Skip to content

HDFS-15961. standby namenode failed to start ordered snapshot deletion is enabled while having snapshottable directories #2881

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 5 commits into from
Apr 27, 2021
Merged

HDFS-15961. standby namenode failed to start ordered snapshot deletion is enabled while having snapshottable directories #2881

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
d5e413d
HDFS-15961. standby namenode failed to start ordered snapshot deletio…
bshashikant Apr 9, 2021
bb6dd12
Addressed review comments.
bshashikant Apr 10, 2021
1e53a49
Fixed indentation.
bshashikant Apr 10, 2021
32c8de1
Removed the termination condition for snapshot trash root provisionin…
bshashikant Apr 20, 2021
aaeeec0
Changed exception log msg.
bshashikant Apr 23, 2021
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
39 changes: 28 additions & 11 deletions ...roject/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/FSNamesystem.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,7 +124,12 @@
import org.apache.hadoop.hdfs.server.namenode.metrics.ReplicatedBlocksMBean;
import org.apache.hadoop.hdfs.server.protocol.SlowDiskReports;
import org.apache.hadoop.ipc.ObserverRetryOnActiveException;
import org.apache.hadoop.util.*;
import org.apache.hadoop.util.Time;
import org.apache.hadoop.util.Daemon;
import org.apache.hadoop.util.DataChecksum;
import org.apache.hadoop.util.ReflectionUtils;
import org.apache.hadoop.util.StringUtils;
import org.apache.hadoop.util.VersionInfo;

import static org.apache.hadoop.util.Time.now;
import static org.apache.hadoop.util.Time.monotonicNow;
Expand Down Expand Up @@ -8559,38 +8564,50 @@ void checkAccess(String src, FsAction mode) throws IOException {
/**
* Check if snapshot roots are created for all existing snapshottable
* directories. Create them if not.
* Only the active NameNode needs to execute this in HA setup once it is out
* of safe mode.
*
* The function gets called while exiting safe mode or post starting the
* services in Active NameNode, but comes into effect post whichever event
* happens later.
*/
@Override
public void checkAndProvisionSnapshotTrashRoots() {
if (isSnapshotTrashRootEnabled) {
public synchronized void checkAndProvisionSnapshotTrashRoots() {
if (isSnapshotTrashRootEnabled && (haEnabled && inActiveState()
|| !haEnabled) && !blockManager.isInSafeMode()) {
SnapshottableDirectoryStatus dirStatus = null;
try {
SnapshottableDirectoryStatus[] dirStatusList =
getSnapshottableDirListing();
if (dirStatusList == null) {
return;
}
for (SnapshottableDirectoryStatus dirStatus : dirStatusList) {
for (SnapshottableDirectoryStatus status : dirStatusList) {
dirStatus = status;
String currDir = dirStatus.getFullPath().toString();
if (!currDir.endsWith(Path.SEPARATOR)) {
currDir += Path.SEPARATOR;
}
String trashPath = currDir + FileSystem.TRASH_PREFIX;
HdfsFileStatus fileStatus = getFileInfo(trashPath, false, false, false);
HdfsFileStatus fileStatus =
getFileInfo(trashPath, false, false, false);
if (fileStatus == null) {
LOG.info("Trash doesn't exist for snapshottable directory {}. " + "Creating trash at {}", currDir, trashPath);
LOG.info("Trash doesn't exist for snapshottable directory {}. "
+ "Creating trash at {}", currDir, trashPath);
PermissionStatus permissionStatus =
new PermissionStatus(getRemoteUser().getShortUserName(), null,
SHARED_TRASH_PERMISSION);
mkdirs(trashPath, permissionStatus, false);
}
}
} catch (IOException e) {
final String msg =
"Could not provision Trash directory for existing "
+ "snapshottable directories. Exiting Namenode.";
ExitUtil.terminate(1, msg);
if (dirStatus == null) {
LOG.error("Failed to get snapshottable directory list", e);
} else {
LOG.error("Could not provision Trash directory for existing "
+ "snapshottable directory {}", dirStatus, e);
}
}

}
}

Expand Down
1 change: 1 addition & 0 deletions ...fs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/NameNode.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2021,6 +2021,7 @@ public HAState getState() {
public void startActiveServices() throws IOException {
try {
namesystem.startActiveServices();
namesystem.checkAndProvisionSnapshotTrashRoots();
startTrashEmptier(getConf());
} catch (Throwable t) {
doImmediateShutdown(t);
Expand Down
37 changes: 37 additions & 0 deletions ...t/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/ha/TestHASafeMode.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -98,6 +98,7 @@ public void setupCluster() throws Exception {
conf.setInt(DFSConfigKeys.DFS_BLOCK_SIZE_KEY, BLOCK_SIZE);
conf.setInt(DFSConfigKeys.DFS_HEARTBEAT_INTERVAL_KEY, 1);
conf.setInt(DFSConfigKeys.DFS_HA_TAILEDITS_PERIOD_KEY, 1);
conf.setBoolean("dfs.namenode.snapshot.trashroot.enabled", false);
Copy link
Member

Choose a reason for hiding this comment

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

Can we move this config in DFSConfigKyes and use it from there, As of now I think it is defined in FsNamesystem which doesn't makes, The value of this I guess is exposed via getServerDefaults also so no point keeping it in FsN

Copy link
Contributor Author

Choose a reason for hiding this comment

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

It is intended to be an hidden config similar to ordered snapshot deletion config. I would prefer to keep it as it is now.


cluster = new MiniDFSCluster.Builder(conf)
.nnTopology(MiniDFSNNTopology.simpleHATopology())
Expand Down Expand Up @@ -909,6 +910,42 @@ public Boolean get() {
assertSafeMode(nn1, 3, 3, 3, 0);
}

@Test
public void testNameNodeCreateSnapshotTrashRootOnHASetup() throws Exception {
DistributedFileSystem dfs = cluster.getFileSystem(0);
final Path testDir = new Path("/disallowss/test2/");
final Path file0path = new Path(testDir, "file-0");
dfs.create(file0path).close();
dfs.allowSnapshot(testDir);
// .Trash won't be created right now since snapshot trash is disabled
final Path trashRoot = new Path(testDir, FileSystem.TRASH_PREFIX);
assertFalse(dfs.exists(trashRoot));
// Set dfs.namenode.snapshot.trashroot.enabled=true
cluster.getNameNode(0).getConf()
.setBoolean("dfs.namenode.snapshot.trashroot.enabled", true);
cluster.getNameNode(1).getConf()
.setBoolean("dfs.namenode.snapshot.trashroot.enabled", true);
restartActive();
cluster.transitionToActive(1);
dfs = cluster.getFileSystem(1);
// Make sure .Trash path does not exist yet as on NN1 trash root is not
// enabled
assertFalse(dfs.exists(trashRoot));
cluster.transitionToStandby(1);
cluster.transitionToActive(0);
dfs = cluster.getFileSystem(0);
// Check .Trash existence, should be created now
assertTrue(dfs.exists(trashRoot));
assertFalse(cluster.getNameNode(0).isInSafeMode());
restartStandby();
// Ensure Standby namenode is up and running
assertTrue(cluster.getNameNode(1).isStandbyState());
// Cleanup
dfs.delete(trashRoot, true);
dfs.disallowSnapshot(testDir);
dfs.delete(testDir, true);
}

/**
* Test transition to active when namenode in safemode.
*
Expand Down