Normal Description
Team, is there any way we can increase the TCP idle timeout value for Windows Container? As per the customer, the pod running on Windows Node is connecting to a database external to the cluster but in the same network where sometimes the Database takes 20+ minutes to respond with results and the client closes the TCP connection i.e. Windows Container.
I am sharing the exact update provided by the customer.
In our Openshift cluster we have a job running on Windows worker node/container which connects to an external Oracle database (running outside of the Openshift cluster, but in the same L2 network as the worker nodes) and if the operation on the Oracle database (select/update/etc) takes multiple minutes to complete on the database the TCP connection on the client side appears to be dropped (by containerd/overlay networking ?). So when Oracle tries to respond with results (after cca 10-20 minutes) it cannot as the client side already closed the TCP connection.
Is it possible to configure a TCP idle connection timeout on containerd/overlay networking on the windows worker node ?
