Hello,
We are seeing that client connections to our database cluster (11.2.0.1 RAC database) do not get released when there is an unexpected failure to the instance. For example, we had an instance abort due to an OS problem and client(s) are still waiting in an OCI library nearly 24 hours after the instance died. There are obviously no active sessions to the cluster from the client machine in question, yet attaching a python debugger showed it was waiting in OCI.
Clients are using python 2.7.6, cx_Oracle 5.1.2, and Oracle client 11.2.0.3.
My question is whether or not there is a timeout setting or some other client configuration that we are missing that is causing this seemingly endless loop. Again, the server process is clearly long gone, yet the client is still waiting for a response (or something) and has been for many hours.
Thank you in advance.
Todd