Skip to content

HDDS-15657. excludeContainersDueToFailure resets every iteration, repeatedly failing containers keep being retried #10892

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
sravani-revuri wants to merge 1 commit into
base: master
Choose a base branch
from
+52 −4
Open

HDDS-15657. excludeContainersDueToFailure resets every iteration, repeatedly failing containers keep being retried #10892

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 14 additions & 2 deletions ...ava/org/apache/hadoop/hdds/scm/container/balancer/ContainerBalancerSelectionCriteria.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,7 @@ public class ContainerBalancerSelectionCriteria {
private Map<ContainerID, DatanodeDetails> containerToSourceMap;
private Set<ContainerID> excludeContainers;
private Set<ContainerID> includeContainers;
private Set<ContainerID> excludeContainersDueToFailure;
private final Set<ContainerID> excludeContainersDueToFailure;
private FindSourceStrategy findSourceStrategy;
private Map<DatanodeDetails, NavigableSet<ContainerID>> setMap;

Expand All @@ -66,12 +66,24 @@ public ContainerBalancerSelectionCriteria(
ContainerManager containerManager,
FindSourceStrategy findSourceStrategy,
Map<ContainerID, DatanodeDetails> containerToSourceMap) {
this(balancerConfiguration, nodeManager, replicationManager, containerManager,
findSourceStrategy, containerToSourceMap, new HashSet<>());
}

public ContainerBalancerSelectionCriteria(
ContainerBalancerConfiguration balancerConfiguration,
NodeManager nodeManager,
ReplicationManager replicationManager,
ContainerManager containerManager,
FindSourceStrategy findSourceStrategy,
Map<ContainerID, DatanodeDetails> containerToSourceMap,
Set<ContainerID> excludeContainersDueToFailure) {
this.balancerConfiguration = balancerConfiguration;
this.nodeManager = nodeManager;
this.replicationManager = replicationManager;
this.containerManager = containerManager;
this.containerToSourceMap = containerToSourceMap;
excludeContainersDueToFailure = new HashSet<>();
this.excludeContainersDueToFailure = excludeContainersDueToFailure;
excludeContainers = balancerConfiguration.getExcludeContainers();
includeContainers = balancerConfiguration.getIncludeContainers();
this.findSourceStrategy = findSourceStrategy;
Expand Down
3 changes: 2 additions & 1 deletion ...cm/src/main/java/org/apache/hadoop/hdds/scm/container/balancer/ContainerBalancerTask.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -108,6 +108,7 @@ public class ContainerBalancerTask implements Runnable {

private Set<DatanodeDetails> selectedTargets;
private Set<DatanodeDetails> selectedSources;
private final Set<ContainerID> excludeContainersDueToFailure = new HashSet<>();
private FindTargetStrategy findTargetStrategy;
private FindSourceStrategy findSourceStrategy;
private Map<ContainerMoveSelection, CompletableFuture<MoveManager.MoveResult>>
Expand Down Expand Up @@ -635,7 +636,7 @@ private boolean initializeIteration() {

selectionCriteria = new ContainerBalancerSelectionCriteria(config,
nodeManager, replicationManager, containerManager, findSourceStrategy,
containerToSourceMap);
containerToSourceMap, excludeContainersDueToFailure);

@chihsuan chihsuan Jul 29, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

REPLICATION_NOT_HEALTHY_AFTER_MOVE may be temporary or target-specific. Would it make sense to persist only the ContainerNotFoundException exclusions across iterations, and keep the transient ones iteration-scoped?

return true;
}

Expand Down
37 changes: 36 additions & 1 deletion ...rc/test/java/org/apache/hadoop/hdds/scm/container/balancer/TestContainerBalancerTask.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -489,7 +489,42 @@ public void testConcurrentMoveCallbacksAccumulateMovedBytesAtomically() throws E
moveCompletionExecutor.shutdownNow();
}
}


@Test
public void testExcludeContainersDueToFailurePersistsAcrossIterations() throws Exception {
when(moveManager.move(any(ContainerID.class), any(DatanodeDetails.class),
any(DatanodeDetails.class)))
.thenReturn(CompletableFuture.completedFuture(
MoveManager.MoveResult.REPLICATION_NOT_HEALTHY_AFTER_MOVE))
.thenReturn(CompletableFuture.completedFuture(MoveManager.MoveResult.COMPLETED));

balancerConfiguration.setThreshold(10);
balancerConfiguration.setIterations(2);
balancerConfiguration.setBalancingInterval(0);
balancerConfiguration.setMaxSizeEnteringTarget(10 * STORAGE_UNIT);
balancerConfiguration.setMaxSizeToMovePerIteration(100 * STORAGE_UNIT);
balancerConfiguration.setMaxDatanodesPercentageToInvolvePerIteration(100);
String includeNodes = nodesInCluster.get(0).getDatanodeDetails().getHostName() + "," +
nodesInCluster.get(nodesInCluster.size() - 1).getDatanodeDetails().getHostName();
balancerConfiguration.setIncludeNodes(includeNodes);

startBalancer(balancerConfiguration);

ArgumentCaptor<ContainerID> containerCaptor = ArgumentCaptor.forClass(ContainerID.class);
verify(moveManager, atLeast(1)).move(containerCaptor.capture(),
any(DatanodeDetails.class), any(DatanodeDetails.class));
ContainerID failedContainerId = containerCaptor.getAllValues().get(0);

assertTrue(containerBalancerTask.getSelectionCriteria()
.getExcludeDueToFailContainers().contains(failedContainerId));

long failedContainerMoveAttempts = containerCaptor.getAllValues().stream()
.filter(id -> id.equals(failedContainerId))
.count();
assertEquals(1, failedContainerMoveAttempts,
"Permanently failed container should not be retried in later iterations");
}

/**
* Generates a range of equally spaced utilization(that is, used / capacity)
* values from 0 to 1.
Expand Down