[KYUUBI apache#4713][TEST] Fix false positive result in SchedulerPool…

…Suite

### _Why are the changes needed?_

fix issuse apache#4713

### _How was this patch tested?_
- [ ] Add some test cases that check the changes thoroughly including negative and positive cases if possible

- [ ] Add screenshots for manual tests if appropriate

- [X] [Run test](https://kyuubi.readthedocs.io/en/master/develop_tools/testing.html#running-tests) locally before make a pull request

Closes apache#4714 from huangzhir/fixtest-schedulerpool.

Closes apache#4713

e66ede2 [huangzhir] fixbug TEST SchedulerPoolSuite  a false positive result

Authored-by: huangzhir <306824224@qq.com>
Signed-off-by: Cheng Pan <chengpan@apache.org>

externals/kyuubi-spark-sql-engine/src/test/scala/org/apache/kyuubi/engine/spark/SchedulerPoolSuite.scala

@@ -19,6 +19,8 @@ package org.apache.kyuubi.engine.spark
 
 import java.util.concurrent.Executors
 
+import scala.concurrent.duration.SECONDS
+
 import org.apache.spark.scheduler.{SparkListener, SparkListenerJobEnd, SparkListenerJobStart}
 import org.scalatest.concurrent.PatienceConfiguration.Timeout
 import org.scalatest.time.SpanSugar.convertIntToGrainOfTime
@@ -80,6 +82,7 @@ class SchedulerPoolSuite extends WithSparkSQLEngine with HiveJDBCTestHelper {
         threads.execute(() => {
           priority match {
             case 0 =>
+              // job name job2
               withJdbcStatement() { statement =>
                 statement.execute("SET kyuubi.operation.scheduler.pool=p0")
                 statement.execute("SELECT java_method('java.lang.Thread', 'sleep', 1500l)" +
@@ -92,17 +95,18 @@ class SchedulerPoolSuite extends WithSparkSQLEngine with HiveJDBCTestHelper {
                 statement.execute("SELECT java_method('java.lang.Thread', 'sleep', 1500l)" +
                   " FROM range(1, 3, 1, 2)")
               }
+              // make sure this job name job1
+              Thread.sleep(1000)
           }
         })
       }
       threads.shutdown()
-      eventually(Timeout(20.seconds)) {
-        // We can not ensure that job1 is started before job2 so here using abs.
-        assert(Math.abs(job1StartTime - job2StartTime) < 1000)
-        // Job1 minShare is 2(total resource) so that job2 should be allocated tasks after
-        // job1 finished.
-        assert(job2FinishTime - job1FinishTime >= 1000)
-      }
+      threads.awaitTermination(20, SECONDS)
+      // because after job1 submitted, sleep 1s, so job1 should be started before job2
+      assert(job1StartTime < job2StartTime)
+      // job2 minShare is 2(total resource) so that job1 should be allocated tasks after
+      // job2 finished.
+      assert(job2FinishTime < job1FinishTime)
     } finally {
       spark.sparkContext.removeSparkListener(listener)
     }