otubo | vijayendra: doesn't openstack have a dhcp service? IIRC, when an instance deployed with cloud-inti is on a ipv4 network, the link local interface comes up with its addr 169.254.0.0/16 to get metadata from OpenStack. | 10:38 |
---|---|---|
otubo | vijayendra: perhaps I don't know exactly how PowerVS works here, does it uses the full implementation of OpenStack? | 10:39 |
vijayendra | otubo, PowerVC/VS doesn't use complete openstack services, It doesn't use DHCP service provided by openstack. VM's are statically configured using ConfigDrive | 11:08 |
otubo | vijayendra: would you have your network yaml configuration to share? I'm implementing the support for cloud-init set some more network options on sysconfig. | 11:09 |
vijayendra | otubo, here is sample network_data.json looks like https://paste.ubuntu.com/p/Tb5vk6SgT4 | 11:10 |
otubo | ok | 11:13 |
vijayendra | otubo, cloud.cfg: https://paste.ubuntu.com/p/KbNGjqHYYq/ | 11:13 |
otubo | vijayendra: I might need the network_data.json as well | 11:14 |
otubo | vijayendra: I need a yaml configuration like this https://cloudinit.readthedocs.io/en/latest/topics/network-config-format-v2.html#examples | 11:15 |
vijayendra | otubo, this is network_data.json only https://paste.ubuntu.com/p/Tb5vk6SgT4 | 11:16 |
vijayendra | you can see step cat 1.json | 11:17 |
vijayendra | otubo, we don't have this network yaml config in our setup. I believe cloudinit just uses this network_data.json | 11:21 |
vijayendra | otubo, cloudinit is not using any other config here https://paste.ubuntu.com/p/8q84fGC74s/ | 11:24 |
vijayendra | otubo, let me check configDrive once again, if it has any network yaml | 11:32 |
vijayendra | otubo, I don't find any network yaml inside config drive too https://paste.ubuntu.com/p/WbmsnyJDG4/ | 11:40 |
=== cpaelzer__ is now known as cpaelzer | ||
Odd_Bloke | rharper: smoser: Do either of you know of a reason why we shouldn't fix https://bugs.launchpad.net/cloud-init/+bug/1900837? It seems clear-cut to me, IDK if I'm missing some historical context though, as this behaviour has been in place since at least 2012. | 14:48 |
ubot5 | Ubuntu bug 1900837 in cloud-init "cloud-init resets permissions on log file after reboot" [High,Triaged] | 14:48 |
smoser | i think there is a permissions setting on it. | 14:51 |
smoser | log_file_persm | 14:51 |
Odd_Bloke | I think that's owner/group. | 14:52 |
smoser | yeah, and its also syslog_fix_perms | 14:52 |
smoser | which may not be the cloud-init log. | 14:52 |
smoser | so i agree that if it is present, we should not change perms. | 14:53 |
Odd_Bloke | rharper: I'm looking at the Azure testing you asked me to do on https://github.com/canonical/cloud-init/pull/608 (apologies for missing the request the first time!); how would you anticipate failure presenting? (My underlying question: would running the SRU script for Azure, which includes advanced networking, be sufficient to confirm that all is good?) | 16:24 |
Odd_Bloke | (From reading through, it looks to me like that is likely sufficient, but some confirmation would be appreciated. :) | 16:25 |
blackboxsw | :q | 16:41 |
blackboxsw | oops | 16:42 |
rharper | Odd_Bloke: on the azure instance, it would make the same duplicate mac complaint | 17:15 |
rharper | Odd_Bloke: so a successful boot up of an instance with cloud-init and that PR is sufficient; the ovs side and verifying normal maas bond/bridge is verified; so just checking that we don't regress azure + advanced networking; | 17:16 |
Odd_Bloke | rharper: Ack, thanks! | 17:49 |
meena | i'm online at a "reasonable" time, and i'm not brain-dead, and i have hardly any real work, so who's up for holding my hand, and spoon feeding me python tests? | 17:57 |
meena | smoser: thanks for the start on the kenv patch | 18:00 |
smoser | you're welcome. | 18:05 |
smoser | the unit tests for ds-identify are ... interesting | 18:06 |
meena | if by interesting you mean gross, then, i agree. | 18:06 |
meena | do i have to recompile my quemu / libvirt / kvm to temporarily fake a different vendor? | 18:12 |
meena | smoser: anyway, i;d also have to refactor util.py / read_dmi_data() | 18:15 |
smoser | meena: yeah, you wil have to do that too. | 18:17 |
meena | anyway, i'd like to get this into the next release, cuz… well, i fucked that up last year, would be good to fix it, too… | 18:18 |
smoser | and by gross... if you can find a better way to get reasonable coverage on a shell script, then go for it. | 18:18 |
meena | smoser: 🙈️🙉️🙊️ | 18:19 |
meena | anyone else really surprised by emails from launchpad bugs, and then really disappointed that it's from you yourself | 21:13 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!