| blackboxsw | rharper: https://github.com/cloud-init/ubuntu-sru/pull/50 | 07:40 |
|---|---|---|
| rharper | thx | 07:40 |
| andrein | hello everyone, i have the following network_data.json: https://www.irccloud.com/pastebin/fYPE11ed/ | 13:20 |
| andrein | the problem is, cloud-init doesn't configure my bond+vlan on first boot. however, using `cloud-init devel net-convert -p /mnt/config/openstack/latest/network_data.json -k network_data.json -d / -D centos -O sysconfig` the network configuration is applied correctly | 13:21 |
| andrein | the one thing that stands out in cloud-init.log (https://seashells.io/p/pS8KbbuB) is `Applying network configuration from fallback bringup=True: {'version': 1, 'config': [{'subnets': [{'type': 'dhcp'}], 'type': 'physical', 'name': 'enp130s0f0', 'mac_address': '0c:c4:7a:1e:63:8e'}]}` | 13:22 |
| smoser | andrein: cloud-init is not finding your config drive. | 14:58 |
| smoser | it is falling back to datasourcenone | 14:58 |
| smoser | not sure why, but some configuration is telling it that its datasource_list is just (i think) None | 15:00 |
| smoser | possibly ds-identify did that... notsure. | 15:00 |
| smoser | you can post a 'cloud-init collect-logs' tarball somewhere it will contain a lot more info | 15:00 |
| andrein | smoser: it's complaining that it can't read `/run/cloud-init/result.json`. looks like a broken symlink | 15:14 |
| andrein | i've ran cloud-init clean / clound-init init multiple times here, maybe i did something wrong at some point? | 15:16 |
| andrein | is cloud-init supposed to read it's config drive directly from /mnt/config? how can it not find my config drive, it's right there :) | 15:17 |
| smoser | andrein: can you just touch /run/cloud-init/result.json and then rujn it ? | 15:21 |
| smoser | its supposed to be a symlink into /var/lib/cloud i think. but not much matter here. | 15:22 |
| smoser | i'd like to see what that collect-logs says. | 15:22 |
| andrein | I've noticed cloud-init status says it's still running, although I don't see any processes matching cloud | 15:24 |
| smoser | if you want, you can let me in and i can poke a round a bit. if thats possible. | 15:25 |
| smoser | where did you get this image? | 15:25 |
| andrein | that's going to be complicated, unfortunately. but we can arrange a skype call or something and I can show you around if that's ok with you | 15:26 |
| andrein | kolla-ansible built it for me usind dib | 15:26 |
| smoser | well, try to get collect-logs to run | 15:28 |
| andrein | i just got it running, gotta configure networking and i'll send it off | 15:29 |
| smoser | or, in liue of that, i guess if you collect | 15:29 |
| smoser | /var/lib/cloud/ | 15:29 |
| smoser | /run/cloud-init | 15:29 |
| smoser | /etc/cloud | 15:30 |
| smoser | and | 15:30 |
| smoser | systemctl list-units | grep cloud | 15:30 |
| andrein | is there something like paste.openstack.org for tarballs? | 15:31 |
| smoser | you can base64 encode | 15:33 |
| smoser | andrein: make sense ? | 15:34 |
| andrein | smoser: sent you the link in private | 15:35 |
| smooth_mess | is bugs where i submit a feature request? | 20:11 |
| smooth_mess | there is a small chance that (in aws) attaching a volume to an instance will cause cloud-int "fs_setup" to fail; so a missing feature would be that there is no "wait" or "retry" functionality. So, to get around it, adding to 'bootcmd' the following: counter=0; while [ ! -b /dev/xvdf ]; do counter=$((counter+1)); printf "%s not found, tried %ds out of 300\n" "/dev/xvdf" $counter; if [ $counter -ge 300 ]; then printf "Did not find | 20:12 |
| smooth_mess | %s" "/dev/xvdf"; exit 1; else sleep 1; fi; done | 20:12 |
| smoser | smooth_mess: yes, submit a bug | 20:29 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!