| powersj | https://paste.ubuntu.com/24377231/ are the new errors we get | 00:02 |
|---|---|---|
| smoser | powersj, i dont understand those E0702 | 00:36 |
| smoser | powersj, https://github.com/PyCQA/pylint/issues/1419 | 02:39 |
| smoser | rharper, fun bond fix spinoff | 17:05 |
| smoser | https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1682871 | 17:05 |
| smoser | blackboxsw, are you actually doing the maas one ? | 17:35 |
| smoser | (bug 1677710) | 17:35 |
| smoser | thats great if you are. | 17:35 |
| * smoser hasn't had a maas around for a while. | 17:36 | |
| blackboxsw | smoser, yeah was going to give it a go | 17:39 |
| rharper | smoser: =( | 18:14 |
| rharper | I saw another thread about vlan_id=0 is somewhat undefined behavior | 18:14 |
| smoser | i didnt try anything other than xenial and yakkety | 18:15 |
| rharper | https://lwn.net/Articles/719297/ | 18:20 |
| rharper | 'In theory, VLAN 0 means "no VLAN". But the Linux kernel currently handles this differently depending on whether the VLAN module is loaded and whether a VLAN 0 interface was created. Sometimes the VLAN tag is stripped, sometimes not. | 18:20 |
| blackboxsw | smoser, per 1677710, I deployed a xenial (!proposed) machine on maas 2.20rc1. I see the following in /run/cloud-init/ds-identify.log: | 18:27 |
| blackboxsw | single entry in datasource_list (MAAS) use that. | 18:27 |
| blackboxsw | [up 4.70s] returning 0 | 18:27 |
| smoser | hm.. | 18:27 |
| blackboxsw | does this actually mean that ds-identify properly parsed the datasource? | 18:27 |
| smoser | no. it means the install only had that enabled | 18:28 |
| smoser | rharper, maybe you were missing that ? | 18:28 |
| smoser | rharper found this when doing some ubuntu core stuff | 18:28 |
| smoser | some versions of maas must pre-seed cloud-init with just MAAS | 18:28 |
| smoser | which, yeah, means that ds-identify shortcuts. | 18:28 |
| smoser | i had wondered why there wasnt much screaming over this | 18:29 |
| smoser | :) | 18:29 |
| rharper | what's going on ? | 18:36 |
| rharper | some versions of maas? how can it preseed the target? | 18:37 |
| smoser | through the cloud-inti preseed | 18:39 |
| rharper | I don't know what that is (ephemeral or deployment) ? | 18:39 |
| smoser | curtin passes the debian preseed through and dpkg-reconfigure | 18:39 |
| rharper | for *ubuntu* images | 18:39 |
| smoser | right. | 18:39 |
| rharper | deb-conf set selections work | 18:39 |
| smoser | and obviously that doesnt work for core images | 18:39 |
| rharper | the issue for the bug is for images that don't use debs | 18:39 |
| rharper | of course no one complained as no one is using core yet | 18:39 |
| smoser | well, yes. | 18:39 |
| smoser | but i didnt' realize it was was working still in ubuntu | 18:40 |
| rharper | you mentioned something that I missed ? | 18:40 |
| smoser | i didnt' think you understood it to be working in ubuntu either | 18:40 |
| rharper | ds-identify detecting maas in ubuntu images without the change to ds-identify to parse cloud.cfg.d/ ? | 18:41 |
| rharper | it entirely depends on where dpkg-reconfigure cloud-init wrote the cfg file | 18:41 |
| rharper | if it updates a file in /etc/cloud/*maas*.cfg; and had MAAS in the file, it would work; the issue was that it didn't search the cloud.cfg.d/*.cfg | 18:42 |
| smoser | hm.. | 18:45 |
| rharper | oh | 18:45 |
| rharper | no | 18:45 |
| smoser | oh. so it was busted. and blackboxsw verified it fixed. | 18:45 |
| rharper | they use pxe kernel cmdline, right ? | 18:45 |
| rharper | in addition to the cloud config | 18:45 |
| smoser | i didnt' realize blackboxsw above said proposed | 18:45 |
| rharper | they had two possible paths to tell it | 18:45 |
| smoser | for ephemeral they do cmdline | 18:45 |
| smoser | and it worked there. | 18:46 |
| rharper | right | 18:46 |
| smoser | but was busted in install images. | 18:46 |
| smoser | but i'm confused why there wasnt more screaming. | 18:46 |
| rharper | hrm | 18:46 |
| smoser | wasnt/isnt/ | 18:46 |
| smoser | oh. | 18:46 |
| smoser | i know | 18:46 |
| smoser | it was just yakkety that was busted | 18:46 |
| smoser | xenial is report only | 18:46 |
| rharper | yes | 18:46 |
| rharper | that's right | 18:46 |
| rharper | so it'd say it found nothing but then work anyhow | 18:46 |
| rharper | ie, not disable itself | 18:46 |
| rharper | that doesn't quite line up with saying it found MAAS though | 18:47 |
| rharper | I'd be happy to see /run/cloud-init/* and /etc/cloud/* | 18:47 |
| rharper | from that instance | 18:47 |
| rharper | then we can sort out what it did and did not find | 18:47 |
| smoser | ? | 18:49 |
| smoser | in above, it found a single datasource configured (MAAS) | 18:49 |
| smoser | which indicates it searched /etc/cloud.cfg.d/ | 18:49 |
| smoser | blackboxsw, you could pastebin /run/cloud-init/ds-identify.log | 18:49 |
| smoser | blackboxsw, ahasenack thanks. you guys have been a huge help | 18:53 |
| rharper | smoser: the newer code does | 18:54 |
| rharper | he said !proposed | 18:54 |
| rharper | so that'd be the previous one, right ? | 18:54 |
| rharper | which didn't search cloud.cfg.d | 18:54 |
| smoser | i dont know | 18:57 |
| smoser | i dont know how it could work. | 18:57 |
| smoser | i'd like to see more. | 18:57 |
| * smoser really wishes he had a maas again | 18:58 | |
| smoser | i do intend at some point to get beisner's work up joined with some updated "virtual-maas" stuff | 18:58 |
| smoser | so that we can easily deploy a maas on diglett | 18:58 |
| rharper | well, I know it didn't search the .cfg dir before; the code was pretty clear about what files it searched | 18:59 |
| rharper | what's not clear is what cloud-init was tested in the above report | 18:59 |
| rharper | report | 18:59 |
| rharper | supposedly not the proposed one | 18:59 |
| rharper | but we'll need the log and cloud dirs | 18:59 |
| smoser | hey. i'm out. have a nice weekend all. | 19:00 |
| rharper | ok | 19:00 |
| rharper | you too | 19:00 |
| blackboxsw | sorry smoser got pulled into a phone call. | 19:07 |
| blackboxsw | geting that now | 19:07 |
| blackboxsw | http://paste.ubuntu.com/24382246 | 19:08 |
| * blackboxsw catches up on backlog. So yeah I had a maas 2.2.0rc1 available and when I deploy xenial to bare metal I didn't see any warnings on login about missing datasources per 1677710. I didn't tweak anything in preseeds on this install as I wanted/hoped to see the original failure mode | 19:11 | |
| blackboxsw | this is commandline login BTW I don't have physical access to the hardware | 19:12 |
| blackboxsw | grabbing cloud-init/* and /etc/cloud/* | 19:15 |
| blackboxsw | rharper, for future reference on this one http://paste.ubuntu.com/24382309/ here are my /var/cloud-init/* files . I'll keep poking at this | 19:20 |
| blackboxsw | I had tested xenial (not xenial-proposed). ii cloud-init 0.7.9-48-g1c795b9-0ubuntu1~16.04.1 all Init scripts for cloud instances | 19:21 |
| rharper | blackboxsw: can you paste "find /etc/cloud" | 19:22 |
| rharper | /etc/cloud/cloud.cfg.d/90_dpkg.cfg set datasource_list: [ MAAS ] | 19:22 |
| blackboxsw | grabing that too. http://paste.ubuntu.com/24382329/ | 19:23 |
| rharper | there may be some issues here with report mode vs the strict handling | 19:24 |
| blackboxsw | http://paste.ubuntu.com/24382335/ | 19:25 |
| blackboxsw | ^ /etc/cloud/cloud.cfg.d/90_dpkg_maas.cfg with keys redacted | 19:25 |
| smoser | ok. that makes sense. | 19:25 |
| smoser | it always read that file | 19:25 |
| smoser | and maas seeded cloud-init with that so it wrote it. | 19:25 |
| smoser | /etc/cloud/cloud.cfg.d/90_dpkg.cfg | 19:25 |
| smoser | later | 19:25 |
| blackboxsw | cat /etc/cloud/cloud.cfg.d/90_dpkg.cfg | 19:26 |
| blackboxsw | # to update this file, run dpkg-reconfigure cloud-init | 19:26 |
| blackboxsw | datasource_list: [ MAAS ] | 19:26 |
| blackboxsw | have a good one smoser | 19:27 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!