On 5/10 @ 12:11 P.M. my AWX instance suddenly changed its own behavior. I can find no changes anywhere in the system. A job that had run hundreds of times successfully every time suddenly started failing 100% of the time. The cause of failure is that 2 variables for which the module supplies default values were suddenly only getting null values, not the necessary defaults. I was able to restore the system to function by explicitly setting those variables in my plays, but what happened? Why do the modules no longer provide the defaults?
Is this a behavior anyone else has seen?
The module is: bigip_cli_script
{
“changed”: false,
“invocation”: {
“module_args”: {
“name”: “get_oauth_client.tcl”,
“content”: “proc script::run {} {puts stdout "Ran to completion"\n}”,
“provider”: {
“server”: “10.xx.xx.198”,
“validate_certs”: false,
“server_port”: null,
“user”: “ansible”,
“password”: “VALUE_SPECIFIED_IN_NO_LOG_PARAMETER”,
“transport”: “rest”,
“timeout”: null,
“no_f5_teem”: null,
“auth_provider”: null
},
“state”: “present”,
“partition”: “Common”,
“description”: null
}
},
“ansible_facts”: {
“discovered_interpreter_python”: “/usr/bin/python3”
},
“_ansible_no_log”: null
}
The variables in the provider section, server_port and no_f5_teem, have been null for months. The module should supply the values “443” and “yes” to those two variables, but it suddenly quit doing so. When I supply those two variables to the play, it runs like a champ.
There were no project updates done at that time. The nodes did not restart. There were no changes to the project even to update. The OpenShift cluster on which it’s running did not change. I cannot think of any reason for it to behave differently.
Thank you,
Kevin