Error while running ansible-playbook on OS 3.1.1

Latest response

After configuring 6 VMs (3 master, 3 node), I am attempting to run the playbook. I'm getting this for each of the master and nodes:

fatal: []: FAILED! => {"msg": "The task includes an option with an undefined variable. The error was: 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'openshift_version'\n\nThe error appears to have been in '/usr/share/ansible/openshift-ansible/playbooks/init/version.yml': line 20, column 5, but may\nbe elsewhere in the file depending on the exact syntax problem.\n\nThe offending line appears to be:\n\n  tasks:\n  - set_fact:\n    ^ here\n"}

Line 20 seems to be:

  - set_fact:
      openshift_version: "{{ l_first_master_openshift_version }}"
      openshift_pkg_version: "{{ l_first_master_openshift_pkg_version }}"
      openshift_image_tag: "{{ l_first_master_openshift_image_tag }}"

openshift-ansible version seems to be: 3.11.59-1.git.0.ba8e948.el7

Any ideas?



I seem to moved past the initial issue and landed on:

TASK [openshift_node_group : Update the sysconfig to group "node-config-compute"] ***********************
Monday 28 January 2019  10:25:19 -0600 (0:00:00.571)       0:08:31.286 ********
fatal: []: FAILED! => {"changed": false, "msg": "Destination /etc/sysconfig/atomic-openshift-node does not exist !", "rc": 257}

Can you please share the steps which you executed to resolve this issue?

Resolved :)

Yeah - wanna share?

Hi, I am experiencing this issue while performing upgrade to OCP 3.11... What was the cause?

My apologies for the late response, I did not get notifications for some reason. In any case, my issue was caused by ansible version. I had to downgrade to ansible 2.6.12. Hope this helps

I ran into same problem when using ansible 2.6.5. Changing to ansible 2.6.12 didn't help. For me, the fatal error message was a bit misleading. It was actually related to the fact that my master/boot node wasn't configured probably to allow ssh to itself without password

make sure you can go to the individual hosts and successfully run yum -y install <something>. There's a chance that ansible couldn't find the package because yum repo is not available.

Thanks Chee Ooi. Totally agree with you and the fact error is misleading.

After verifying the master host was not able to SSH itself was the real cause of the problem.. After adding it's own (Master01) public key to the authorized_keys file and able to ssh itself and reran the prerequisite check. Everything went smooth.