search

k3s-io / k3s-ansible

Apache License 2.0
1.89k stars 780 forks source link

Errors encountered when airgap_dir is undefined #257

Closed jon-stumpf closed 7 months ago

jon-stumpf commented 7 months ago

I attempted to create a simple cluster with three nodes. Here is my inventory.yml file.

---
k3s_cluster:
  children:
    server:
      hosts:
        10.140.240.20:
    agent:
      hosts:
        10.140.240.21:
        10.140.240.22:

  vars:
    ansible_port: 22
    ansible_user: pi
    k3s_version: v1.26.9+k3s1
    token: "mytoken"  # Use ansible vault if you want to keep it secret
    api_endpoint: "{{ hostvars[groups['server'][0]]['ansible_host'] | default(groups['server'][0]) }}"
    extra_server_args: ""
    extra_agent_args: ""

When running ansible-playbook playbook/site.yml, I get the following errors from role/airgap:

TASK [airgap : Distribute K3s amd64 images [Airgap]] **************************************************************************************************************************************************
Tuesday 28 November 2023  12:35:25 -0500 (0:00:00.191)       0:00:44.666 ****** 
fatal: [10.140.240.20]: FAILED! => {"msg": "No file was found when using first_found. Use errors='ignore' to allow this task to be skipped if no files are found"}
fatal: [10.140.240.21]: FAILED! => {"msg": "No file was found when using first_found. Use errors='ignore' to allow this task to be skipped if no files are found"}
fatal: [10.140.240.22]: FAILED! => {"msg": "No file was found when using first_found. Use errors='ignore' to allow this task to be skipped if no files are found"}

This error will occur at the following locations:

I can add ignore_errors: true to the ansible.builtin.copy tasks to allow the playbook to proceed, but this does not seem the correct thing to do (given the error messages still produced). And, since airgap_dir is not defined in my case, I don't understand why it is in the block to begin with. See https://github.com/k3s-io/k3s-ansible/blob/1e8bfb0d3967defe66929616f50a2f40d2470a87/roles/airgap/tasks/main.yml#L3

dereknola commented 7 months ago

I am unable to reproduce this issue:

Tested with 

---
k3s_cluster:
  children:
    server:
      hosts:
        192.168.1.91:

  # Required Vars
  vars:
    ansible_port: 22
    ansible_user: pi
    k3s_version: v1.26.9+k3s1
    token: "mytoken"  # Use ansible vault if you want to keep it secret
    api_endpoint: "{{ hostvars[groups['server'][0]]['ansible_host'] | default(groups['server'][0]) }}"
    extra_server_args: ""
    extra_agent_args: ""

Running ansible-playbook -i pi.yml ./playbook/site.yml, I see the airgap tasks being skipped as designed. (Only change is I removed the [Airgap] on all the task names).

TASK [prereq : Copy manifests] ***********************************************************************************************************************
skipping: [192.168.1.91]

TASK [prereq : Make config directory] ****************************************************************************************************************
skipping: [192.168.1.91]

TASK [prereq : Copy config values] *******************************************************************************************************************
skipping: [192.168.1.91]

TASK [airgap : Download k3s install script] **********************************************************************************************************
skipping: [192.168.1.91]

TASK [airgap : Distribute K3s install script] ********************************************************************************************************
skipping: [192.168.1.91]

TASK [airgap : Distribute K3s binary] ****************************************************************************************************************
skipping: [192.168.1.91]

TASK [airgap : Distribute K3s SELinux RPM] ***********************************************************************************************************
skipping: [192.168.1.91]

TASK [airgap : Install K3s SELinux RPM] **************************************************************************************************************
skipping: [192.168.1.91]

TASK [airgap : Make images directory] ****************************************************************************************************************
skipping: [192.168.1.91]

TASK [airgap : Determine Architecture] ***************************************************************************************************************
skipping: [192.168.1.91]

TASK [airgap : Distribute K3s amd64 images] **********************************************************************************************************
skipping: [192.168.1.91] => (item=/home/derek/rancher/k3s-ansible/roles/airgap/) 
skipping: [192.168.1.91]

TASK [airgap : Distribute K3s arm64 images] **********************************************************************************************************
skipping: [192.168.1.91] => (item=/home/derek/rancher/k3s-ansible/roles/airgap/) 
skipping: [192.168.1.91]

TASK [airgap : Distribute K3s arm images] ************************************************************************************************************
skipping: [192.168.1.91] => (item=/home/derek/rancher/k3s-ansible/roles/airgap/) 
skipping: [192.168.1.91]

TASK [airgap : Run K3s Install [server]] *************************************************************************************************************
skipping: [192.168.1.91]

TASK [airgap : Run K3s Install [agent]] **************************************************************************************************************
skipping: [192.168.1.91]
dereknola commented 7 months ago

The skip line at https://github.com/k3s-io/k3s-ansible/blob/master/roles/airgap/tasks/main.yml#L73 is specifically designed to avoid the errors you are receiving, but the without the airgap_dir because of https://github.com/k3s-io/k3s-ansible/blob/1e8bfb0d3967defe66929616f50a2f40d2470a87/roles/airgap/tasks/main.yml#L3, that entire airgap task should be skipped.

jon-stumpf commented 7 months ago

:-| Checking further.

jon-stumpf commented 7 months ago

I am at a loss. I have debug statements within the block that are appropriately not executing and it is skipping all previous tasks like your output. However, for some reason, it is evaluating this task. Trying to find the difference.

I am running:

ansible-playbook 2.10.8 config file = \~/turingpi/k3s-ansible-latest/ansible.cfg configured module search path = ['\~/.ansible/plugins/modules', '/usr/share/ansible/plugins/modules'] ansible python module location = /usr/lib/python3/dist-packages/ansible executable location = /usr/bin/ansible-playbook python version = 3.10.12 (main, Nov 20 2023, 15:14:05) [GCC 11.4.0]

@dereknola , what are you running?

jon-stumpf commented 7 months ago

And, the below works. Yet, it should be equivalent.

diff --git a/roles/airgap/tasks/main.yml b/roles/airgap/tasks/main.yml
index 378f61b..32cf7ac 100644
--- a/roles/airgap/tasks/main.yml
+++ b/roles/airgap/tasks/main.yml
@@ -60,13 +60,14 @@
     - name: Distribute K3s amd64 images [Airgap]
       when: ansible_architecture == 'x86_64'
       ansible.builtin.copy:
-        src: "{{ item }}"
+        src: "{{ lookup('first_found', params) }}"
         dest: /var/lib/rancher/k3s/agent/images/k3s-airgap-images-amd64.tar
         owner: root
         group: root
         mode: 0755
-      with_first_found:
-        - files:
+      vars:
+        params:
+          files:
             - "{{ airgap_dir }}/k3s-airgap-images-amd64.tar.zst"
             - "{{ airgap_dir }}/k3s-airgap-images-amd64.tar.gz"
             - "{{ airgap_dir }}/k3s-airgap-images-amd64.tar"
dereknola commented 7 months ago

My core version is several releases newer, could that be the difference?

ansible-playbook [core 2.15.6]
  config file = None
  configured module search path = ['/home/derek/.ansible/plugins/modules', '/usr/share/ansible/plugins/modules']
  ansible python module location = /home/derek/.local/lib/python3.10/site-packages/ansible
  ansible collection location = /home/derek/.ansible/collections:/usr/share/ansible/collections
  executable location = /home/derek/.local/bin/ansible-playbook
  python version = 3.10.12 (main, Nov 20 2023, 15:14:05) [GCC 11.4.0] (/usr/bin/python3)
  jinja version = 3.0.3
  libyaml = True
dereknola commented 7 months ago

I just tried with ansible 2.10.17 and now can see the same error you do. I'll do more testing and see when this gets resolved.

dereknola commented 7 months ago

Okay moving up to ansible core 2.12.10 resolved the issue. I think expecting users to have a more modern ansible release (2.12 came out 08 Nov 2021 and is already EOL) is fine. I will update the README accordingly.

jon-stumpf commented 7 months ago

ansible 2.10.7 is the version for Ubuntu 22.04.03 LTS. I will update it from another repository. Thanks.

zen commented 7 months ago

I'm on ansible-playbook [core 2.15.2] and still encountering this issue :(

zen commented 7 months ago 
TASK [airgap : Determine Architecture] **************************************************************************************************************************************************************************************************
task path: /Users/zen/projects/work/al/infra/k3s-ansible/roles/airgap/tasks/main.yml:62
skipping: [10.0.60.50] => {
    "changed": false,
    "false_condition": "airgap_dir is defined",
    "skip_reason": "Conditional result was False"
}
skipping: [10.0.60.51] => {
    "changed": false,
    "false_condition": "airgap_dir is defined",
    "skip_reason": "Conditional result was False"
}

TASK [airgap : Distribute K3s amd64 images] *********************************************************************************************************************************************************************************************
task path: /Users/zen/projects/work/al/infra/k3s-ansible/roles/airgap/tasks/main.yml:66
fatal: [10.0.60.50]: FAILED! => {
    "msg": "No file was found when using first_found."
}
fatal: [10.0.60.51]: FAILED! => {
    "msg": "No file was found when using first_found."
}

PLAY RECAP ******************************************************************************************************************************************************************************************************************************
10.0.60.50                 : ok=8    changed=0    unreachable=0    failed=1    skipped=28   rescued=0    ignored=0
10.0.60.51                 : ok=8    changed=0    unreachable=0    failed=1    skipped=28   rescued=0    ignored=0

❯
❯
k3s_cluster:
  children:
    server:
      hosts:
        10.0.60.50:
    agent:
      hosts:
        10.0.60.51:
        #10.0.60.52:

  # Required Vars
  vars:
    ansible_ssh_common_args: xxx
    ansible_user: ubuntu
    k3s_version: v1.28.3+k3s2
    token: "xx"
    api_endpoint: "{{ hostvars[groups['server'][0]]['ansible_host'] | default(groups['server'][0]) }}"
    extra_server_args: ""
    extra_agent_args: ""
zen commented 7 months ago

Bumping ansible to 2.15.6 solved the issue. Very weird.

lefeverd commented 6 months ago

Same error for me, upgrading to 2.16.2 solved it. Maybe the README should be updated ?