Created attachment 1718161 [details]
0-failure.log

Description of problem:
receptor-satellite allows inconsistent combination of status codes. There are the following codes in the message from receptor to c.rh.c:

satellite_connection_code (set wrong password in receptor.conf): 0 if connection between Receptor and Satellite is successful, non-zero otherwise
connection_code (delete foreman-proxy's ssh key from host's authorized_keys): 0 if connection between Satellite and the host is successful, non-zero otherwise
execution_code (change binary used in remediation to 'exit 1'): 0 if playbook finished successfully, non-zero otherwise
satellite_infrastructure_code (not tested): should return 0 when connection between satellite and connection is successful, non-zero otherwise

The one weird thing I encountered during testing is that if connection between Receptor and Satellite fails (I actually set wrong password in receptor.conf), then satellite_connection_code!=0 and connection_code==0. Connection between Satellite and host could, however, never have succeeded because Satellite wasn't instructed to do anything in the first place. Log of this case attached.
It doesn't make much sense to me, in combination with status==failure. Also, c.rh.c team confirmed they would prefer something that is more consistent, like connection_code==null.


Version-Release number of selected component (if applicable):
Sat 6.8. Not a regression - not supported in previous versions.

How reproducible:
Deterministic

Steps to Reproduce:
1. Setup receptor and host, upload inventory
2. Create remediation
3. Set wrong password in receptor.conf
4. Try to apply the remediation

Actual results:
Error, satellite_connection_code!=0 and connection_code==0

Expected results:
Error, satellite_connection_code!=0 and connection_code==null