Questions regarding the submission of jar files

[shashihoushengqia]

I found that every time I submitted a Spark on K8s task through Kyuubi, a file named "kyuubi-spark-sql-engine_2.12-1.10.3.jar" would be uploaded to S3/HDFS. This would occupy a large amount of space on S3/HDFS, and uploading it every time would also slow down the startup speed of Spark. So I placed this jar file in /opt/spark/jars/ in all the Spark containers, and also in /opt/kyuubi/jars in the Kyuubi container. However, Kyuubi still uploads this jar file to S3/HDFS. So, experts, how should we solve this problem?