zhuzhurk commented on a change in pull request #11857:
URL: https://github.com/apache/flink/pull/11857#discussion_r414264754
##########
File path:
flink-runtime/src/main/java/org/apache/flink/runtime/executiongraph/failover/flip1/RestartPipelinedRegionFailoverStrategy.java
##########
@@ -84,34 +77,8 @@ public RestartPipelinedRegionFailoverStrategy(
ResultPartitionAvailabilityChecker
resultPartitionAvailabilityChecker) {
this.topology = checkNotNull(topology);
- this.regions = Collections.newSetFromMap(new
IdentityHashMap<>());
- this.vertexToRegionMap = new HashMap<>();
this.resultPartitionAvailabilityChecker = new
RegionFailoverResultPartitionAvailabilityChecker(
resultPartitionAvailabilityChecker);
-
- // build regions based on the given topology
- LOG.info("Start building failover regions.");
- buildFailoverRegions();
- }
- //
------------------------------------------------------------------------
- // region building
- //
------------------------------------------------------------------------
-
- private void buildFailoverRegions() {
- final Set<Set<SchedulingExecutionVertex>> distinctRegions =
-
PipelinedRegionComputeUtil.computePipelinedRegions(topology);
-
- // creating all the failover regions and register them
- for (Set<SchedulingExecutionVertex> regionVertices :
distinctRegions) {
- LOG.debug("Creating a failover region with {}
vertices.", regionVertices.size());
- final FailoverRegion failoverRegion = new
FailoverRegion(regionVertices);
- regions.add(failoverRegion);
- for (SchedulingExecutionVertex vertex : regionVertices)
{
- vertexToRegionMap.put(vertex.getId(),
failoverRegion);
- }
- }
-
- LOG.info("Created {} failover regions.", regions.size());
Review comment:
Previously it is INFO level. And I'd still prefer to make it INFO.
Regions are basic components now but users can get little information on that
at the moment. This log can help to verify if regions are generated as
expected, especially there are multiple GlobalDataExchangeModes which can lead
to different region building results.
Also it is a short and one time log so it would not mess up the logs or
cause any performance issue.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
