[04:38] Bug #1644229 changed: MAAS 2.1.1 - Curtin - Failed to deploy CentOS7 [08:53] Hi everyone, I'm trying to enlist and commission a virtual machine but it is failing with the error '{"power_parameters": ["Power ID: This field is required."], "power_type": ["Select a valid choice. virsh is not one of the available choices."]}' [08:56] here is the regiond log http://paste.ubuntu.com/24169761/ [08:57] any idea what might be the issue here, I'm facing this issue randomly in a pipeline where configuration of MAAS is automated. [11:04] junaidali: how are you trying to add it ? === degville_ is now known as degville [11:50] Why can't I find the boot-images command on the MaaS cli ? [11:50] am I missing something ? [11:51] myra, have you been emailing maas-devel about this? [11:52] roaksoax: I'm using this command [11:52] maas "${MAASCLI}" machines create autodetect_nodegroup='yes' hostname="${JUJU_POWER_ID}" tags='bootstrap' power_type='virsh' mac_addresses="${JUJU_MAC}" power_parameters_power_address='qemu+ssh://'${HOST_USER}'@'${HOST_IP}'/system' architecture='amd64/generic' power_parameters_power_id="${JUJU_POWER_ID}" [11:54] junaidali, maybe we need the proper packages installed to use the virsh driver [11:54] e.g. libvirt-bin, virsh [11:56] brendand: restarting the regiond fixed the issue for me. [11:56] but I've seen the issue hitting multiple times [11:59] it seems like there is a sync issue between regiond and rackd. Is there any way (a CLI command etc) that we can check if both are synced properly or not? [11:59] junaidali: it is because the rack controller is not connected [12:00] junaidali: and it cannot validate [12:01] roaksoax: is there a way to check if it is connected? In MAAS UI under controller, the status was green [12:02] but the logs were showing connection error. [12:03] junaidali: did you check the detailed cntroller page ? [12:03] and there is, check the rack controller endpoint [12:03] iirc [12:06] roaksoax: cant find iirc under services in controller page http://imgur.com/a/tZDXZ [12:08] I've installed MAAS on a VM with 8GB RAM and 4 vcpus, do you think the issue might be due to the specs? [12:10] One specific thing about my environment is, I'm changing MAAS IP to the IP on eth1 interface [12:11] sudo maas-region local_config_set --maas-url http://${MAAS_IP}:5240/MAAS [12:11] and then updating rackd [12:12] sudo service maas-regiond restart; sudo sed -i "s/localhost/${MAAS_IP}/g" /etc/maas/rackd.conf; sudo service maas-rackd restart [12:24] Bug #1672353 opened: [2.2b3] Cannot see IP address of machine while in Hardware Testing [12:24] Bug #1672354 opened: [2.2b3] Cannot see IP address of machine while in Hardware Testing [12:33] Bug #1672353 changed: [2.2b3] Cannot see IP address of machine while in Hardware Testing [12:33] Bug #1672354 changed: [2.2b3] Cannot see IP address of machine while in Hardware Testing [12:45] Bug #1672353 opened: [2.2b3] Cannot see IP address of machine while in Hardware Testing [12:45] Bug #1672354 opened: [2.2b3] Cannot see IP address of machine while in Hardware Testing [12:45] Bug #1672363 opened: [2.2b3] 'deploy' API endpoint doesn't automatically allocate a machine [13:15] Bug #1672370 opened: [2.2b3] Hardware testing doesn't actively sends output to MAAS [13:20] junaidali: or maas-rack config --region-url && sudo maas-rackd restart [13:50] thanks roaksoax I will change that. [15:06] Bug #1672414 opened: Custom storage on Centos [17:04] kiko, did you see my updates to the Bug #1669548? [17:25] <_flx> hi. I have a MaaS 1.9.4 setup on ubuntu 14.04, unmanaged cluster. The default nameserver on deployed instances is the maas ip. Is there any way to tell maas to use the DNS provided by the router? [17:28] <_flx> https://bugs.launchpad.net/maas/+bug/1581155 [17:28] <_flx> i see [18:07] /wi/win 5 [18:24] _flx, you can set MaaS' upstream dns [18:25] _flx, that way it will act as a caching nameserver as well as a resolver [18:27] kiko, you around? [18:35] <_flx> stormmore, do I need to commission/deploy again? [18:36] _flx, shouldn't need to since it is just changing what MaaS will do with DNS requests, not the nodes themselves [18:38] <_flx> ok. testing... [18:47] Bug #1667141 changed: [FFe] Standing FFe for MAAS 2.2 [18:47] <_flx> stormmore, doesn't work. can't resolve hostname [18:48] <_flx> the default nameserver on the node is the maas server ip [18:48] <_flx> and i can't resolve the hostname [18:49] <_flx> the documentation of that feature says "Only used when MAAS is running its own DNS server." [20:13] Bug #1671320 opened: MaaS should set /etc/environment with appropriate proxy settings for rescue mode [20:15] _flx: /win 6 [20:15] err [21:12] hmmm I alway use MaaS as it's own DNS server so that it can update it with the hosts [21:13] anyone familiar with how tag definitions worth? what would it look like to search the hostname for a substring? [21:44] stormmore, re tags, i'm curious whether you read the docs and found them lacking [21:51] stormmore: tags definition to search the hostname for a substring? tag definitions only work for lshw output that maas collects and to autotag commmon machines [21:51] stormmore: https://docs.ubuntu.com/maas/2.1/en/manage-cli-tags [21:51] for example there, you can tag all machines that have a CUDA GPU for the example [21:52] tags dont apply to machine hostnames [21:54] pmatulis, yeah I always read the docs before asking questions ;) [21:55] I was really looking for a way to auto tag nodes that get the hostname worker* as worker, decided to manually tag them instead for now