Client errors
Client errors indicate that a problem occurred within the driver.
Driver timeout
All drivers can throw a driver timeout that indicates a timeout occurred when executing a CQL statement.
The name of this error varies by driver.
For example, the Java driver throws DriverTimeoutException, and the Python driver throws OperationTimedOut.
- Cause
-
A driver timeout means the client-side timeout limit was reached before the driver received a response from the server-side coordinator. For example, a driver timeout can occur when a read query requests a large amount of data or when batch write sizes are large or span multiple partitions.
- Resolution
-
To avoid driver timeouts, make sure the value of the client-side timeout is greater than the server-side read/write timeouts plus network latency:
x = max(Server read timeout,Server write timeout) Client timeout > (x + Network latency)The client side timeout is commonly configured in an execution profile or on the query execution method.
When reading or writing large amounts of data, avoid timeouts by making queries smaller and more efficient. For example, you can fetch large partitions in batches and execute queries asynchronously.
- Retry policy
-
When a driver timeout occurs, you might not know if the data was modified. The requests might have been partially or fully processed by the server, but the client didn’t receive a response. While driver timeouts aren’t a subject for retry policies, you can implement custom logic based on these exceptions. For more information, see Cassandra driver retry policies and Configure request tracing.
Other client errors
For other possible client errors, see your driver’s documentation.
