smoser | powersj: https://jenkins.ubuntu.com/server/view/cloud-init,%20curtin,%20streams/job/cloud-init-integration-lxd-c/ | 15:21 |
---|---|---|
smoser | how often / how does that run ? | 15:22 |
smoser | it looks like it just stopped running on the 20th | 15:22 |
powersj | lxd is daily | 15:22 |
powersj | hmmm | 15:22 |
powersj | hmm xenial is running daily, but not set to kick the bionic test | 15:24 |
powersj | doh: set to kick project: cloud-init-integration-lxd-a | 15:24 |
smoser | hm.. | 15:25 |
powersj | fixed | 15:26 |
[42] | can you point me somewhere for getting started creating a debian template image to be used with qemu (proxmox)? | 16:15 |
blackboxsw | smoser, time to reset cloud-init status meeting date. it doesn't look like we had one since 07/02 | 16:33 |
blackboxsw | and it looks like someone kicked off a meeting start a couple days ago by accident | 16:34 |
blackboxsw | https://ubottu.com/meetingology/logs/cloud-init/2018/cloud-init.2018-07-23-16.05.log.txt | 16:34 |
blackboxsw | #endmeeting | 16:34 |
meetingology | Meeting ended Mon Jul 30 16:34:15 2018 UTC. | 16:34 |
meetingology | Minutes: http://ubottu.com/meetingology/logs/cloud-init/2018/cloud-init.2018-07-23-16.05.moin.txt | 16:34 |
blackboxsw | smoser: think we should host the status meeting next monday? | 16:35 |
smoser | blackboxsw: that sfine with me... | 16:41 |
smoser | (long delayed) | 16:41 |
blackboxsw | lotta vacation | 16:43 |
=== blackboxsw changed the topic of #cloud-init to: Reviews: http://bit.ly/ci-reviews | Meeting minutes: https://goo.gl/mrHdaj | Next status meeting: Monday 8/6 16:00 UTC | cloud-init 18.3 released (06/20/2018) | ||
TJ- | I'm trying to diagnose why a vagrant ubuntu/bionic64 box doesn't appear to be getting its network configured. As I understand it cloud-init should be doing that on first boot. Are there docs that'll help me figure it out? | 17:33 |
blackboxsw | TJ http://cloudinit.readthedocs.io/en/latest/topics/network-config.html should be an overview of what cloud-init does | 17:34 |
blackboxsw | generally speaking if you are on cloud-init version 17.2 or greater you can run "cloud-init status --long" on the commandline if you can access the serial console of the box to see if there are errors | 17:34 |
blackboxsw | otherwise look in /var/log/cloud-init.log for Tracebacks | 17:35 |
blackboxsw | on bionic you should have 18.2 or later cloud-init so should be ok on the version | 17:35 |
TJ- | Thanks; it seems there's no "network:" defined. I presume that is something to do with the vagrant box | 17:38 |
TJ- | no log whatsoever; systemctl status shows all cloud-* services dead | 17:38 |
blackboxsw | hrm, I would've expected "systemctl list-dependencies" to list all cloud-init jobs/units in green status unless dependency chains were broken by some other units/services that are running in the environment | 18:17 |
blackboxsw | root@myb1:~# systemctl list-dependencies | grep cloud | 18:17 |
blackboxsw | ● ├─cloud-init.target | 18:17 |
blackboxsw | ● │ ├─cloud-config.service | 18:17 |
blackboxsw | ● │ ├─cloud-final.service | 18:17 |
blackboxsw | ● │ ├─cloud-init-local.service | 18:17 |
blackboxsw | ● │ └─cloud-init.service | 18:17 |
blackboxsw | if dependency chains were broken due to ordering cycles, you'd see "Breaking ordering cycle" from "journalctl -b" which may tell you if cloud-init was not run or disabled by the booting system | 18:19 |
blackboxsw | sudo cloud-init collect-logs will collect a tarfile of cloud-init logs that you could post to me if you'd like me to peek at what may be the symptom. but generally "cloud-init status --long" of look in /var/log/cloud-init.log for any Tracebacks | 18:20 |
TJ- | I'm not sure that vagrant relies on cloud-init - lack of authoritative docs though. I've just read an overview says vagrant box /should/ be configured to use DHCP. My situation is even the vagrant management interface isn't UP, let alone running dhclient to get an IP address. | 18:25 |
blackboxsw | so artful/bionic should have switched from using /etc/network/interfaces to /etc/netplan. I'd expect you'd see /etc/netplan/50-cloud-init.yaml if cloud-init was in charge of your networking. | 18:29 |
blackboxsw | if your vagrant image contains network configuration in /etc/network/interfaces(.d) it's likely misconfigured | 18:30 |
blackboxsw | netplan *should* be relying on networkd to do all your dhclient work from artful/bionic or later | 18:31 |
TJ- | right, no netplan config, nothing under /run/systemd/network/ or whatever | 18:35 |
TJ- | This is the ubuntu/bionic64 vagrant box , so very weird | 18:36 |
TJ- | OK, I've solved it! The problem is vagrant-mutate. The Ubuntu boxes are for VirtualBox and have 2 vmdk disk images, 1 of which is cloudimage-configdrive. vagrant-mutate only copies over the root-fs so there's no config data for cloud-init | 19:09 |
TJ- | looks like vagrant-mutate is no longer maintained by its author either. I'll have to fork and see if I can get a handle on the code | 19:10 |
TJ- | Thanks for the pointers; helped me reduce the problem space considerably | 19:11 |
arthurl | hi guys- I'm trying to pass a cloud-init.yml to packer to build a custom ami in aws ec2- the ami is being built but i see none of the cloud-init actually gets run... | 21:27 |
blackboxsw | thanks for the write back TJ | 22:10 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!