Skip to content

Remove some printed errMsg when stop DN #18799

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

Merged
JackieTien97 merged 3 commits into from
Oct 10, 2026
Merged

Remove some printed errMsg when stop DN #18799

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
dcf5a92
fix some
Wei-hao-Li Oct 8, 2026
330a85a
fix some
Wei-hao-Li Oct 10, 2026
2b01d11
Merge branch 'master' of github.com:apache/iotdb into fixStopDN
Wei-hao-Li Oct 10, 2026
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
11 changes: 8 additions & 3 deletions ...in/java/org/apache/iotdb/db/queryengine/execution/fragment/FragmentInstanceExecution.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@
package org.apache.iotdb.db.queryengine.execution.fragment;

import org.apache.iotdb.calc.exception.MemoryNotEnoughException;
import org.apache.iotdb.commons.conf.CommonDescriptor;
import org.apache.iotdb.commons.utils.FileUtils;
import org.apache.iotdb.db.conf.IoTDBConfig;
import org.apache.iotdb.db.conf.IoTDBDescriptor;
Expand Down Expand Up @@ -332,7 +333,9 @@ private void initialize(IDriverScheduler scheduler, boolean isExplainAnalyze) {
try {
clearShuffleSinkHandle(newState);
} catch (Throwable t) {
LOGGER.error(DataNodeQueryMessages.ERRORS_RELEASING_SINK, t);
if (!CommonDescriptor.getInstance().getConfig().isStopping()) {
LOGGER.error(DataNodeQueryMessages.ERRORS_RELEASING_SINK, t);
}
}

// close the driver after sink is aborted or closed because in driver.close() it
Expand All @@ -358,8 +361,10 @@ private void initialize(IDriverScheduler scheduler, boolean isExplainAnalyze) {
exchangeManager.deRegisterFragmentInstanceFromMemoryPool(
instanceId.getQueryId().getId(), instanceId.getFragmentInstanceId(), true);
} catch (Throwable t) {
LOGGER.error(
DataNodeQueryMessages.ERRORS_DEREGISTER_FI_FROM_MEMORY_POOL, newState, t);
if (!CommonDescriptor.getInstance().getConfig().isStopping()) {
LOGGER.error(
DataNodeQueryMessages.ERRORS_DEREGISTER_FI_FROM_MEMORY_POOL, newState, t);
}
}

try {
Expand Down
9 changes: 7 additions & 2 deletions ...ode/src/main/java/org/apache/iotdb/db/queryengine/execution/schedule/DriverScheduler.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,7 @@
import org.apache.iotdb.calc.execution.schedule.queue.IndexedBlockingQueue;
import org.apache.iotdb.calc.execution.schedule.queue.IndexedBlockingReserveQueue;
import org.apache.iotdb.commons.concurrent.ThreadName;
import org.apache.iotdb.commons.conf.CommonDescriptor;
import org.apache.iotdb.commons.conf.IoTDBConstant;
import org.apache.iotdb.commons.exception.StartupException;
import org.apache.iotdb.commons.queryengine.common.SessionInfo;
Expand Down Expand Up @@ -394,7 +395,9 @@ private void clearDriverTask(DriverTask task) {
try {
task.getDriver().failed(task.getAbortCause().get());
} catch (Exception e) {
logger.error(DataNodeQueryMessages.CLEAR_DRIVERTASK_FAILED, e);
if (!CommonDescriptor.getInstance().getConfig().isStopping()) {
logger.error(DataNodeQueryMessages.CLEAR_DRIVERTASK_FAILED, e);
}
}
}
if (task.getStatus() == DriverTaskStatus.ABORTED) {
Expand All @@ -405,7 +408,9 @@ private void clearDriverTask(DriverTask task) {
task.getDriverTaskId().getFragmentId().getId(),
task.getDriverTaskId().getFragmentInstanceId().getInstanceId()));
} catch (Exception e) {
logger.error(DataNodeQueryMessages.CLEAR_DRIVERTASK_FAILED, e);
if (!CommonDescriptor.getInstance().getConfig().isStopping()) {
logger.error(DataNodeQueryMessages.CLEAR_DRIVERTASK_FAILED, e);
}
}
}
} finally {
Expand Down
2 changes: 1 addition & 1 deletion iotdb-core/datanode/src/main/java/org/apache/iotdb/db/service/DataNodeShutdownHook.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -112,11 +112,11 @@ public void run() {
DNAuditLogger.getInstance().log(fields, () -> logMessage);

startWatcher();
CommonDescriptor.getInstance().getConfig().setStopping(true);
// Stop external rpc service firstly.
ExternalRPCService.getInstance().stop();

// Reject write operations to make sure all tsfiles will be sealed
CommonDescriptor.getInstance().getConfig().setStopping(true);
CommonDescriptor.getInstance()
.getConfig()
.setNodeStatusWithReason(NodeStatus.ReadOnly, NodeStatus.STOPPING);
Expand Down
80 changes: 80 additions & 0 deletions ...de/src/test/java/org/apache/iotdb/db/queryengine/execution/exchange/SourceHandleTest.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,7 @@
import org.apache.iotdb.commons.client.sync.SyncDataNodeMPPDataExchangeServiceClient;
import org.apache.iotdb.commons.memory.MemoryManager;
import org.apache.iotdb.db.conf.IoTDBDescriptor;
import org.apache.iotdb.db.exception.runtime.MemoryLeakException;
import org.apache.iotdb.db.queryengine.common.FragmentInstanceId;
import org.apache.iotdb.db.queryengine.execution.exchange.MPPDataExchangeManager.SourceHandleListener;
import org.apache.iotdb.db.queryengine.execution.exchange.source.SourceHandle;
Expand All @@ -47,10 +48,12 @@
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Future;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicInteger;
import java.util.concurrent.atomic.AtomicReference;
import java.util.stream.Collectors;
import java.util.stream.Stream;

Expand Down Expand Up @@ -83,6 +86,83 @@ public void testAbortStopsFragmentFetchAfterInFlightRpc() throws Exception {
testCancellationStopsFragmentFetchAfterInFlightRpc(true);
}

@Test
public void reproduceAbortAfterFragmentInstanceMemoryIsForceDeregistered() throws Exception {
final String queryId = "q0";
final String fragmentInstanceId = "0.0";
final String localPlanNodeId = "exchange_0";
final TEndPoint remoteEndpoint =
new TEndPoint("remote", IoTDBDescriptor.getInstance().getConfig().getMppDataExchangePort());
final TFragmentInstanceId remoteFragmentInstanceId = new TFragmentInstanceId(queryId, 1, "0");
final TFragmentInstanceId localFragmentInstanceId = new TFragmentInstanceId(queryId, 0, "0");
final MemoryPool memoryPool =
new MemoryPool("test", new MemoryManager(2 * MOCK_TSBLOCK_SIZE), 2 * MOCK_TSBLOCK_SIZE);
final LocalMemoryManager localMemoryManager = Mockito.mock(LocalMemoryManager.class);
Mockito.when(localMemoryManager.getQueryPool()).thenReturn(memoryPool);
final ExecutorService executorService = Mockito.mock(ExecutorService.class);
final SourceHandle sourceHandle =
new SourceHandle(
remoteEndpoint,
remoteFragmentInstanceId,
localFragmentInstanceId,
localPlanNodeId,
0,
localMemoryManager,
executorService,
Utils.createMockTsBlockSerde(MOCK_TSBLOCK_SIZE),
Mockito.mock(SourceHandleListener.class),
Mockito.mock(IClientManager.class));

sourceHandle.isBlocked();
sourceHandle.updatePendingDataBlockInfo(0, List.of(MOCK_TSBLOCK_SIZE));
Assert.assertEquals(MOCK_TSBLOCK_SIZE, sourceHandle.getBufferRetainedSizeInBytes());
Assert.assertEquals(MOCK_TSBLOCK_SIZE, memoryPool.getReservedBytes());

final CountDownLatch memoryDeregistered = new CountDownLatch(1);
final AtomicReference<Throwable> deregisterFailure = new AtomicReference<>();
final AtomicReference<Throwable> abortFailure = new AtomicReference<>();
final Thread fragmentInstanceCleanupThread =
new Thread(
() -> {
try {
memoryPool.deRegisterFragmentInstanceFromQueryMemoryMap(
queryId, fragmentInstanceId, true);
} catch (Throwable t) {
deregisterFailure.set(t);
} finally {
memoryDeregistered.countDown();
}
});
final Thread driverSchedulerCleanupThread =
new Thread(
() -> {
try {
if (!memoryDeregistered.await(10, TimeUnit.SECONDS)) {
abortFailure.set(
new AssertionError("Timed out waiting for FI memory deregistration"));
return;
}
sourceHandle.abort();
} catch (Throwable t) {
abortFailure.set(t);
}
});

fragmentInstanceCleanupThread.start();
driverSchedulerCleanupThread.start();
fragmentInstanceCleanupThread.join(TimeUnit.SECONDS.toMillis(10));
driverSchedulerCleanupThread.join(TimeUnit.SECONDS.toMillis(10));

Assert.assertFalse(fragmentInstanceCleanupThread.isAlive());
Assert.assertFalse(driverSchedulerCleanupThread.isAlive());
Assert.assertTrue(deregisterFailure.get() instanceof MemoryLeakException);
Assert.assertTrue(abortFailure.get() instanceof IllegalArgumentException);
Assert.assertEquals(
"RelatedMemoryReserved can't be null when freeing memory", abortFailure.get().getMessage());
Assert.assertEquals(MOCK_TSBLOCK_SIZE, sourceHandle.getBufferRetainedSizeInBytes());
Assert.assertEquals(MOCK_TSBLOCK_SIZE, memoryPool.getReservedBytes());
}

private void testCancellationStopsFragmentFetchAfterInFlightRpc(boolean abort) throws Exception {
final String queryId = "q0";
final TEndPoint remoteEndpoint =
Expand Down
Loading