=== frankban|afk is now known as frankban [08:45] question: I don't quite understand what configuring fabrics and subnets in maas do [08:45] does it alter interfaces and routing tables? [10:21] Bug #1690781 opened: Compose machine failure: Start tag expected, '<' not found, line 1, column 1 [10:45] Bug #1690784 opened: An ip address from a reserved range is used for auto-assigning [11:12] DesktopMan: you configure networking for machines in maas [12:57] Bug #1690810 opened: Failed to probe and enlist VMware nodes === dannf` is now known as dannf [13:24] Bug #1690784 changed: An ip address from a reserved range is used for auto-assigning [13:27] Bug #1690810 changed: Failed to probe and enlist VMware nodes === gimmic_ is now known as gimmic [15:05] Hey guys, what is this new MaaS' feature, called Pods? [15:06] Bug #1649501 opened: No rack controllers can access the BMC of node [15:07] Can I deploy a bare metal server with MaaS, turn it into a KVM server by installing "apt install ubuntu-virt-server", then, use the Pods feature to manage its VMs from MaaS? [15:09] ThiagoCMC: yes [15:09] Wow! [15:09] ThiagoCMC: the only caveat is that you need the vm's to PXE boot from maas itself [15:09] That's awesome! [15:09] Hmmm... [15:09] Right, so, at the bare metal, I need to reconfigure the /etc/network/interfaces and add a Linux Bridge there? [15:10] So the VMs can PXE boot on the same segment as its bare metal... [15:10] Oh, wait... [15:10] ThiagoCMC: we added virsh as a pod for testing and development. So if you have a KVM/libvirt running out there, as long as the networking can pxe boot of maas, it will jsut work [15:10] ThiagoCMC: not necessarily, but at a minimum yes [15:10] ThiagoCMC: you could have, say connect the vm's on a different vlan where you have another rack controller running [15:11] ThiagoCMC: but yes, at the very minimum, your vm's need to be able to pxe boot off maas for this to work [15:11] But if the VMs PXE boot, MaaS will "think" that it is just another bare metal server, isn't it? [15:11] ThiagoCMC: yes, but it will a vm' inside the pod [15:11] or a machine inside the pod [15:12] Right, so, before PXE booting the VM, I need to first, install KVM/Libvirt and "Add pod" on MaaS, is that correct? [15:12] ThiagoCMC: honestly, I have not tested deploying a machine with maas, then adding it as a pod. but i dont see why would that not work [15:12] ThiagoCMC: that's correct. say you have a kvm/libvirt, where you configure a bridge in the same vlan as MAAS' pxe. You can add a pod in MAAS [15:13] ThiagoCMC: and you will be able to create VM's and such [15:13] :-D [15:13] Thank you! I'll try that now... [15:25] roaksoax, what the "virsh address" syntax looks like? [15:25] I mean... How it looks like (sorry about my English). [15:49] I'm trying to "Add pod" using a syntax like this: "qemu+ssh://root@hail.cloud.example.com/system" but, the following error appear: "Failed talking to pod: The command was not found or was not executable: virsh" [15:50] but virsh is there, user "ubuntu", not "root", it works... [15:51] ThiagoCMC, user 'ubuntu' works but not 'root'? [15:52] (the user needs to be in group 'libvirtd') [15:52] both ubuntu and root works lolcally, but not via "add pod" [15:53] Mayeb this: "qemu+ssh://ubuntu@hot-alpapa.maas/system" is not a correct syntax. [15:54] From MaaS, I can "ssh ubuntu@hot-alpapa.maas" and run virsh. [15:54] But "add pod" doesn't work... [15:55] And there is no Pods documentation here: https://docs.ubuntu.com/maas/devel/en/ [15:56] ThiagoCMC, yeah, use 2.2 branch [15:56] I'm on it! [15:57] maas 2.2.0~rc4+bzr6048-0ubuntu1~16.04.1 [15:57] ThiagoCMC, i mean the 2.2 docs [15:57] Bug #1690848 opened: [VLAN details] In the Subnets on this VLAN table the (Unassigned) space is a link that links to the device discovery [15:57] devel is behind 2.2 for the next little while [15:57] Where is the "Add pod / virsh" at the 2.2 doc? [15:58] I can olny find this: https://docs.ubuntu.com/maas/2.2/en/intel-rsd [15:58] Pod but not via virsh. [15:58] right. docs need to be added for virsh Pods. i opened a doc bug a little while ago [15:59] ThiagoCMC, maybe you want to track it and add any notes you may have - https://github.com/CanonicalLtd/maas-docs/issues/430 [16:00] Nice... I'll, soon as I figure it out... =P [16:03] Bug #1690848 changed: [VLAN details] In the Subnets on this VLAN table the (Unassigned) space is a link that links to the device discovery [16:09] Bug #1690848 opened: [VLAN details] In the Subnets on this VLAN table the (Unassigned) space is a link that links to the device discovery === frankban is now known as frankban|afk [17:54] Bug #1690878 opened: Unreliable PXE-booting of Quanta S910-X31E [17:57] Bug #1690878 changed: Unreliable PXE-booting of Quanta S910-X31E [18:03] Bug #1690878 opened: Unreliable PXE-booting of Quanta S910-X31E [18:03] Bug #1690882 opened: Send hostname of Pod to driver to allow better log messages [18:03] Bug #1690883 opened: Drivers cannot post notifications [18:06] Bug #1690882 changed: Send hostname of Pod to driver to allow better log messages [18:06] Bug #1690883 changed: Drivers cannot post notifications [18:15] Bug #1690882 opened: Send hostname of Pod to driver to allow better log messages [18:15] Bug #1690883 opened: Drivers cannot post notifications [18:17] pmatulis, commented there on maas-docs issue 430, not working yet. [18:19] ThiagoCMC, alright. i'll try to get to this soon [18:20] Thank you! [18:27] Bug #1690884 opened: [2.2] apparmor_parser: enable to replace "/usr/sbin/dhcpd" [spurious error during dist-upgrade in container?] [18:57] Bug #1690884 changed: [MAAS 2.2] apparmor_parser: enable to replace "/usr/sbin/dhcpd" [spurious error during dist-upgrade in container?] [19:03] Bug #1690884 opened: [MAAS 2.2] apparmor_parser: enable to replace "/usr/sbin/dhcpd" [spurious error during dist-upgrade in container?] [19:06] Bug #1690884 changed: [MAAS 2.2] apparmor_parser: enable to replace "/usr/sbin/dhcpd" [spurious error during dist-upgrade in container?] [19:20] clear [19:25] Hello, i have an issue with maas. I cannot ssh to some nodes deployed with maas [19:36] atm_it: Hi there. for the machines to be 'deployed', they must have obtained keys from MAAS allowing you to SSH into them [19:37] atm_it: that said, are your machines in 'deployed' state ? [19:42] yes, they are in deployed state and i've uploaded the keys to maas [19:43] i noticed i can ssh to juju and autopilot but not the openstack nodes [19:44] This was the error message i got from Landscape when trying to deploy openstack "juju ended with exit code 1 (out='', err='Adding contents of "/var/lib/landscape/juju-homes/3/ssh/juju_id_rsa.pub" to authorized-keys Creating Juju controller "3" on landscape-maas Loading image metadata Looking for packaged Juju agent version 2.1.2 for amd64 Starting new instance for initial controller Launching controller instance(s) on landscape-maas... [19:45] atm_it: could be that it caused the keys not to be imported [19:45] atm_it: ssh ubuntu@ wont allow you to ssh in ? [19:45] it doesnt allow me to ssh in [19:45] strange [19:46] atm_it: check /var/log/maas/rsyslog/machine-name/date/messages and see if there show any error before importing ? [19:46] okay [19:46] this is the error i get when i try to ssh [19:46] ubuntu@autopilot:~$ ssh ubuntu@node1 Permission denied (publickey). [19:52] atm_it: my guessis that something is failing preventing cloud-init from injecting the keys [19:52] can you pastebin the requested log ? [19:54] okay, will do that [20:18] here's the log [20:18] https://pastebin.com/B3Hq5dtK [20:20] atm_it: in that log it doesn't seem as if the machine contacted maas and run metadata after install [20:24] let me send the log of another deployed node which i cant ssh into [20:27] atm_it: so, what I think the issue is that cloud-init is failing before it actually imports the ssh keys for maas [20:29] is there a way to determine if cloud-init completed successfully? [20:48] i'll redeploy the affected nodes and see if it solves the issue [20:49] also, can you show me what to look for in the logs, for a successful deployment? [20:49] thanks [21:25] Bug #1690919 opened: Lots of "maasserver.websockets.base.HandlerDoesNotExistError" in regiond.log [21:30] Guys, MaaS is creating a RAID1 on top of /dev/sda1 and /dev/sdb! Why? [21:30] Why not sda1 + sdb1 ? [21:30] MaaS did: "md0 : active raid1 sda1[0] sdb[1]" =/ [21:31] It is syncing and I'll ignore this for now... [21:34] Bug #1690919 changed: Lots of "maasserver.websockets.base.HandlerDoesNotExistError" in regiond.log [21:37] Bug #1690919 opened: Lots of "maasserver.websockets.base.HandlerDoesNotExistError" in regiond.log