blackboxsw | FWIW, seems to "work fine"= on the cosmic azure instance I deployed. I could see the encoded content in the proper kvp_pool file | 01:46 |
---|---|---|
otubo | smoser: stupid question: that's my first time contributing on Launchpad (I'm from the mailing-list era). Shall I create a new branch a new pull-request? Or just push on top of that branch you already reviewed? | 14:19 |
smoser | otubo: welcome to the cloud era ;) | 14:20 |
smoser | you can push --force over whatever branch you have | 14:20 |
smoser | launchpad will do the right thing | 14:20 |
smoser | or you can 'git commit -a -m "add stuff smoser requested"' | 14:21 |
smoser | and then git push (without --force) | 14:21 |
smoser | generally i prefer the latter as it then is obvious what you've done to address that. | 14:21 |
otubo | smoser: anything else after that? Like, create a new pull-request? Just push -f and we're good? | 14:21 |
smoser | otubo: sorry, looked away. yeah, if you push over then launchpad upates the existing merge proposal | 14:49 |
smoser | blackboxsw: in your jinja branch... should i be able to reference 'v1.cloud_name' ? | 19:16 |
smoser | or is that not yet done. | 19:16 |
blackboxsw | smoser: I think I found a bug in render, I'm trying to fix it now. | 19:17 |
blackboxsw | rendered instance-data.json isn't being converted, I need to add a unit test for this | 19:17 |
blackboxsw | smoser: and yes you should be able to | 19:17 |
blackboxsw | I was just testing it trying to draw up slides | 19:18 |
blackboxsw | ohh I forgot to back out our persisting underscore values in initial instance-data.json and move it back into convert_jinja_instance_data | 19:20 |
smoser | yeah, that seems to be it. if i changed my instance_data to have underscores then it works. | 19:21 |
blackboxsw | yeah, and I think we decided we didn't want to persist underscores in instance-data.json as snap relies on v1 keys that don't have underscores | 19:22 |
smoser | yeah. | 19:22 |
smoser | thats unfortunate though | 19:22 |
blackboxsw | yeah costly to do that processing on each part render | 19:23 |
blackboxsw | smoser: alternative: maybe at instance-data.json write time (in ds.get_data()) we could write the jinja processed content | 19:25 |
blackboxsw | then we do it just once | 19:25 |
blackboxsw | and souce that processed jinja_vars file when rendering jinja content? | 19:25 |
blackboxsw | *source* rather | 19:25 |
smoser | we can figure something out there. | 19:27 |
blackboxsw | ok pushing a fix, but I need to sort a unit test failure still | 19:33 |
=== apollo13_ is now known as apollo13 | ||
smoser | blackboxsw: for your consideration http://paste.ubuntu.com/p/QqGC8SdkgG/ | 20:49 |
blackboxsw | smoser: for your consideration https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/353530 | 20:51 |
blackboxsw | thanks smoser will play with that and pull it in. | 20:53 |
smoser | blackboxsw: imds is https://docs.microsoft.com/en-us/azure/virtual-machines/windows/instance-metadata-service ? | 20:56 |
smoser | shouldnt we have a version in our data then | 20:56 |
blackboxsw | correct. smoser. after our discussion yesterday I started a branch to better represent the raw azure data (which has already landed) | 20:58 |
blackboxsw | I'm not done w/ that branch yet to restucture instance-data.json on azure | 20:59 |
* blackboxsw thinks it was yesterday, or maybe it was friday | 20:59 | |
smoser | :) | 20:59 |
smoser | well, i hit a comment there. to that effect. | 21:00 |
smoser | i have to run now. | 21:00 |
blackboxsw | in any event, I think openstack and azure ds's both need a little tweak | 21:00 |
blackboxsw | see you | 21:00 |
smoser | blackboxsw: fyi, there is a 0821 cosmic build now \o/ | 21:11 |
blackboxsw | woot! | 21:11 |
blackboxsw | getting easier | 21:11 |
blackboxsw | each time | 21:11 |
* smoser heads out. | 21:11 | |
socket- | Can anyone help me understand if/how cloudinit sets the hostname to ip-xxx-xxx-xxx-xxx? We want to change the default hostname as we spin up images to be 15 random hex characters instead of ip based. | 22:15 |
blackboxsw | socket-: cloud-init generally gets hostname from the cloud's metadata service if present and applies it. For setting or generating hostname a user can specify a static hostname with a #cloud-config yaml file at instance creation time. | 22:22 |
blackboxsw | https://pastebin.ubuntu.com/p/DZ43ptvtPT/ | 22:23 |
blackboxsw | example ^ | 22:23 |
socket- | blackboxsw: I'm not sure how familiar you are with AWS, but for example, we are using the ubuntu AMI (which I believe Ubuntu created) and it's default cloud init. Are you saying I would just modify the cloud-config to say something like hostname: $rand(15,[A-Z,a-z,0-9], save it as a new image, and each time someone launches one, they will get a random hostname? | 22:26 |
socket- | looking at the default cloud.cfg, the only hostname option i see is "- set_hostname" under cloud_init_modules | 22:28 |
blackboxsw | socket-: when the following branch is published you'll be able to specify something like the following: https://pastebin.ubuntu.com/p/jcVvmgZSPh/ | 22:30 |
blackboxsw | socket-: correct, via aws CLI or web UI you could currently a user-data file containing your "#cloud-config\nhostname: <random-value>" statically generated content | 22:31 |
blackboxsw | socket-: sorry branch in question is https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/335290 with some examples of rendering jinja templates. | 22:34 |
blackboxsw | hrm. socket- I so you want a custom AMI that will by default set the hostname to a random string right... you aren't in control at the time the instance is created is that right? | 22:37 |
blackboxsw | hrm, gotta bail for dinner | 22:47 |
socket- | sorry man, got called away from my desk, yeah your right I want a custom ami that wukk set the hostname to a random string, and I am not in control of the time its created | 23:02 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!