If you have cases where you find this to be a breaking change, I would like to hear them. In most cases, we silently switch to TCP/IP. There are a few cases, likely fringe cases, where we know some minor configuration changes will be required. Assuming client & server are running as different users, then the following cases are known to require configuration changes:

1. The server explicitly starts only shmem via "-x shmem" or "-x none" on the command line. In this case, the server will not start a tcpip port and fall-back will not be available. Reconfigure the server by either removing the "-x" parameter or explicitly adding the tcpip port along with the desired tcpip configuration parameters.
2. The client specifies commlinks=shmem or commlinks=none: In this case, fall-back will not be available. Reconfigure the connection string to remove the commlinks parameter or change it to specify allow tcpip along with the desired tcpip configuration parameters.
3. The server uses "-ec TLS..." to require TLS communications and also specifies "-es" to allow non-encrypted shmem communications. Since tcpip will be used as a fall-back in this scenario, the local client will need to connect using TLS via tcpip and will need to provide a valid certificate.