[00:04] Hey there. I added a custom image in Maas for CoreOS, using the openstack.img file from coreos. The file is listed, I can select the image when deploying. But I see in the PXE boot that it's still the Ubuntu image that is pushed. [00:04] I have the same problem with a custom image for Debian 9 [00:06] I uploaded the custom images using just the "maas boot-resources create" command. I don't have a custom repo. [00:08] once the installation failed (from maas point of view), in the next boot it shows the following error [00:08] Loading custom/amd64/ga-16.04/coreos/no-such-image/boot-kernel... failed: No such file or directory [00:08] notice there is a "ga-16.04" in the path which I think is not where it should be [00:08] I use Maas 2.2.2 (6099-g8751f91-0ubuntu1~16.04.1) [00:09] mimizone_: maas will use ubuntu to install images of other os' [00:09] mimizone_: that said, maas doesn't really support coreos nor debian [00:09] haaa.... [00:09] so i ca't guarantee that would work [00:13] other unrelated question. After doing multiple commissioning on a node when trying different hardware configuration of my storage chassis attached to it, the commissioning gets stuck in the block devices check, getting a HTTP BAD request on the metadata server. [00:13] the only solution I found has been to delete the node and re-add/enlist it. [00:14] is that an expected behavior? [00:14] mimizone_: definitely not [00:15] it would be worth trying to figure out what triggers it [00:16] I'll try to see. not easy to log what's happening on the IPMI console of my old OCP hardware [00:17] should I run maas in debug mode? how do I do that? [00:22] mimizone_: you should have logs in /var/log/maas/rsyslog/machine-date/etc/etc [00:22] machine-name* [00:25] sure. but it doesn't go into the details of the BAD REQUEST [00:25] for example "cloud-init[2773]: request to http://172.30.72.130:5240/MAAS/metadata//2012-03-01/ failed. sleeping 1.: HTTP Error 400: BAD REQUEST" === frankban|afk is now known as frankban [07:51] Bug # opened: 1650554, 1650575, 1650580, 1659613 [09:01] Bug #1709849 opened: juju bridge creation with MAAS provider ruins static routes on host [13:37] Bug #1642298 opened: UEFI Xenial install sets computer to boot from hard disk [15:32] roaksoax: i will be in hamburg on saturday. so we have to setup a session for debugging and looking at vcenter 6.5 on sunday somehow. if you still remember [15:33] roaksoax: and i hope that resolv.conf problem will be solved soon. i need my maas funktionality back [15:44] hello comunity [15:44] any documentation about How to custom ubuntu xenial maas image ephemeral-v3 [15:44] ?? [15:44] thanks [15:52] dont you need a subscription for maas in order to customize images? [15:52] ybaumy: the resolv.conf is being looked at. Not a maas bug thought, It is a bug in ubuntu it seems [15:53] alian: maas does not publish any documentation to customize ubuntu images. I do believe there's documentation on how to create your own ubuntu images in the internet, which should work with MAAS [15:53] but we dont support nor recommend image customization [15:53] roaksoax: have you a estimate on how long this will take to fix? [15:54] alian: what maas supports, however, is to allow you to pass user_data for cloud-init to customize your installation [15:54] alian: which is the same thing you would do in any other cloud [15:54] ybaumy: idk TBH, this is not being worked under my team but is under someone else's which are currently looking at it [15:55] roaksoax: ok. how about the other thing. will there be somebody there on sunday evening CEST in order to debug the memory problem? [15:56] roaksoax: i created already an account for vpn access at my lab [15:57] roaksoax: but i want to happen that under my supervison (if that is the correct term) [15:59] ybaumy: i need to circle back on that one. We have some other discussions wrt that [16:00] roaksoax: you know its a chance to clear some things up in maas vmware support. i just want to say that. [16:00] roaksoax: just drop me a message if you or a member of the team will look at it [16:02] roaksoax: one thing that also needs to be looked at IMO is eg. the power status. which is not always correct [16:02] roaksoax: i havent reported that yet but it exists [16:02] can help me find the doc reference to create my own ephemeral-v3 image to import on maas? [16:05] roaksoax: we sure can postpone it. i just want you to know im there for you guys if you need a setup [16:06] roaksoax: its a win win situation for both of us i guess === frankban is now known as frankban|afk [17:03] ybaumy: it is indeed, but we have other priorities we are working through that we havent been able to clear. Although, this is still on the list [19:11] i have commissions that keep timing out during storage collect [19:12] how do i debug that via resuce? what am i looking for? [19:12] rescue [19:17] I have a problem with the first PXE boot with our new setup - is anyone about who might be able to lend a hand? (sorry for my lack of terminology) I get as far as PXE boot happening, but none of the maas related bits seem to succeed. it's booting up to a login and not being discovered. [19:33] only clue i have so far is errors about reaching the region controller. getting 404s back. [19:33] sorry. not 404. 400. bad request. [19:34] and the metadata url ends with 2012-03-01 [19:34] is that a date? [19:34] time on the node is correct... [19:50] crud i bet its because of this tiny FlexFlash drive [19:51] based on bug 1541030 [20:07] hello how to build/edit ubuntu xenial ephemeral-v3 image to deploy using metal as a service? thanks for any help [20:29] how do i disable a devicde [20:30] i want mass to ignore during commission? [20:30] maas [20:40] xygnal, did you sort it? tick the device on the nodes page, from the "take action" list choose "release" [20:47] huh? [20:47] its not listed. [20:47] sorry, I misunderstood. [22:37] Bug #1714106 opened: Allow more user permission granularity so a user cannot reboot a machine