Orphaned Files on Job Failure Break Future Jobs on self-hosted runners

[exp-hc]

When an error occurs during execution, a collection is orphaned and breaks future runs until cleaned up manually. This is on a self-hosted runner which works fine until an error occurs.

The following snippet is from the failed job:

Run echo ::group::Extracting the collection metadata from "'galaxy.yml'"...
Extracting the collection metadata from 'galaxy.yml'...
Run import yaml
/home/github/actions-runner/_work/_temp/01dd8358-3030-4067-9a41-b4b8a3b66d60.py:5: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.
  collection_meta = yaml.load(galaxy_yml)
Run echo ::endgroup::

Run echo ::group::Move "'expedient.smc'" collection to ansible_collections/expedient/smc...
Move 'expedient.smc' collection to ansible_collections/expedient/smc...
Run set -x ; mkdir -pv "ansible_collections/expedient" ; mv -v ".tmp-ansible-collection-checkout" "ansible_collections/expedient/smc" ; set +x
+ mkdir -pv ansible_collections/expedient
+ mv -v .tmp-ansible-collection-checkout ansible_collections/expedient/smc
mv: cannot move '.tmp-ansible-collection-checkout' to 'ansible_collections/expedient/smc/.tmp-ansible-collection-checkout': Directory not empty
Error: Process completed with exit code 1.

Here is the original failure leading to the orphaned files

Installing collected packages: ansible-core
    Running setup.py install for ansible-core: started
    ERROR: Command errored out with exit status 1:
    Running setup.py install for ansible-core: finished with status 'error'
     command: /home/github/actions-runner/_work/_tool/Python/3.9.9/x64/bin/python -u -c 'import io, os, sys, setuptools, tokenize; sys.argv[0] = '"'"'/tmp/pip-req-build-042atxsa/setup.py'"'"'; __file__='"'"'/tmp/pip-req-build-042atxsa/setup.py'"'"';f = getattr(tokenize, '"'"'open'"'"', open)(__file__) if os.path.exists(__file__) else io.StringIO('"'"'from setuptools import setup; setup()'"'"');code = f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' install --record /tmp/pip-record-anvypspo/install-record.txt --single-version-externally-managed --user --prefix= --compile --install-headers /home/github/.local/include/python3.9/ansible-core
         cwd: /tmp/pip-req-build-042atxsa/
    Complete output (27 lines):
    Traceback (most recent call last):
      File "<string>", line 1, in <module>
      File "/tmp/pip-req-build-042atxsa/setup.py", line 120, in <module>
        _validate_install_ansible_core()
      File "/tmp/pip-req-build-042atxsa/setup.py", line 91, in _validate_install_ansible_core
        raise RuntimeError(
    RuntimeError:
    
        ****************************************************************************
    
        Cannot install ansible-core with a pre-existing ansible-base==2.10.15.post0
        installation.
    
        Installing ansible-core with ansible-2.9 or older, or ansible-base-2.10
        currently installed with pip is known to cause problems. Please uninstall
        ansible-base and install the new version:
    
            pip uninstall ansible-base
            pip install ansible-core
    
        If you want to skip the conflict checks and manually resolve any issues
        afterwards, set the ANSIBLE_SKIP_CONFLICT_CHECK environment variable:
    
            ANSIBLE_SKIP_CONFLICT_CHECK=1 pip install ansible-core
    
        ****************************************************************************
    
    ----------------------------------------
ERROR: Command errored out with exit status 1: /home/github/actions-runner/_work/_tool/Python/3.9.9/x64/bin/python -u -c 'import io, os, sys, setuptools, tokenize; sys.argv[0] = '"'"'/tmp/pip-req-build-042atxsa/setup.py'"'"'; __file__='"'"'/tmp/pip-req-build-042atxsa/setup.py'"'"';f = getattr(tokenize, '"'"'open'"'"', open)(__file__) if os.path.exists(__file__) else io.StringIO('"'"'from setuptools import setup; setup()'"'"');code = f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' install --record /tmp/pip-record-anvypspo/install-record.txt --single-version-externally-managed --user --prefix= --compile --install-headers /home/github/.local/include/python3.9/ansible-core Check the logs for full command output.
Error: Process completed with exit code 1.

[webknjaz]

@exp-hc where is that install error comming from? This action? Or have you performed some of the installation procedures yourself?

[exp-hc]

@webknjaz The error was due to different version of ansible being installed locally, I'm using a self-hosted runner. Once I pip uninstalled ansible, the 'original failure' went away.

[webknjaz]

Interesting... I suppose it'd be best for us to maintain our own virtualenv, then.