Environment

• Covalent version: 0.216.0rc0
• Covalent-SSH plugin version: 0.20.0
• Python version: 3.8.15
• Operating system: MacOS (Darwin Kernel Version 21.6.0)

What is happening?

The following code block, starting at line 501, checks for a non-empty stderr, else it assumes task failure:

        if result_err := result.stderr.strip():
            app_log.warning(result_err)
            return self._on_ssh_fail(function, args, kwargs, result_err)

However, it does not distinguish warnings from errors. Warning messages will therefore result in a "failed" task.

How can we reproduce the issue?

Issue a warning message inside any task that uses the SSHExecutor.

What should happen?

The task should be marked as failed only if the returncode is non-zero - as opposed to assuming this if stderr is not empty.

Any suggestions?

Log the message all the same, but check the returncode before deferring to _on_ssh_fail:

         if result_err := result.stderr.strip():
             app_log.warning(result_err)
             if result.returncode != 0:
                 return self._on_ssh_fail(function, args, kwargs, result_err)