OpenAM plugin for Apache reverse-proxy fails after sometime

Question

We have OpenAM plugin for Apache Http server. Here Apache Http server works as reverse-proxy. OpenAM plugin validates user requests for authentication and authorisation then forwards the requests to service. This works fine for some days. But suddenly, OpenAM plugin fails to authenticate and authorise and requests directly goes to service. Only fix is to restart the Apache HTTP server.

But this is not the correct fix. We would like to find the real route cause of this problem and fix. There is no logging reported from http server.

Thanks.

Answer 1

There is no logs reported. Totally clueless. It is happening once in two days or >so.

You should set the Agent level to message and check the debug.log (in agent root --> instances --> debu --> debug.log) and trace the requests in the agent log files that are not working. There may also be relevant information in the system logs in the agent root log folders.

But suddenly, OpenAM plugin fails to authenticate and authorise and requests directly goes to service.

What HTTP status code is returned during this period? By default, if the Agent cannot communicate to AM for policy advice (unless it has preconfigured fallback not enforced URLs) will be to 403 each request. You'd really need to check your debug and apache access logs to look at these requests hitting the reverse proxy (or whether they hit them at all).