When I try to execute the ping command on an external server (authentication via ssh), a connection timeout message is displayed. When creating credentials, the type was selected “Machine” and the corresponding ssh key was added. In addition, the ping to the external VM also did not work via ssh, while it went through the password successfully. I want the connection via ssh to go through successfully.
Deployment details:
OS version: Ubuntu 20 04
Client Version: v1.25.5+k3s2
Kustomize Version: v4.5.7
Server Version: v1.25.5+k3s2
The server is a VM created on KVM
AWX version - 21.5.0
In addition:
In the deployment version on docker (awx version 17.1.0) the problem described above does not occur.- After installing ansible locally on the server where the problem occurs, it is possible to execute the ad-hoc command
After installing ansible locally on the server where the problem occurs, it is possible to execute the ad-hoc command
Hello,
We would like to gather some more information from you. Could you run the ping module against the host again with verbosity set to 3 and provide us with that output? Thank you for that additional information!
Are you able to ssh directly from the awx control node command line? If so, how long does that connection typically take? If not, you should investigate any proxy or firewall that could be in the middle.
do you mind providing the output of /api/v2/ad_hoc_jobs/ (or /api/v2/jobs if you ran a playbook) for that ping? we can take a closer look (be sure to remove sensitive info before uploading)
can you successfully establish a connection to your host from inside the awx-ee container? You can use podman run to run awx-ee container (outside of awx), exec into it and try to ssh into your host? that is closer to how k8s will be doing it.
do other regular jobs (based on job templates) run okay (i.e. is it just ad hoc commands that cause issues?)
Hi,
First for all we want to apologize for long break. Some others tasks came to us that were need to solve. We think this is the output that you’ve asked for. We’re not using podeman. This output was intercepted by entering thru crictl to a running awx-ee container. Added in attachment.
BR
What does your deployment look like? the automation job pods are running in the same cluster as your control node correct? are you using a container group to connect out to a different cluster?
Hi,
In screens below, I’ve attached (I think) whole info about our deployment. Tbh I’m not master of k8s/k3s but from the tutorial that we used for setting up this environment I can deduce automation job pods are running in the same cluster as control node.
