Apache
/
hadoop
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

HDFS-3693. JNStorage should read its storage info even before a writer becomes active. Contributed by Todd Lipcon. 

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/HDFS-3077@1365794 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Todd Lipcon 2012-07-25 21:47:19 +00:00
parent d2d0736de4
commit b17018e4b8
7 changed files with 48 additions and 39 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
hadoop-hdfs-project/hadoop-hdfs/CHANGES.HDFS-3077.txt
View File

@ -6,3 +6,5 @@ HDFS-3077. Quorum-based protocol for reading and writing edit logs. Contributed
HDFS-3694. Fix getEditLogManifest to fetch httpPort if necessary (todd)
HDFS-3692. Support purgeEditLogs() call to remotely purge logs on JNs (todd)
HDFS-3693. JNStorage should read its storage info even before a writer becomes active (todd)

13
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/qjournal/server/GetJournalEditServlet.java
View File

@ -120,12 +120,13 @@ public class GetJournalEditServlet extends HttpServlet {
if (theirStorageInfoString != null
&& !myStorageInfoString.equals(theirStorageInfoString)) {
response.sendError(HttpServletResponse.SC_FORBIDDEN,
"This node has storage info " + myStorageInfoString
+ " but the requesting node expected "
+ theirStorageInfoString);
LOG.warn("Received an invalid request file transfer request "
+ " with storage info " + theirStorageInfoString);
String msg = "This node has storage info '" + myStorageInfoString
+ "' but the requesting node expected '"
+ theirStorageInfoString + "'";
response.sendError(HttpServletResponse.SC_FORBIDDEN, msg);
LOG.warn("Received an invalid request file transfer request from " +
request.getRemoteAddr() + ": " + msg);
return false;
}
return true;

52
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/qjournal/server/JNStorage.java
View File

@ -28,6 +28,8 @@ import org.apache.hadoop.hdfs.server.namenode.FileJournalManager;
import org.apache.hadoop.hdfs.server.namenode.NNStorage;
import org.apache.hadoop.hdfs.server.protocol.NamespaceInfo;
import com.google.common.base.Preconditions;
/**
* A {@link Storage} implementation for the {@link JournalNode}.
*
@ -38,18 +40,21 @@ class JNStorage extends Storage {
private final FileJournalManager fjm;
private final StorageDirectory sd;
private boolean lazyInitted = false;
private StorageState state;
/**
* @param logDir the path to the directory in which data will be stored
* @param errorReporter a callback to report errors
* @throws IOException
*/
protected JNStorage(File logDir, StorageErrorReporter errorReporter) {
protected JNStorage(File logDir, StorageErrorReporter errorReporter) throws IOException {
super(NodeType.JOURNAL_NODE);
sd = new StorageDirectory(logDir);
this.addStorageDir(sd);
this.fjm = new FileJournalManager(sd, errorReporter);
analyzeStorage();
}
FileJournalManager getJournalManager() {
@ -107,34 +112,27 @@ class JNStorage extends Storage {
}
}
void analyzeStorage(NamespaceInfo nsInfo) throws IOException {
if (lazyInitted) {
checkConsistentNamespace(nsInfo);
return;
}
StorageState state = sd.analyzeStorage(StartupOption.REGULAR, this);
switch (state) {
case NON_EXISTENT:
case NOT_FORMATTED:
public void formatIfNecessary(NamespaceInfo nsInfo) throws IOException {
if (state == StorageState.NOT_FORMATTED ||
state == StorageState.NON_EXISTENT) {
format(nsInfo);
// In the NORMAL case below, analyzeStorage() has already locked the
// directory for us. But in the case that we format it, we have to
// lock it here.
// The directory is unlocked in close() when the node shuts down.
sd.lock();
break;
case NORMAL:
// Storage directory is already locked by analyzeStorage() - no
// need to lock it here.
readProperties(sd);
checkConsistentNamespace(nsInfo);
break;
analyzeStorage();
assert state == StorageState.NORMAL :
"Unexpected state after formatting: " + state;
} else {
Preconditions.checkState(state == StorageState.NORMAL,
"Unhandled storage state in %s: %s", this, state);
assert getNamespaceID() != 0;
default:
LOG.warn("TODO: unhandled state for storage dir " + sd + ": " + state);
checkConsistentNamespace(nsInfo);
}
}
private void analyzeStorage() throws IOException {
this.state = sd.analyzeStorage(StartupOption.REGULAR, this);
if (state == StorageState.NORMAL) {
readProperties(sd);
}
lazyInitted = true;
}
private void checkConsistentNamespace(NamespaceInfo nsInfo)

4
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/qjournal/server/Journal.java
View File

@ -79,7 +79,7 @@ class Journal implements Closeable {
private final FileJournalManager fjm;
Journal(File logDir, StorageErrorReporter errorReporter) {
Journal(File logDir, StorageErrorReporter errorReporter) throws IOException {
storage = new JNStorage(logDir, errorReporter);
File currentDir = storage.getSingularStorageDir().getCurrentDir();
@ -152,7 +152,7 @@ class Journal implements Closeable {
// If the storage is unformatted, format it with this NS.
// Otherwise, check that the NN's nsinfo matches the storage.
storage.analyzeStorage(nsInfo);
storage.formatIfNecessary(nsInfo);
if (epoch <= getLastPromisedEpoch()) {
throw new IOException("Proposed epoch " + epoch + " <= last promise " +

2
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/qjournal/server/JournalNode.java
View File

@ -64,7 +64,7 @@ public class JournalNode implements Tool, Configurable {
*/
private int resultCode = 0;
synchronized Journal getOrCreateJournal(String jid) {
synchronized Journal getOrCreateJournal(String jid) throws IOException {
QuorumJournalManager.checkJournalId(jid);
Journal journal = journalsById.get(jid);

3
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/qjournal/server/JournalNodeHttpServer.java
View File

@ -115,7 +115,8 @@ public class JournalNodeHttpServer {
DFSConfigKeys.DFS_JOURNALNODE_HTTP_ADDRESS_KEY);
}
public static Journal getJournalFromContext(ServletContext context, String jid) {
public static Journal getJournalFromContext(ServletContext context, String jid)
throws IOException {
JournalNode jn = (JournalNode)context.getAttribute(JN_ATTRIBUTE_KEY);
return jn.getOrCreateJournal(jid);
}

11
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/qjournal/server/TestJournal.java
View File

@ -112,10 +112,17 @@ public class TestJournal {
QJMTestUtil.createTxnData(1, 2));
// Don't finalize.
String storageString = journal.getStorage().toColonSeparatedString();
System.err.println("storage string: " + storageString);
journal.close(); // close to unlock the storage dir
// Now re-instantiate, make sure history is still there
journal = new Journal(TEST_LOG_DIR, mockErrorReporter);
// The storage info should be read, even if no writer has taken over.
assertEquals(storageString,
journal.getStorage().toColonSeparatedString());
assertEquals(1, journal.getLastPromisedEpoch());
NewEpochResponseProtoOrBuilder newEpoch = journal.newEpoch(FAKE_NSINFO, 2);
assertEquals(1, newEpoch.getLastSegmentTxId());
@ -135,9 +142,8 @@ public class TestJournal {
// Journal should be locked
GenericTestUtils.assertExists(lockFile);
Journal journal2 = new Journal(TEST_LOG_DIR, mockErrorReporter);
try {
journal2.newEpoch(FAKE_NSINFO, 2);
new Journal(TEST_LOG_DIR, mockErrorReporter);
fail("Did not fail to create another journal in same dir");
} catch (IOException ioe) {
GenericTestUtils.assertExceptionContains(
@ -147,6 +153,7 @@ public class TestJournal {
journal.close();
// Journal should no longer be locked after the close() call.
Journal journal2 = new Journal(TEST_LOG_DIR, mockErrorReporter);
journal2.newEpoch(FAKE_NSINFO, 2);
}