=== ejat_ is now known as ejat === frankban|afk is now known as frankban [09:03] Bug #1762344 opened: JS tests random failure with "test caused a page reload" error [15:40] Bug #1762461 opened: [2.4] MAAS becomes unresponsive when there's RPC errors [15:46] hi guys, I am new to maas. Is there a way that I can change the default DNS record "maas(default)" to a custom record? [16:02] is there a way to exclude subnets from the Dashboard? [16:03] Also.. why do I not see Observed Nodes in the Dashboard? [16:24] Hey_: you can disable active discovery from the dashboard, but only if active discovery is enabled [16:24] srihas: yes, change the domain name over the api [16:25] roaksoax: "over the api" means? :) [16:27] srihas: you can update the domain name over the api [16:28] roaksoax: But I have 3 subnets.. I only want to disable it for 2 subnets [16:28] roaksoax: as GUI talks over API. it can update as well, right? [16:31] Hey_: you can do that under each subnet if active discovery is enabled. passive discover only listens for arp broadcasts [16:31] srihas: the GUI doesn't support changing the domain name [16:31] srihas: you have to do it via the API (or the CLI, which is just a client to the API) [16:32] roaksoax: aha, I will look for commands, thank you [16:32] one more thing, I cannot delete "fabric-1" , its showing like "Can't delete fabric; the following interfaces are still connected: eth0 (unknown) on " [16:33] can it be forced from CLI as well? [16:34] roaksoax: so.. under subnet, I should Enable Managedallocation and disable active mapping? [16:35] at the moment.. I have them bot disabled for the subnets I don't want to see in Dashboard [16:35] Managed Allocation, Active Mapping / Disabled [16:37] Hey_: you can't disable passive discover per subnet [16:37] Hey_: becuase that listens to arps across all the network [16:37] Hey_: you can only disable "active" mapping, which either does ping or nmap on your subnet [16:37] roaksoax: ohh.. understood. [16:38] Hey_: so you could enable active mapping in the settings page, but disable it on the subnets you want disabled [16:38] Hey_: IIRC, that should not do "passive" mapping [16:38] roaksoax: under the subnets..I see observed nodes.. but I don't see some of those nodes in the dashboard [16:39] Hey_: there's an open bug for that, but those obversed nodes are nodes that dhcpd' from maas [16:39] I want to image those === frankban is now known as frankban|afk [16:50] roaksoax, do I need to create a comissioning script for my custom win10 image? [16:50] The partition scheme is default [17:08] Hey_: windows images are only support if created by canonical [17:09] roaksoax, I was able to create an image successfully. [17:09] ohh.. your saying you won't discuss it. [17:09] understood. [21:27] Bug #1762555 opened: ipmi-config connection timeout [21:33] Bug #1762555 changed: ipmi-config connection timeout [21:35] Two MAAS 2.3 questions: [21:36] 1) Once deployed, can the machines be disassociated from MAAS? (i.e., not try to cloud-init contact the MAAS server) [21:39] Bug #1762555 opened: ipmi-config connection timeout [21:40] wdennis: sure, you should be able to uninstall cloud-init from the deployed machine [21:41] wdennis: a deployed system, AFAIK,is not re-running anything in cloud-init on boot [21:50] 2) there's just no way of spec-ing a different root disk partition? [21:52] roaksoax: It seems that cloud-init drops a .cfg file into /etc/network/interfaces.d that will have the wrong IP for my target deploy network... Have to re-write/erase it [21:55] wdennis: for ubuntu there is [21:55] wdennis: ? cloud-nit should be writing correct networking based on what maas is configured [21:57] It does write correct Ip for the network I do OS installs on, which is not the same as the target deployment network, which is elsewhere [21:58] Maybe MAAS is not the correct tool to do deploys for machines that will not stay connected to MAAS? [22:00] Also, it seems that the machine does run cloud-init each time it boots, but now where it's deployed, all the messages say "handlers.py[WARNING]: failed posting event: {...etc...}" [22:06] wdennis: cloud-init, in maas or everwhere else, will run on first boot [22:07] wdennis: that said cloud-init is configured to report back to maas messages [22:07] wdennis: that's what you see there [22:07] wdennis: after a mcahine is 'deployed' [22:07] you can uninstall cloud-init [22:07] wdennis: as far as the networking, i'm not quite sure what you mean nor have enough context of what you are doing [22:30] roaksoax: I'm deploying machines' OS on a lab network, then will post-configure them with Ansible on that same network, and then disconnecting them from that network, and deploying them on the target network which has no access to the lab (MAAS) network [22:31] I also now see that it configures Ubuntu to get updates from the (now-inaccessible) MAAS server [22:33] So, it seems like there's almost a requirement (perhaps assumption) that the delpoyed machine will stay in contact with the MAAS rack server [22:35] wdennis: ok, so you disconnect the machine from the maas network and then you basically reconfigure networking ? [22:36] yes [22:36] wdennis: ok, and is this with centos [22:36] ? [22:36] No, Ubuntu [22:36] wdennis: ok, and if you reconfigure the network, and you reboot the machine, it gets reconfigured back to how it ws deployed by MAAS and your changes are lost ? [22:37] wdennis: on the updates part. MAAS configures an APT proxy. You *can* disbale the use of the proxy or you can change the proxy to a different one [22:38] No, I was not expecting (due to MAAS n00b status) that MAAS was dropping an interface config via cloud-init (which I'm also a n00b on) into /etc/network/interfaces.d, and also trying to post back cloud-init results to MAAS [22:39] So I have to re-write networking config (differently from what I was doing) and also 'apt-get purge cloud-*' to remove all the cloud-init stuff [22:40] SO I think I get what needs to be done there [22:41] Also need to remove the apt proxy [22:41] Now I'm wondering what else? [22:42] I'm just wondering if I'm fighting against the design of MAAS (i.e. if MAAS assumes continued connectivity/mgmt of the nodes it's deployed) [22:43] Was previously using Cobbler to deploy nodes, but it's not really that Ubuntu-friendly, and development seems to be on life support... [22:44] Cobblem didn't seem to assume ongoing connecitivity/mgmt of the nodes it deployed, though [22:44] well, MAAS is a cloud-like system that has a metadata server [22:44] so maas works very similarly to how instances in the cloud work via cloud-init [22:45] wdennis: that said, it has sane defaults, like the proxy case (you can disable it in the Settings page) [22:46] Yup, I see... Cloud instances have an external controller by needs [22:46] wdennis: right, the one thing, however, is that once the machine is deployed (after it boots for the first time onto the installed os) [22:46] cloud-init just basically does nothing with MAAS [22:47] OK, except try to report back status, right? [22:48] wdennis: yeah, but on if it failes to report status back, after it deployed, it doens't really matter since cloud-init wont really fial if it cannot report status back [22:49] wdennis: you can see those messages in the machine event log (e.g. Machine details page > 'Events' tab [23:17] Bug #1762575 opened: bmc-config fails to set maas user password on Cisco IMC during commission on MAAS 1.9.5 [23:53] Bug #1762579 opened: HA region controllers offer different NTP servers