Closed discostur closed 4 months ago
Ok debugged it the last days and it seems to be related to slirp4netns
. Maybe its a race condition or something like that. However, with podman 3 and slirp4netns 1.0.1 there seems to be many ipv6 issues (when you scroll through the github issues). Since my receptor node (and also podman) is running on debian 11 i won't get any newer versions with bugfixes.
However i found a workaround to disable slirp4netns and enable host networking. At the moment this cannot be configured on a job or template based option - you can just set it global via job settings (in awx ui):
old:
[
"--network",
"slirp4netns:enable_ipv6=true"
]
new:
[
"--network=host"
]
I found someone who was able to set it via extra_settings but that didn't work for me:
https://forum.ansible.com/t/awx-instance-receptor-podman-bug/2599/5 https://github.com/ansible/awx-operator/blob/devel/docs/user-guide/advanced-configuration/extra-settings.md
Podman then was always running a command like
/usr/bin/podman run --rm --tty --interactive --workdir /runner/project -v /tmp/awx_1193_ycvbr77j/:/runner/:Z --env-file /tmp/awx_1193_ycvbr77j/artifacts/1193/env.list --quiet --name ansible_runner_1193 --user=root - - n e t w o r k = h o s t
which seems to be a syntax issue. However in the configmap on awx it look correct.
It would also be nice to configure such podman run options on a job or template base level and not just globally. It seems there are some tickets already open but nothing happend:
https://github.com/ansible/awx/issues/11552 https://github.com/ansible/awx/issues/12339
So for the moment it works for me via global job configuration (network=host).
Please confirm the following
security@ansible.com
instead.)Bug Summary
I deployed AWX (v24.5.0) in k8s via helm. I added a receptor node (execution). If the server which i want to provision has an ipv4 address everything is working normal.
If the server i want to run my playbooks against as an ipv6 address, sometimes the job finishes successful and sometimes it failes. It is exactly the same job template. If i just click on re-run it sometimes works and sometimes not. Cannot see any pattern when it works and when not ...
AWX version
24.5.0
Select the relevant components
Installation method
kubernetes
Modifications
no
Ansible version
No response
Operating system
Debian
Web browser
Firefox
Steps to reproduce
Expected results
Working
Actual results
Broken / Network unreachable
Additional information
I even did a tcpdump on the interface on my receptor node and i can see that when awx says
Network is unreachable
it DOES NOT try to connect. I can not see any tcp packet on the interface.When it is working i see normal tcp packets on the interface.
I did some ICMP / manual SSH connection attempts and it always works. I run a ICMP test for about 60 minutes from the receptor node to the target server and no single package got dropped. Network is table and running without any issues.