Hello everyone,
I created a test environment to prepare for the RHCE exam and for some days now I`m having a problem with Ansible Navigator. The problem began when I updated the packages on my RHEL 9.4 control node, after that when I tried to execute my scripts with ansible-navigator, scripts which should run on the control node where also ansible-navigator is installed, gave me the following error:
[student@control-station my_project]$ ansible-navigator run -mstdout ping-all.yml -vvvv
<cut>
debug1: Control socket \"/home/runner/.ansible/cp/5e0488efdc\" does not exist
debug2: resolving \"control-station\" port 22
debug2: ssh_connect_direct
debug1: Connecting to control-station [192.168.122.164] port 22.
debug2: fd 3 setting O_NONBLOCK
debug1: connect to address 192.168.122.164 port 22: Connection refused
ssh: connect to host control-station port 22: Connection refused",
"unreachable": true
<cut>
I checked the /etc/hosts, ~/.ssh/config, inventory files but everything is fine with them.
[student@control-station ~]$ cat /etc/hosts
127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6
192.168.122.164 control-station
[student@control-station ~]$ cat .ssh/config
Host control-station
Hostname control-station
User student
Port 22
IdentityFile ~/.ssh/ansible_key
[student@control-station my_project]$ cat inventory
[workstation]
control-station
On one forum they suggested that if you define a custom control path in ansible.cfg it will fix the problem; I tried it:
cat ansible.cfg
[defaults]
inventory=./inventory
remote_user = student
gathering = explicit
[ssh_connection]
control_path_dir = /home/student/.ansible/cp/
control_path = %(directory)s
the error remains the same, only the path changes in 'debug1: Control socket \"/home/runner/.ansible/cp/5e0488efdc\" does not exist' to the path defined in the configuration file;
while this configuration was in place, I observerd a strange behaviour: I added a remote machine to the inventory file and changed the current directory into /home/student/.ansible/cp/ and ran the ansible script, in this case ansible-navigator created the connection socket for the remote machine, but for the local machine, control node, it didn`t do anything.
On another forum, it was suggested to set the control_path variable to none, then the "Control socket..." line dissapears from the output, but the rest remains the same.
I also looked at the /var/log/secure, /var/log/messages files; the /var/log/secure file does not contain anything usefull, the /var/log/messages file contains the same messages which are written on the screen by running ansible-navigator run -mstdout ping-all.yml -vvvv
Thinking that maybe when I updated the packages on the system, something did not update correctly, or some packages are too new for ansible-navigator, I created a new VM, with a bare RHEL 9.4 installation, installed the latest ansible-navigator and downloaded the ee-supported-rhel8:latest image from registry.redhat.io/ansible-automation-platform-25/ee-supported-rhel8:latest; the problem persisted...
Also tried older versions of anisble-navigator images from registry.redhat.io/ansible-automation-platform-22/23/24/ee-supported-rhel8:latest, on the bare system and also on my current control node, which has been updated, the problem still persists.
Tried to run same playbook with ansible-playbook and it runs as expected:
[student@control-station my_project]$ ansible-playbook ping-all.yml -v
<cut>
ok: [control-station] => {"ansible_facts": {"discovered_interpreter_python": "/usr/bin/python3"}, "changed": false, "ping": "pong"}
<cut>
Has anybody encountered this problem with ansible-navigator? If so, how did you solve it?
Thanks in advance
