Spark ERROR client.TransportResponseHandler: Still have 1 requests outstanding when connection from

The spark task reports an error (though the calculated result is correct).

21/04/29 14:38:38 ERROR client.TransportResponseHandler: Still have 1 requests outstanding when connection from /172.16.4.156:37528 is closed
21/04/29 14:38:39 ERROR cluster.YarnScheduler: Lost executor 3 on cdh-slave1.test.com: Container marked as failed: container_1618397085347_0529_01_000004 on host: cdh-slave1.test.com. Exit status: 137. Diagnostics: [2021-04-29 14:38:39.112]Container killed on request. Exit code is 137
[2021-04-29 14:38:39.117]Container exited with a non-zero exit code 137.
[2021-04-29 14:38:39.119]Killed by external signal

Looking at the code, I found that hive is not used in the spark program, but hive support is set during spark initialization.

The program works fine after deleting this line of code.

Read More: