Skip to content

HDFS-17821. Fix the SNN repeatedly checkpoint after fsimage transfer failure on one of the multiple NNs #7876

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
merged 1 commit into from
Oct 22, 2025
Merged

HDFS-17821. Fix the SNN repeatedly checkpoint after fsimage transfer failure on one of the multiple NNs #7876

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
2 changes: 1 addition & 1 deletion ...oop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/ha/StandbyCheckpointer.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -342,7 +342,7 @@ public TransferFsImage.TransferResult call()
throw ie;
}

if (!ioes.isEmpty()) {
if (ioes.size() > activeNNAddresses.size() / 2) {
throw MultipleIOException.createIOException(ioes);
}
}
Expand Down
47 changes: 45 additions & 2 deletions ...-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/ha/TestStandbyCheckpoints.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -721,8 +721,51 @@ private void doCreate() throws IOException {
out.write(42);
out.close();
}



/**
* Test checkpoint still succeeds when no more than half of the fsimages upload failed.
*/
@Test
@Timeout(value = 300)
public void testPutFsimagePartFailed() throws Exception {
for (int i = 1; i < NUM_NNS; i++) {
cluster.shutdownNameNode(i);

// Make true checkpoint for DFS_NAMENODE_CHECKPOINT_PERIOD_KEY
cluster.getConfiguration(i).setInt(DFSConfigKeys.DFS_NAMENODE_CHECKPOINT_PERIOD_KEY, 3);
cluster.getConfiguration(i).setInt(DFSConfigKeys.DFS_NAMENODE_CHECKPOINT_TXNS_KEY, 1000);
}
doEdits(0, 10);
cluster.transitionToStandby(0);

for (int i = 1; i < NUM_NNS; i++) {
cluster.restartNameNode(i, false);
}
cluster.waitClusterUp();
setNNs();

for (int i = 0; i < NUM_NNS; i++) {
// Once the standby catches up, it should do a checkpoint
// and save to local directories.
HATestUtil.waitForCheckpoint(cluster, i, ImmutableList.of(12));
}

long snnCheckpointTime1 = nns[1].getNamesystem().getStandbyLastCheckpointTime();
cluster.transitionToActive(0);
cluster.transitionToObserver(2);
cluster.shutdownNameNode(2);

doEdits(11, 20);
nns[0].getRpcServer().rollEditLog();
HATestUtil.waitForCheckpoint(cluster, 0, ImmutableList.of(23));

long snnCheckpointTime2 = nns[1].getNamesystem().getStandbyLastCheckpointTime();

// Make sure that standby namenode checkpoint success and update the lastCheckpointTime
// even though it send fsimage to nn2 failed because nn2 is shut down.
assertTrue(snnCheckpointTime2 > snnCheckpointTime1);
}

/**
* A codec which just slows down the saving of the image significantly
* by sleeping a few milliseconds on every write. This makes it easy to
Expand Down