Aerospike Java client TIMEOUT error propagation when cancellation is requested

Direct Answer

Calling Future.cancel(true) on an asynchronous Aerospike operation does not inherently abort ongoing retries. The client treats the resulting thread interruption as a timeout error (AerospikeResultCode.TIMEOUT). Whether further retries occur depends entirely on your configured retry policy: if retryOnTimeout is set to true, the client may continue retrying until maxRetries is reached, regardless of the cancellation request. The client does not provide a distinct error code to differentiate between a socket timeout and a total timeout when a cancellation is requested.

Likely Explanation vs. Confirmed Facts

Confirmed Behavior

• Error Propagation: Synchronous socket timeouts and asynchronous cancellations both propagate as an AerospikeException with resultCode == AerospikeResultCode.TIMEOUT (error code -1) and the message "Timeout".
• Cancellation Mechanism: Future.cancel(true) attempts to interrupt the Netty worker thread. This is a best-effort operation and does not guarantee that the in-flight request is aborted on the server side.
• Retry Logic: The cancellation signal does not reset or disable the retry policy. The operation is treated as a non-retryable timeout only if the policy's retryOnTimeout flag is false.

Assumptions and Uncertainties

It is assumed that the behavior observed is consistent with the standard Netty-based asynchronous implementation of the Java client. However, because thread interruption is best-effort, the exact timing of when the TIMEOUT error is thrown relative to server processing can vary under heavy load, potentially leading to inconsistent observations of "fatal" vs "retryable" behavior.

Verification Steps

1. Enable DEBUG level logging for the Aerospike client to trace thread interruptions and retry attempts.
2. Execute an asynchronous operation with a very short socketTimeout (e.g., 10ms) while simulating server delay.
3. Invoke Future.cancel(true) on the returned future.
4. Catch the AerospikeException and inspect exception.getCause() to determine if the underlying trigger was a java.util.concurrent.TimeoutException or a Netty-specific timeout.
5. Monitor AerospikeMetrics (specifically retry_count and timeout_count) to verify if a retry was triggered despite the cancellation.

Missing Diagnostic Detail

To determine if your specific configuration treats cancellation as fatal, please provide the current value of retryOnTimeout in your WritePolicy or ReadPolicy. If this is true, the client will not treat cancellation as a fatal error that prevents further retries.