-
Notifications
You must be signed in to change notification settings - Fork 51
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
CI: add 2.14 tests, create ignore-2.15 to prevent CI failure #139
Conversation
I tried to check why these failed but it appears that the old logs have been deleted. This was perhaps erroring due to #141 but unsure. |
I'll re-test and see where Ubuntu fails with fresh logs... |
@csmart I read on another forum that perhaps ssl1.1 isn't available via apt on Ubuntu 22.04. It looks like this PR introduces 22.04? Looks like it maybe v3 now... https://packages.ubuntu.com/jammy/libssl-dev Edit: 20.04 looks like version 1.1.1 .. https://packages.ubuntu.com/focal/libssl-dev |
Looks like it's failing when trying to talk to an IPv6 localhost? Maybe something changed with Docker 2.14 images..
|
Ci 2 15
Pipeline taking ages today! Must be working it too hard. :) |
@csmart I'm so sorry, I was adding a PR via the github web interface on your forked repo and ci_2_15 branch... github gave me a 404 error. Then out of curiosity I tried the other option fully expecting it to fail, but, it allowed me to update your forked repo! Really sorry about that... it is probably because I'm on the maintainers list? Edit: CI is now failing due to not being able to connect to 192.168.250.1 in the lookup_rabbitmq plugin. I'm afraid to touch git right now but the |
@Im0 oh, no need to apologise! I don't think you did anything wrong My only concern I had before about the change that I introduced checking if connection failed, I'm not sure if that's a valid response to the request, or whether that might mask a problem where we can't talk to rabbitmq for another reason... I tried to run the tests locally to check, but I ran into other issues and they wouldn't run... And then I ran out of time... So please feel free to try whatever you think and we'll go from there. Thanks for the help! |
Ci 2 15 x2
I'll take a look at these today. |
Test 1: "Connection issue: module 'collections' has no attribute 'Callable'"
Thought # 1: This is fixed already in the exchange patch (PR #140) Try # 2: Copy rabbitmq_publish and module_utils/rabbitmq.py from fix_138_publish_to_exchange for testing Test 2: "Connection issue: Connection to ::1:5672 failed: [Errno 99] Cannot assign requested address"
Try # 3: Test 3: "Connection issue: Connection to ::1:5672 failed: [Errno 99] Cannot assign requested address" Suggest merging PR #140 and rebasing csmart:ci_2_15 |
Updating to pull in recent merge
Well, we got a CI success! But only one check. I can't see how our change would have dropped all other checks. |
Currently 20.04 is causing devel to fail. This maybe due to Ubuntu 20.04 running python 3.8, however, ansible-test requires 3.9+. This means ansible test spins up a controller and target container which is probably why rabbitmq_publish is not able to connect to rabbitmq on localhost.
@csmart Hi mate, I've disabled 20.04 from 2.14 and devel as @felixfontein (thanks heaps) suggested that the issue maybe that Ubuntu 20.04 uses python 3.8 but ansible-test requires 3.9+. Ansible-test therefore uses a controller and target container which is probably why we can't hit it with localhost in rabbitmq_publish. I think we have good coverage with what we have and am happy with this PR as is. I'm happy to merge if you are. However maybe we create an issue to try and fix 20.04 on 2.14/devel? I got some container information in another CI run which maybe useful:
I'm not sure if 172.19.0.3 will be the controller or target... I'm assuming controller as otherwise we'd be able to connect to 127.0.0.1. |
The IP 172.19.0.3 is only valid in this specific CI run, it might be different on every other one. You probably need to dynamically query the IP of the target container (some ansible facts should help for that) and use that. |
@felixfontein I got some ansible facts out here: https://dev.azure.com/ansible/community.rabbitmq/_build/results?buildId=58759&view=logs&j=1d520954-9606-5520-b594-0b935de5aeca&t=cbeffed3-49d8-5f25-fa26-6591d04d4a87 I don't understand if this was generated by the controller or target? I presume target? In which case we could use ansible_facts['ansible_eth0']['ipv4.address'] ? |
@Im0 that should be for the target. The debug in line 3691 shows that |
Brilliant, thanks! I'll try that. |
Wow, this MR got complicated 😄 Thanks @Im0 |
Fixes #137