:06] should maas be able to use qemu://system as a power address? [12:07] hmm, it seems it insists on having an IP there [12:57] cnf, the maas instance is on the kvm host? [12:58] brendand: the kvm host is the maas host [12:58] so uhm, yes [12:58] brendand: i just need a place to run the juju controller [13:21] zeestrat: k, seems to work [13:21] i think :P === frankban|afk is now known as frankban [14:15] Hi guys! === Flint is now known as Guest54478 [14:19] Hey folks, any URL where I can have an informations related to "How to build custom images on MAAS" or "Explanation of a MAAS image" [14:19] ?? [14:25] Guest54478, Ubuntu and CentOS 6.6 and 7.0 images are included with MAAS for free, custom images require a support agreement [14:31] where can i find out why WoL powertype was removed from maas 2.x? i really miss it. [15:06] Bug #1437024 changed: Failure to PXE-boot from secondary NIC [15:41] Hi brendand, thanks for the reply, I know that Custom image require support agreement, but here, I'm looking for how the supported images are builded and what MAAS is expecting to find on the images like does it need cloud-init or another tool etc [16:31] can I have maas ignore cert failures for a VMWare power module? === frankban is now known as frankban|afk [18:44] when does maas add ssh keys to a deployed node? can I add a key in maas after and connect with it? [18:54] DesktopMan, when the node is deployed. no [18:54] at which point are they added? commision or deploy? [18:54] (not sure if deploy wipes what's there from comissioning...) [18:54] deployed [18:55] commissioning does not touch the node's disk [18:56] ah [18:56] thanks [18:57] the keys are added under the ubuntu user? or is there a user created for the maas user that deploys it [18:57] (if it's ubuntu) [18:58] in the web UI you are logged in via a user. that's the user whose keys will get copied when that user deploys [18:58] (each web UI user has keys imported into their account/profile) [18:59] but is the user also named after the maas user? or is it always ubuntu [18:59] oh, you mean when you ssh to the node? [18:59] yeah [18:59] yes, exactly, it is always to the 'ubuntu' account [19:00] great, thanks. should know enough to connect now :) [19:01] DesktopMan, it would be great if you could follow along here and open a docs bug if something is not clear: [19:02] https://docs.ubuntu.com/maas/2.1/en/installconfig-deploy-nodes [19:02] there is a link at the bottom of the left menu for doing that [19:04] DesktopMan, it's important to realize that this only works when you deploy directly from maas. i.e. deploying something via juju:maas will not have the key (in maas) copied. you need to do that via a juju method (i.e. add key to a juju "model") [19:07] thanks. I haven't quite figured how juju deployment differs yet [19:16] hmm dumped to a grub prompt after Fetching Netbook Image. that's not good [19:16] anyone have any suggesions as to why when installing maas using the maas package it doesn't always configure the rack controller? [19:21] Integrated NIC 1: Intel(R) 2P X520/2P I350 rNDC, any known issues pxe-booting from that? [19:29] stormmore, I'm very curious about that -- if it's the case then it's an ugly bug [19:29] stormmore, apt-get install maas doesn't leave rackd running? [19:30] kiko, it is running but not registered to the region controller [19:31] kiko, so when I try and implement dhcp through the cli I get "dhcp can only be turned on when a primary rack controller is set." [19:32] kiko, and it isn't consistent which is the weird part, one run of my automation fails and then a wipe and rerun might work or fail [19:32] stormmore, and you've set up the VLANs correctly? [19:32] when it fails, what do the logs say? [19:32] kiko, not using VLANs at the moment [19:32] I have long suspected there's a bug in the rack registration [19:32] MAAS 2.x? [19:32] yes MaaS 2.1 [19:33] not sure which log would be good to look at, can you advise me so then when it happens again I can look? [19:39] well, /var/log/maas/*log and the rackd/regiond log in particular would tell you if the rackd ever registered or if it's a registration failure happening early in the process [19:41] OK I am sure I will be reinstalling maas surely. just doing a bit of bash coding for a script to initialize an environment [19:43] also, note that MAAS is currently sensitive and bad about guessing what NIC you want MAAS to talk on [19:43] stormmore, so if your machines are multi-homed, make sure we're configuring it to do the right thing [19:44] kiko, the inconsistency suggests it isn't a system configuration problem [19:46] well, you could get interfaces unpredictably named in the old days, and maybe that is your case.. [19:46] i.e. pre-16.04/systemd [19:47] yeah I am using 16.04 lts [19:48] i'm having some trouble getting maas working on lxd. The problem seems to be that twisted can't bind and listen on localhost:69 [19:49] i followed the lxd-install page exactly [19:49] i'm running on xenial [19:53] nc is able to listen on the port. I suspect it's something else [19:58] last 2 installs have worked :-/ [20:00] cholcombe, do you have the right interface name configured? [20:01] cholcombe, what's the log entry you are seeing? [20:38] cholcombe: is it listing on localhost:69 or ::69? [20:40] because some services really really want ipv6 localhost for some reason. And if that isn't configured for your lxd, then that could give some trouble === beisner- is now known as beisner [21:06] kiko: i'll gist in a sec [21:06] kiko: https://gist.github.com/cholcombe973/376aa769285493ea81e95fbadcfeb25d === Siva is now known as Guest19683 [22:20] Bug #1668731 opened: Total pod resources not matching used pod resources [23:42] kiko so far today I only seen the "bug" once but of course I am still nowhere close to finishing what I am doing so I am sure I will see it again and will investigate further