| Jingli | Hello :] . | 08:30 |
|---|---|---|
| Jingli | We are using cloudinit to provision our servers. And we meed a problem of vendor-data, it works in ubuntu but in Debian and CentOS doesn't work. | 08:31 |
| Jingli | Can someone help, please ? | 08:32 |
| meena | Jingli: can you be a little more specific with "It doesn't work" | 09:02 |
| meena | do you have logs on how exactly it's failing? | 09:02 |
| meena | are you trying to use the same vendor data for different OSes? | 09:02 |
| Jingli | Yes, I tried on all ubuntu and Debian Buster 10 | 09:16 |
| Jingli | Uploaded file: https://uploads.kiwiirc.com/files/a0b0cfe36ff9e709c036f9db48fd513f/pasted.txt | 09:16 |
| Jingli | All the OSes are using the same vendor-data above | 09:16 |
| Jingli | I checked /var/log/cloud-init-output.log . it doesn't show any log which is related to the modules which is in the vendor-data, I have the felling , the vendor-data is not applied at all. | 09:18 |
| Jingli | user-data is empty | 09:19 |
| meena | You got logging fully enabled? | 09:27 |
| Jingli | Ah, nope. I will do it. | 09:29 |
| meena | do you know how or do you need help? | 09:36 |
| meena | Jingli: is the user-data empty on purpose or is it not discovered by the machines? | 09:37 |
| Jingli | user-data empty is on purpose. | 09:37 |
| meena | aye | 09:38 |
| Jingli | ops | 09:38 |
| Jingli | should not ? | 09:38 |
| meena | if vendor-data is all you need so far, it makes sense. but please be aware that user-data can override vendor-data; which often makes perfect sense | 09:39 |
| Jingli | yes. | 09:39 |
| Jingli | Uploaded file: https://uploads.kiwiirc.com/files/bdda3aa05103b1b0e2f85ed6526a98d5/pasted.txt | 09:40 |
| Jingli | And to enable log. should add this to vendor-data for example? | 09:40 |
| meena | I have full debug logging enabled always, because, well, i hack on cloud-init | 09:45 |
| Jingli | Nice. I would like the same =D Is it on your repo ? | 09:47 |
| meena | Jingli: it's just a config option | 10:00 |
| meena | https://github.com/canonical/cloud-init/blob/main/config/cloud.cfg.d/05_logging.cfg and it's the default if this log file exists or is named correctly | 10:01 |
| Jingli | 👍 Thank you ! | 10:06 |
| Jingli | Uploaded file: https://uploads.kiwiirc.com/files/299bcf85c6e92d4c2bb1a44518248b41/pasted.txt | 11:22 |
| Jingli | 👍enabled logs found this | 11:22 |
| Jingli | But it's set `manage_etc_hosts:true` in vendor-data | 11:23 |
| meena | you might need a space between : and true | 12:27 |
| Jingli | Yep, still no luck. | 13:13 |
| Xe | is there any cloud-init equivalent for Android? | 15:10 |
| rharper | Xe: I'm not aware of such a thing; but that doesn't mean there's not something like it; I've never gone looking for something like that. | 15:22 |
| blackboxsw | falcojr: the other thing I was thinking with the py version branch was this attempt to deserialize check as well and blow it away if we know we can't. https://paste.ubuntu.com/p/Vp27bsphmD/ But, I'm not sure that really solves the underlying problem Moustafa mentioned about ancillary stdlib serialization issues across python upgrades | 16:08 |
| blackboxsw | maybe this would solve the initial concern moustafa mentioned, I'm not sure if their error was during pickle deserialization w/ mime-type user-data or not. | 16:11 |
| falcojr | blackboxsw: we can ask moustafa, but the two issues we dealt with recently (https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1899299) and (https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1922739) both happened after deserialization and on access | 16:15 |
| ubottu | Launchpad bug 1899299 in cloud-init (Ubuntu Hirsute) "do-release-upgrade from Trusty to Xenial breaks cloud-init init-local on reboot" [Undecided, Fix Released] | 16:15 |
| ubottu | Launchpad bug 1922739 in cloud-init "AttributeError: 'DataSourceNoCloud' object has no attribute 'vendordata2_raw'" [High, Fix Released] | 16:15 |
| falcojr | so I'm not sure that sort of addition would really buy us much | 16:16 |
| blackboxsw | yeah kinda feels like this would be the same type of case. access time, not deserialization time | 16:16 |
| falcojr | right | 16:16 |
| blackboxsw | community-notice: cloud-init office hours is beginning in a moment. Next office hours will be listed in the topic "July 13 17:15 UTC" | 16:23 |
| blackboxsw | #startmeeting cloud-init bi-weekly office-hours | 16:23 |
| meetingology | Meeting started at 16:23:41 UTC. The chair is blackboxsw. Information about MeetBot at https://wiki.ubuntu.com/meetingology | 16:23 |
| meetingology | Available commands: action, commands, idea, info, link, nick | 16:23 |
| blackboxsw | Hi cloud-init devs, Our biweekly office hours meeting has begun. We keep this channel open during the designated time on the channel #topic for drop in questions, feature or bug discussions and prodding committers folks for reviews. | 16:25 |
| blackboxsw | All questions welcome and we'll have a couple of upstream committers with eyes on the channel to field any discussions brought up as best we can. | 16:27 |
| blackboxsw | Just prior to this meeting, someone will post recent happenings in cloud-init upstream which may be of value to the community. | 16:27 |
| blackboxsw | #link https://discourse.ubuntu.com/t/cloud-init-status-06-29-2021/22981 | 16:27 |
| blackboxsw | As far as upstream is concerned, the only schedule detail of note is we are looking to create an official 21.3 upstream release August 28th. Before that time we hope to have some hot plug network device handling functional in cloud-init. | 16:29 |
| blackboxsw | my plan today is to get falcojr a review on his foundational hotplug support work at https://github.com/canonical/cloud-init/pull/919 which should allow cloud-init to support actively interacting with network config utilities to actively enable individual interfaces on device discovery after the network renderers have finished initial network bringup. | 16:34 |
| ubottu | Pull 919 in canonical/cloud-init "Add new network configurers to bring up interfaces (SC-85)" [Open] | 16:34 |
| === blackboxsw changed the topic of #cloud-init to: 21.2 (May 6) | file a bug: https://bugs.launchpad.net/cloud-init/+filebug | pull-requests: https://git.io/JeVed | meeting minutes: https://goo.gl/mrHdaj | next office hours: July 13 17:15 UTC | ||
| blackboxsw | ok, that wraps things up for "office hours" for today. Thanks for checking in. | 17:35 |
| blackboxsw | #endmeeting | 17:35 |
| meetingology | Meeting ended at 17:35:33 UTC. Minutes at https://new.ubottu.com/meetingology/logs/cloud-init/2021/cloud-init.2021-06-29-16.23.moin.txt | 17:35 |
| blackboxsw | "meeting" minutes posted to https://cloud-init.github.io/ | 17:44 |
| === mamercad3 is now known as mamercad | ||
| blackboxsw | falcojr: at long last another round of bike-sheds and I think a fix for gentoo.py needed I think https://github.com/canonical/cloud-init/pull/919/files | 21:48 |
| ubottu | Pull 919 in canonical/cloud-init "Add new network configurers to bring up interfaces (SC-85)" [Open] | 21:48 |
| falcojr | blackboxsw: thanks for the review! I'll probably get back to the bikeshedding tomorrow though :) | 21:50 |
| blackboxsw | +1 falcojr I'm game to talk through the bike-shed options whenever and we can settle and move forward, 'nuff delay on my naming pricklies | 21:54 |
| falcojr | Haha. Took a quick look and I like NetworkActivator. I'm good with that! | 21:56 |
| blackboxsw | woo hoo! | 22:17 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!