Jae Park [X] - It looks like you are executing a synchronous subdoc lookup operation and expecting it to fast-fail once the operation timeout has been reached, is that correct?
If yes, I'm not sure this test can work as you expect right now. The reason for this is the SDK sends the operation to the cluster, using the default connection pool timeout of 15 seconds. Once a response is received, or a timeout occurs, the subdoc operations is checked to see if a timeout occurred (2.5 seconds). Below are the default timeouts for an operation and connection pool send timeout.
Default operation timeout: 2500ms
Default connection pool send timeout: 15000ms
Jeff Morris - Do you think this constitutes a bug? KV operations do not fast-fast operation timeout but instead use the connection pool send timeout (which by default is much larger) and check for operation afterwards.