| === juergh_ is now known as juergh | ||
| === hjensas is now known as hjensas|afk | ||
| === cpaelzer__ is now known as cpaelzer | ||
| === hjensas|afk is now known as hjensas | ||
| oznt | hi guys, just a quick question: how does cloud init translate "set-passwords" in cloud_config_module to "cc_set_passwords" ? Can't figure this one out | 12:35 |
|---|---|---|
| oznt | Well, I found out myself : https://github.com/canonical/cloud-init/blob/7c88f96245535aacfac9b0b04d9c0d82ac2067f4/cloudinit/config/__init__.py | 12:47 |
| lucasmoura | blackboxsw, thansk for the review :) I will start working on the proposed changes today | 13:12 |
| Odd_Bloke | oznt: Was that your entire question, or is there anything else we can help you out with? | 13:36 |
| oznt | Odd_Bloke, that was the question | 13:38 |
| Odd_Bloke | OK, great. :) | 14:18 |
| Odd_Bloke | falcojr: I've pushed the requested changes to both https://github.com/canonical/cloud-init/pull/354 and https://github.com/canonical/cloud-init/pull/349 :) | 17:06 |
| falcojr | cool, I'll re-review now | 17:31 |
| falcojr | I've technically already approved on the second one, you just need another approval from a committer | 17:32 |
| usrdev | hello all - hope you are safe and well! just a quick question. when utilizing cloud-init to create templates, i have noticed that the startup script for network configuration is severely delayed. i have cloud-init to pull a DHCP lease, and this works without issue for directly standing up VM from an ISO. i am utilizing proxmox as the HV. | 17:33 |
| usrdev | is there something i am missing to make this DHCP lease request, quicker... | 17:33 |
| Odd_Bloke | blackboxsw: Can you add https://github.com/canonical/cloud-init/pull/349 to your list of PRs to double-check as core committer, please? | 17:42 |
| Odd_Bloke | We also didn't assign reviews today, and I'm first so I've taken #358. | 17:44 |
| blackboxsw | Odd_Bloke: yes I can/will | 17:59 |
| AnhVoMSFT | @Odd_Bloke, did you have a chance to take a look into the issue with the test_conftest having pylint error https://pastebin.ubuntu.com/p/QMkhNtShjW/ | 18:40 |
| Odd_Bloke | AnhVoMSFT: I didn't, apologies! | 19:30 |
| Odd_Bloke | Let me take a quick look now. | 19:30 |
| Odd_Bloke | AnhVoMSFT: Hmm, could you give me a more recent log? I'm seeing differences in package versions, and it's hard to tell if that's just because there have been releases in the past ~week. | 19:36 |
| Odd_Bloke | AnhVoMSFT: Hmm, actually, this is stranger than that. I can reproduce the error if I give the exact path (i.e. `.tox/pylint/bin/python -m pylint cloudinit/tests/test_conftest.py`), but not if I only specify the directory (i.e. `.tox/pylint/bin/python -m pylint cloudinit/tests/`). I'm also running Python 3.8 (instead of, I think, 3.5 in these logs). | 19:51 |
| Odd_Bloke | AnhVoMSFT: And I also can't reproduce on 3.5, using tox. :/ | 19:59 |
| AnhVoMSFT | @Odd_Bloke yeah, I am having a hard time reproducing on my local machine - let me try with the .tox/pylint/bin/python. That should be the same version though | 20:08 |
| AnhVoMSFT | yes I can repro using that particular command | 20:10 |
| AnhVoMSFT | and I am on 3.5.2 on my xenial box (python3 --version) | 20:10 |
| AnhVoMSFT | the build server gets a new container every time it runs the test, so whatever is in .tox isn't cached between runs | 20:10 |
| Odd_Bloke | AnhVoMSFT: OK, I think we're dealing with some sort of pylint bug here: I see different output for `.tox/pylint/bin/python -m pylint -j1 cloudinit.tests.test_conftest cloudinit.tests.test_util` and `.tox/pylint/bin/python -m pylint -j1 cloudinit.tests.test_util cloudinit.tests.test_conftest`. Can you confirm you see different behaviour also? | 20:46 |
| Odd_Bloke | AnhVoMSFT: I'm filing a pylint issue now, will drop the link here once I have done so. | 20:56 |
| Odd_Bloke | AnhVoMSFT: https://github.com/PyCQA/pylint/issues/3611 | 21:11 |
| AnhVoMSFT | yes @Odd_Bloke I can confirm I am seeing different behaviors between those two commands | 21:20 |
| AnhVoMSFT | how did the different ordering could come up in one build environment vs the others though? | 21:21 |
| AnhVoMSFT | Oh, is this because the container we were running the test had only one core ? | 21:21 |
| AnhVoMSFT | @Odd_Bloke: I guess the question now is: which one of them was the right one? (I know we want a consistent behavior, but which one is the "right" consistent behavior?) | 21:24 |
| Odd_Bloke | AnhVoMSFT: So I think it's actually more likely that the container had many cores: I think the error is only reported if test_conftest is the _first_ module that a worker process tests (as I've just posted to the issue). | 21:39 |
| Odd_Bloke | (The number of CPUs in the host is almost certainly what pylint will discover.) | 21:40 |
| Odd_Bloke | AnhVoMSFT: So in this case, that test is specifically testing for behaviour when the wrong arguments are called (i.e. exactly what this pylint check is warning us about :p). So it's a valid pylint error, which we should be ignoring in this specific case. | 21:41 |
| Odd_Bloke | However, I bet we also have a whole load of other instances of this error (or others which are similarly buggy). | 21:41 |
| blackboxsw | lucasmoura: thanks one minor nit on https://github.com/canonical/cloud-init/pull/359 | 21:52 |
| blackboxsw | then we can land that | 21:52 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!