Ansible error handling rescue block does not have any indication that the block's failure is being handled by rescue.

[sekharvajjula]

ISSUE TYPE

• Bug Report

COMPONENT NAME

playbook/block.py

ANSIBLE VERSION

ansible 2.5.1

CONFIGURATION

N/A

OS / ENVIRONMENT

N/A

SUMMARY

Ansible error handling rescue block does not give any indication in the output log, that the block's failure is being handled in rescue. Rather the output logs show that one task has failed, but still it is continuing the execution with next task. Just by looking at the logs it does not give any information that rescue tasks are under execution unless we add debug task indicating the same (or) by browsing through the playbooks. This make life difficult in debugging playbooks with multiple roles in identify real failure.

If we take the example of retry tasks failures have clear prefix "FAILED - RETRYING: ", if we take the "ignore_errors" case there is a clear print statement "...ignoring" before proceeding to the next task. It would make debugging easy if we have similar sort of indication for block-rescue tasks as well.

STEPS TO REPRODUCE

Run the below playbook with ansible-playbook command.

---
- name: Testing block-rescue
  hosts: localhost
  tasks:
    - block:
        - name: read /tmp/test.data
          slurp:
            src: /tmp/test.data
          register: tmp_rabbit_cookie

      rescue:
        - name: create /tmp/test.data
          copy:
            content: "Testing data"
            dest: /tmp/test.data

EXPECTED RESULTS

ACTUAL RESULTS

[root@controller-1 playbooks]# ansible-playbook test.yaml
[WARNING]: provided hosts list is empty, only localhost is available

PLAY [Testing block-rescue] ************************************************************************************************************************************

TASK [Gathering Facts] *****************************************************************************************************************************************
ok: [localhost]

TASK [read /tmp/test.data] *************************************************************************************************************************************
fatal: [localhost]: FAILED! => {"changed": false, "failed": true, "msg": "file not found: /tmp/test.data"}

TASK [create /tmp/test.data] ***********************************************************************************************************************************
changed: [localhost]

PLAY RECAP *****************************************************************************************************************************************************
localhost : ok=3 changed=1 unreachable=0 failed=1

[root@controller-1 playbooks]# echo $?
0

[ansibot]

Files identified in the description:
None

If these files are inaccurate, please update the component name section of the description or use the !component bot command.

click here for bot help

[ansibot]

Files identified in the description:

• lib/ansible/playbook/block.py

If these files are inaccurate, please update the component name section of the description or use the !component bot command.

click here for bot help

[sekharvajjula]

Any updates on this ticket?

[sbernhard]

I agree. Looks like that BLOCK .. RESCUE is not working completely. Had the same issue with the junit callback which marks such BLOCK .. RESCUE as a failed test - which it shouldn't be in my eyes.

[DownRangeDevOps]

This makes the block: ... rescue: syntax useless when running playbooks from Jenkins. The Jenkins job will still fail because it recognizes that failed tasks is > 0.

[samdoran]

Added in #48418 is the number of rescued tasks in the play summary.