gazelle failed to run with spark local

[jackylee-ch]

If we run with spark local, ExecutorManager.scala#L46 will throw index out of range exception.

This is because there is no YarnCoarseGrainedExecutorBackend in our environment, thus we use Random.nextInt to get random index.

[zhouyuan]

@stczwd Thanks for reporting! the numa binding feature is not so robust, will look into this

[jackylee-ch]

@stczwd Thanks for reporting! the numa binding feature is not so robust, will look into this

Thanks for your attension. spark.oap.sql.columnar.numaBinding=false worked with me. I have change the code with some style changes, please check it again, thanks