wanglijie95 commented on a change in pull request #18050:
URL: https://github.com/apache/flink/pull/18050#discussion_r771171607
##########
File path:
flink-runtime/src/test/java/org/apache/flink/runtime/executiongraph/IntermediateResultPartitionTest.java
##########
@@ -140,6 +150,127 @@ public void testBlockingPartitionResetting() throws
Exception {
assertFalse(consumedPartitionGroup.areAllPartitionsFinished());
}
+ @Test
+ public void testGetNumberOfSubpartitionsForNonDynamicAllToAllGraph()
throws Exception {
+ testGetNumberOfSubpartitions(7, DistributionPattern.ALL_TO_ALL, false,
Arrays.asList(7, 7));
+ }
+
+ @Test
+ public void testGetNumberOfSubpartitionsForNonDynamicPointWiseGraph()
throws Exception {
+ testGetNumberOfSubpartitions(7, DistributionPattern.POINTWISE, false,
Arrays.asList(4, 3));
+ }
+
+ @Test
+ public void
testGetNumberOfSubpartitionsFromConsumerParallelismForDynamicAllToAllGraph()
+ throws Exception {
+ testGetNumberOfSubpartitions(7, DistributionPattern.ALL_TO_ALL, true,
Arrays.asList(7, 7));
Review comment:
Yes, it exists, users can set parallelism of a certain vertex. The
adaptive batch scheduler only decide parallelism for vertices whose parallelim
is not set by user(the vertex's parallelism in jobgraph is -1). For the
vertices whose parallelism is set by user or set in compile phase (the vertex's
parallelism in jobgraph is > 0), adaptive batch scheduler will not change it's
parallelism.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@flink.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
