aoscx_config is not failing in case of configuration error

[ssasso]

In case aoscx_config is supplied a config file with the src param, the playbook execution continues even if the source config file has invalid configuration in it.

I would have expected a module failure in case of invalid config, but this is not happening.

[ssasso]

Quick & dirty proposal(s):

The config is sent to the device here:

aoscx-ansible-collection/plugins/modules/aoscx_config.py

Line 492 in 9315901

load_config(module, commands)

and

https://github.com/aruba/aoscx-ansible-collection/blob/master/plugins/module_utils/aoscx.py#L124

Option (a):

It would be possible to return from load_config the stdout/stderr of every command invocation (out and err) and then:
if any of out (or err) contains the string ^Invalid input: (which is returned by the CLI in case of config command error) return an error instead of success - or, at least, return all the stdout/stderr to the user as a module output, and let the user decide what to do.

Option (b):

Directly edit the load_config to trigger an error (returing also the content of out - which will be useful to understand the broken command) in case it contains ^Invalid input: for a command.

Example (not tested nor validated):

        try:
            out = exec_command(module, command)
        except ConnectionError as exc:
            module.fail_json(
                msg="unable to enter configuration mode",
                err=to_text(exc, errors="surrogate_then_replace"),
            )
        if out.startswith("Invalid input:"):
            module.fail_json(
                msg="unable to send command",
                err=to_text(out, errors="surrogate_then_replace"),
            )

[tchiapuziowong]

hi @ssasso! at this time the collection does not support CLI validation before execution - this will simply execute a CLI command and then expect to return to the prompt # after - I can include this to be an enhancement to be considered for future releases

[ssasso]

Hi @tchiapuziowong!

No need for pre-validation before execution, but it would be important IMHO to return errors when the CLI reports errors (intercept cli errors)
(Like the code example above)