Apache
/
nifi
mirror of https://github.com/apache/nifi.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

NIFI-4774: Implementation of SequentialAccessWriteAheadLog and updates to WriteAheadFlowFileRepository to make use of the new implementation instead of MinimalLockingWriteAheadLog. 

Signed-off-by: Matthew Burgess <mattyb149@apache.org>

This closes #2416
This commit is contained in:
Mark Payne 2018-01-17 14:24:04 -05:00 committed by Matthew Burgess
parent 14fef2de14
commit 0bcb241db3
26 changed files with 2960 additions and 48 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

56
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/BlockingQueuePool.java Normal file
View File

@ -0,0 +1,56 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import java.util.concurrent.BlockingQueue;
import java.util.concurrent.LinkedBlockingQueue;
import java.util.function.Consumer;
import java.util.function.Predicate;
import java.util.function.Supplier;
public class BlockingQueuePool<T> implements ObjectPool<T> {
private final BlockingQueue<T> queue;
private final Supplier<T> creationFunction;
private final Predicate<T> reuseCheck;
private final Consumer<T> returnPreparation;
public BlockingQueuePool(final int maxSize, final Supplier<T> creationFunction, final Predicate<T> reuseCheck, final Consumer<T> returnPreparation) {
this.queue = new LinkedBlockingQueue<>(maxSize);
this.creationFunction = creationFunction;
this.reuseCheck = reuseCheck;
this.returnPreparation = returnPreparation;
}
@Override
public T borrowObject() {
final T existing = queue.poll();
if (existing != null) {
return existing;
}
return creationFunction.get();
}
@Override
public void returnObject(final T somethingBorrowed) {
if (reuseCheck.test(somethingBorrowed)) {
returnPreparation.accept(somethingBorrowed);
queue.offer(somethingBorrowed);
}
}
}

45
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/ByteArrayDataOutputStream.java Normal file
View File

@ -0,0 +1,45 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import java.io.ByteArrayOutputStream;
import java.io.DataOutputStream;
/**
* A wrapper around a DataOutputStream, which wraps a ByteArrayOutputStream.
* This allows us to obtain the DataOutputStream itself so that we can perform
* writeXYZ methods and also allows us to obtain the underlying ByteArrayOutputStream
* for performing methods such as size(), reset(), writeTo()
*/
public class ByteArrayDataOutputStream {
private final ByteArrayOutputStream baos;
private final DataOutputStream dos;
public ByteArrayDataOutputStream(final int intiialBufferSize) {
this.baos = new ByteArrayOutputStream(intiialBufferSize);
this.dos = new DataOutputStream(baos);
}
public DataOutputStream getDataOutputStream() {
return dos;
}
public ByteArrayOutputStream getByteArrayOutputStream() {
return baos;
}
}

366
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/HashMapSnapshot.java Normal file
View File

@ -0,0 +1,366 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import java.io.BufferedInputStream;
import java.io.BufferedOutputStream;
import java.io.DataInputStream;
import java.io.DataOutputStream;
import java.io.EOFException;
import java.io.File;
import java.io.FileInputStream;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.OutputStream;
import java.nio.file.Files;
import java.util.Collection;
import java.util.Collections;
import java.util.HashMap;
import java.util.HashSet;
import java.util.Map;
import java.util.Set;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.ConcurrentMap;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.wali.SerDe;
import org.wali.SerDeFactory;
import org.wali.UpdateType;
public class HashMapSnapshot<T> implements WriteAheadSnapshot<T>, RecordLookup<T> {
private static final Logger logger = LoggerFactory.getLogger(HashMapSnapshot.class);
private static final int ENCODING_VERSION = 1;
private final ConcurrentMap<Object, T> recordMap = new ConcurrentHashMap<>();
private final SerDeFactory<T> serdeFactory;
private final Set<String> swapLocations = Collections.synchronizedSet(new HashSet<>());
private final File storageDirectory;
public HashMapSnapshot(final File storageDirectory, final SerDeFactory<T> serdeFactory) {
this.serdeFactory = serdeFactory;
this.storageDirectory = storageDirectory;
}
private SnapshotHeader validateHeader(final DataInputStream dataIn) throws IOException {
final String snapshotClass = dataIn.readUTF();
logger.debug("Snapshot Class Name for {} is {}", storageDirectory, snapshotClass);
if (!snapshotClass.equals(HashMapSnapshot.class.getName())) {
throw new IOException("Write-Ahead Log Snapshot located at " + storageDirectory + " was written using the "
+ snapshotClass + " class; cannot restore using " + getClass().getName());
}
final int snapshotVersion = dataIn.readInt();
logger.debug("Snapshot version for {} is {}", storageDirectory, snapshotVersion);
if (snapshotVersion > getVersion()) {
throw new IOException("Write-Ahead Log Snapshot located at " + storageDirectory + " was written using version "
+ snapshotVersion + " of the " + snapshotClass + " class; cannot restore using Version " + getVersion());
}
final String serdeEncoding = dataIn.readUTF(); // ignore serde class name for now
logger.debug("Serde encoding for Snapshot at {} is {}", storageDirectory, serdeEncoding);
final int serdeVersion = dataIn.readInt();
logger.debug("Serde version for Snapshot at {} is {}", storageDirectory, serdeVersion);
final long maxTransactionId = dataIn.readLong();
logger.debug("Max Transaction ID for Snapshot at {} is {}", storageDirectory, maxTransactionId);
final int numRecords = dataIn.readInt();
logger.debug("Number of Records for Snapshot at {} is {}", storageDirectory, numRecords);
final SerDe<T> serde = serdeFactory.createSerDe(serdeEncoding);
serde.readHeader(dataIn);
return new SnapshotHeader(serde, serdeVersion, maxTransactionId, numRecords);
}
@Override
public SnapshotRecovery<T> recover() throws IOException {
final File partialFile = getPartialFile();
final File snapshotFile = getSnapshotFile();
final boolean partialExists = partialFile.exists();
final boolean snapshotExists = snapshotFile.exists();
// If there is no snapshot (which is the case before the first snapshot is ever created), then just
// return an empty recovery.
if (!partialExists && !snapshotExists) {
return SnapshotRecovery.emptyRecovery();
}
if (partialExists && snapshotExists) {
// both files exist -- assume NiFi crashed/died while checkpointing. Delete the partial file.
Files.delete(partialFile.toPath());
} else if (partialExists) {
// partial exists but snapshot does not -- we must have completed
// creating the partial, deleted the snapshot
// but crashed before renaming the partial to the snapshot. Just
// rename partial to snapshot
Files.move(partialFile.toPath(), snapshotFile.toPath());
}
if (snapshotFile.length() == 0) {
logger.warn("{} Found 0-byte Snapshot file; skipping Snapshot file in recovery", this);
return SnapshotRecovery.emptyRecovery();
}
// At this point, we know the snapshotPath exists because if it didn't, then we either returned null
// or we renamed partialPath to snapshotPath. So just Recover from snapshotPath.
try (final DataInputStream dataIn = new DataInputStream(new BufferedInputStream(new FileInputStream(snapshotFile)))) {
// Ensure that the header contains the information that we expect and retrieve the relevant information from the header.
final SnapshotHeader header = validateHeader(dataIn);
final SerDe<T> serde = header.getSerDe();
final int serdeVersion = header.getSerDeVersion();
final int numRecords = header.getNumRecords();
final long maxTransactionId = header.getMaxTransactionId();
// Read all of the records that we expect to receive.
for (int i = 0; i < numRecords; i++) {
final T record = serde.deserializeRecord(dataIn, serdeVersion);
if (record == null) {
throw new EOFException();
}
final UpdateType updateType = serde.getUpdateType(record);
if (updateType == UpdateType.DELETE) {
logger.warn("While recovering from snapshot, found record with type 'DELETE'; this record will not be restored");
continue;
}
logger.trace("Recovered from snapshot: {}", record);
recordMap.put(serde.getRecordIdentifier(record), record);
}
// Determine the location of any swap files.
final int numSwapRecords = dataIn.readInt();
final Set<String> swapLocations = new HashSet<>();
for (int i = 0; i < numSwapRecords; i++) {
swapLocations.add(dataIn.readUTF());
}
this.swapLocations.addAll(swapLocations);
logger.info("{} restored {} Records and {} Swap Files from Snapshot, ending with Transaction ID {}",
new Object[] {this, numRecords, swapLocations.size(), maxTransactionId});
return new StandardSnapshotRecovery<>(recordMap, swapLocations, snapshotFile, maxTransactionId);
}
}
@Override
public void update(final Collection<T> records) {
// This implementation of Snapshot keeps a ConcurrentHashMap of all 'active' records
// (meaning records that have not been removed and are not swapped out), keyed by the
// Record Identifier. It keeps only the most up-to-date version of the Record. This allows
// us to write the snapshot very quickly without having to re-process the journal files.
// For each update, then, we will update the record in the map.
for (final T record : records) {
final Object recordId = serdeFactory.getRecordIdentifier(record);
final UpdateType updateType = serdeFactory.getUpdateType(record);
switch (updateType) {
case DELETE:
recordMap.remove(recordId);
break;
case SWAP_OUT:
final String location = serdeFactory.getLocation(record);
if (location == null) {
logger.error("Received Record (ID=" + recordId + ") with UpdateType of SWAP_OUT but "
+ "no indicator of where the Record is to be Swapped Out to; these records may be "
+ "lost when the repository is restored!");
} else {
recordMap.remove(recordId);
this.swapLocations.add(location);
}
break;
case SWAP_IN:
final String swapLocation = serdeFactory.getLocation(record);
if (swapLocation == null) {
logger.error("Received Record (ID=" + recordId + ") with UpdateType of SWAP_IN but no "
+ "indicator of where the Record is to be Swapped In from; these records may be duplicated "
+ "when the repository is restored!");
} else {
swapLocations.remove(swapLocation);
}
recordMap.put(recordId, record);
break;
default:
recordMap.put(recordId, record);
break;
}
}
}
@Override
public int getRecordCount() {
return recordMap.size();
}
@Override
public T lookup(final Object recordId) {
return recordMap.get(recordId);
}
@Override
public SnapshotCapture<T> prepareSnapshot(final long maxTransactionId) {
return new Snapshot(new HashMap<>(recordMap), new HashSet<>(swapLocations), maxTransactionId);
}
private int getVersion() {
return ENCODING_VERSION;
}
private File getPartialFile() {
return new File(storageDirectory, "checkpoint.partial");
}
private File getSnapshotFile() {
return new File(storageDirectory, "checkpoint");
}
@Override
public synchronized void writeSnapshot(final SnapshotCapture<T> snapshot) throws IOException {
final SerDe<T> serde = serdeFactory.createSerDe(null);
final File snapshotFile = getSnapshotFile();
final File partialFile = getPartialFile();
// We must ensure that we do not overwrite the existing Snapshot file directly because if NiFi were
// to be killed or crash when we are partially finished, we'd end up with no viable Snapshot file at all.
// To avoid this, we write to a 'partial' file, then delete the existing Snapshot file, if it exists, and
// rename 'partial' to Snaphsot. That way, if NiFi crashes, we can still restore the Snapshot by first looking
// for a Snapshot file and restoring it, if it exists. If it does not exist, then we restore from the partial file,
// assuming that NiFi crashed after deleting the Snapshot file and before renaming the partial file.
//
// If there is no Snapshot file currently but there is a Partial File, then this indicates
// that we have deleted the Snapshot file and failed to rename the Partial File. We don't want
// to overwrite the Partial file, because doing so could potentially lose data. Instead, we must
// first rename it to Snapshot and then write to the partial file.
if (!snapshotFile.exists() && partialFile.exists()) {
final boolean rename = partialFile.renameTo(snapshotFile);
if (!rename) {
throw new IOException("Failed to rename partial snapshot file " + partialFile + " to " + snapshotFile);
}
}
// Write to the partial file.
try (final FileOutputStream fileOut = new FileOutputStream(getPartialFile());
final OutputStream bufferedOut = new BufferedOutputStream(fileOut);
final DataOutputStream dataOut = new DataOutputStream(bufferedOut)) {
// Write out the header
dataOut.writeUTF(HashMapSnapshot.class.getName());
dataOut.writeInt(getVersion());
dataOut.writeUTF(serde.getClass().getName());
dataOut.writeInt(serde.getVersion());
dataOut.writeLong(snapshot.getMaxTransactionId());
dataOut.writeInt(snapshot.getRecords().size());
serde.writeHeader(dataOut);
// Serialize each record
for (final T record : snapshot.getRecords().values()) {
logger.trace("Checkpointing {}", record);
serde.serializeRecord(record, dataOut);
}
// Write out the number of swap locations, followed by the swap locations themselves.
dataOut.writeInt(snapshot.getSwapLocations().size());
for (final String swapLocation : snapshot.getSwapLocations()) {
dataOut.writeUTF(swapLocation);
}
// Ensure that we flush the Buffered Output Stream and then perform an fsync().
// This ensures that the data is fully written to disk before we delete the existing snapshot.
dataOut.flush();
fileOut.getChannel().force(false);
}
// If the snapshot file exists, delete it
if (snapshotFile.exists()) {
if (!snapshotFile.delete()) {
logger.warn("Unable to delete existing Snapshot file " + snapshotFile);
}
}
// Rename the partial file to Snapshot.
final boolean rename = partialFile.renameTo(snapshotFile);
if (!rename) {
throw new IOException("Failed to rename partial snapshot file " + partialFile + " to " + snapshotFile);
}
}
public class Snapshot implements SnapshotCapture<T> {
private final Map<Object, T> records;
private final long maxTransactionId;
private final Set<String> swapLocations;
public Snapshot(final Map<Object, T> records, final Set<String> swapLocations, final long maxTransactionId) {
this.records = records;
this.swapLocations = swapLocations;
this.maxTransactionId = maxTransactionId;
}
@Override
public final Map<Object, T> getRecords() {
return records;
}
@Override
public long getMaxTransactionId() {
return maxTransactionId;
}
@Override
public Set<String> getSwapLocations() {
return swapLocations;
}
}
private class SnapshotHeader {
private final SerDe<T> serde;
private final int serdeVersion;
private final int numRecords;
private final long maxTransactionId;
public SnapshotHeader(final SerDe<T> serde, final int serdeVersion, final long maxTransactionId, final int numRecords) {
this.serde = serde;
this.serdeVersion = serdeVersion;
this.maxTransactionId = maxTransactionId;
this.numRecords = numRecords;
}
public SerDe<T> getSerDe() {
return serde;
}
public int getSerDeVersion() {
return serdeVersion;
}
public long getMaxTransactionId() {
return maxTransactionId;
}
public int getNumRecords() {
return numRecords;
}
}
}

27
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/JournalRecovery.java Normal file
View File

@ -0,0 +1,27 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
public interface JournalRecovery {
int getUpdateCount();
long getMaxTransactionId();
boolean isEOFExceptionEncountered();
}

35
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/JournalSummary.java Normal file
View File

@ -0,0 +1,35 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
public interface JournalSummary {
/**
* @return the Transaction ID of the first transaction written
*/
long getFirstTransactionId();
/**
* @return the Transaction ID of the last transaction written
*/
long getLastTransactionId();
/**
* @return the number of transactions that were written to the journal
*/
int getTransactionCount();
}

478
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/LengthDelimitedJournal.java Normal file
View File

@ -0,0 +1,478 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import java.io.BufferedInputStream;
import java.io.BufferedOutputStream;
import java.io.ByteArrayOutputStream;
import java.io.DataInputStream;
import java.io.DataOutputStream;
import java.io.EOFException;
import java.io.File;
import java.io.FileInputStream;
import java.io.FileNotFoundException;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.OutputStream;
import java.nio.ByteBuffer;
import java.text.DecimalFormat;
import java.util.Collection;
import java.util.HashMap;
import java.util.HashSet;
import java.util.Map;
import java.util.Set;
import org.apache.nifi.stream.io.ByteCountingInputStream;
import org.apache.nifi.stream.io.LimitingInputStream;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.wali.SerDe;
import org.wali.SerDeFactory;
import org.wali.UpdateType;
public class LengthDelimitedJournal<T> implements WriteAheadJournal<T> {
private static final Logger logger = LoggerFactory.getLogger(LengthDelimitedJournal.class);
private static final JournalSummary INACTIVE_JOURNAL_SUMMARY = new StandardJournalSummary(-1L, -1L, 0);
private static final int JOURNAL_ENCODING_VERSION = 1;
private static final byte TRANSACTION_FOLLOWS = 64;
private static final byte JOURNAL_COMPLETE = 127;
private static final int NUL_BYTE = 0;
private final File journalFile;
private final long initialTransactionId;
private final SerDeFactory<T> serdeFactory;
private final ObjectPool<ByteArrayDataOutputStream> streamPool;
private SerDe<T> serde;
private FileOutputStream fileOut;
private BufferedOutputStream bufferedOut;
private long currentTransactionId;
private int transactionCount;
private boolean headerWritten = false;
private volatile boolean poisoned = false;
private volatile boolean closed = false;
private final ByteBuffer transactionPreamble = ByteBuffer.allocate(12); // guarded by synchronized block
public LengthDelimitedJournal(final File journalFile, final SerDeFactory<T> serdeFactory, final ObjectPool<ByteArrayDataOutputStream> streamPool, final long initialTransactionId) {
this.journalFile = journalFile;
this.serdeFactory = serdeFactory;
this.serde = serdeFactory.createSerDe(null);
this.streamPool = streamPool;
this.initialTransactionId = initialTransactionId;
this.currentTransactionId = initialTransactionId;
}
private synchronized OutputStream getOutputStream() throws FileNotFoundException {
if (fileOut == null) {
fileOut = new FileOutputStream(journalFile);
bufferedOut = new BufferedOutputStream(fileOut);
}
return bufferedOut;
}
@Override
public synchronized void writeHeader() throws IOException {
try {
final DataOutputStream outStream = new DataOutputStream(getOutputStream());
outStream.writeUTF(LengthDelimitedJournal.class.getName());
outStream.writeInt(JOURNAL_ENCODING_VERSION);
serde = serdeFactory.createSerDe(null);
outStream.writeUTF(serde.getClass().getName());
outStream.writeInt(serde.getVersion());
try (final ByteArrayOutputStream baos = new ByteArrayOutputStream();
final DataOutputStream dos = new DataOutputStream(baos)) {
serde.writeHeader(dos);
dos.flush();
final int serdeHeaderLength = baos.size();
outStream.writeInt(serdeHeaderLength);
baos.writeTo(outStream);
}
outStream.flush();
} catch (final Throwable t) {
poison(t);
final IOException ioe = (t instanceof IOException) ? (IOException) t : new IOException("Failed to create journal file " + journalFile, t);
logger.error("Failed to create new journal file {} due to {}", journalFile, ioe.toString(), ioe);
throw ioe;
}
headerWritten = true;
}
private synchronized SerDeAndVersion validateHeader(final DataInputStream in) throws IOException {
final String journalClassName = in.readUTF();
logger.debug("Write Ahead Log Class Name for {} is {}", journalFile, journalClassName);
if (!LengthDelimitedJournal.class.getName().equals(journalClassName)) {
throw new IOException("Invalid header information - " + journalFile + " does not appear to be a valid journal file.");
}
final int encodingVersion = in.readInt();
logger.debug("Encoding version for {} is {}", journalFile, encodingVersion);
if (encodingVersion > JOURNAL_ENCODING_VERSION) {
throw new IOException("Cannot read journal file " + journalFile + " because it is encoded using veresion " + encodingVersion
+ " but this version of the code only understands version " + JOURNAL_ENCODING_VERSION + " and below");
}
final String serdeClassName = in.readUTF();
logger.debug("Serde Class Name for {} is {}", journalFile, serdeClassName);
final SerDe<T> serde;
try {
serde = serdeFactory.createSerDe(serdeClassName);
} catch (final IllegalArgumentException iae) {
throw new IOException("Cannot read journal file " + journalFile + " because the serializer/deserializer used was " + serdeClassName
+ " but this repository is configured to use a different type of serializer/deserializer");
}
final int serdeVersion = in.readInt();
logger.debug("Serde version is {}", serdeVersion);
if (serdeVersion > serde.getVersion()) {
throw new IOException("Cannot read journal file " + journalFile + " because it is encoded using veresion " + encodingVersion
+ " of the serializer/deserializer but this version of the code only understands version " + serde.getVersion() + " and below");
}
final int serdeHeaderLength = in.readInt();
final InputStream serdeHeaderIn = new LimitingInputStream(in, serdeHeaderLength);
final DataInputStream dis = new DataInputStream(serdeHeaderIn);
serde.readHeader(dis);
return new SerDeAndVersion(serde, serdeVersion);
}
@Override
public void update(final Collection<T> records, final RecordLookup<T> recordLookup) throws IOException {
if (!headerWritten) {
throw new IllegalStateException("Cannot update journal file " + journalFile + " because no header has been written yet.");
}
if (records.isEmpty()) {
return;
}
checkState();
final ByteArrayDataOutputStream bados = streamPool.borrowObject();
try {
for (final T record : records) {
final Object recordId = serde.getRecordIdentifier(record);
final T previousRecordState = recordLookup.lookup(recordId);
serde.serializeEdit(previousRecordState, record, bados.getDataOutputStream());
}
final ByteArrayOutputStream baos = bados.getByteArrayOutputStream();
final OutputStream out = getOutputStream();
final long transactionId;
synchronized (this) {
transactionId = currentTransactionId++;
transactionCount++;
transactionPreamble.clear();
transactionPreamble.putLong(transactionId);
transactionPreamble.putInt(baos.size());
out.write(TRANSACTION_FOLLOWS);
out.write(transactionPreamble.array());
baos.writeTo(out);
out.flush();
}
logger.debug("Wrote Transaction {} to journal {} with length {} and {} records", transactionId, journalFile, baos.size(), records.size());
} catch (final Throwable t) {
poison(t);
throw t;
} finally {
streamPool.returnObject(bados);
}
}
private void checkState() throws IOException {
if (poisoned) {
throw new IOException("Cannot update journal file " + journalFile + " because this journal has already encountered a failure when attempting to write to the file. "
+ "If the repository is able to checkpoint, then this problem will resolve itself. However, if the repository is unable to be checkpointed "
+ "(for example, due to being out of storage space or having too many open files), then this issue may require manual intervention.");
}
if (closed) {
throw new IOException("Cannot update journal file " + journalFile + " because this journal has already been closed");
}
}
private void poison(final Throwable t) {
this.poisoned = true;
try {
if (fileOut != null) {
fileOut.close();
}
closed = true;
} catch (final IOException innerIOE) {
t.addSuppressed(innerIOE);
}
}
@Override
public synchronized void fsync() throws IOException {
checkState();
try {
if (fileOut != null) {
fileOut.getChannel().force(false);
}
} catch (final IOException ioe) {
poison(ioe);
}
}
@Override
public synchronized void close() throws IOException {
if (closed) {
return;
}
closed = true;
try {
if (fileOut != null) {
if (!poisoned) {
fileOut.write(JOURNAL_COMPLETE);
}
fileOut.close();
}
} catch (final IOException ioe) {
poison(ioe);
}
}
@Override
public JournalRecovery recoverRecords(final Map<Object, T> recordMap, final Set<String> swapLocations) throws IOException {
long maxTransactionId = -1L;
int updateCount = 0;
boolean eofException = false;
logger.info("Recovering records from journal {}", journalFile);
final double journalLength = journalFile.length();
try (final InputStream fis = new FileInputStream(journalFile);
final InputStream bufferedIn = new BufferedInputStream(fis);
final ByteCountingInputStream byteCountingIn = new ByteCountingInputStream(bufferedIn);
final DataInputStream in = new DataInputStream(byteCountingIn)) {
try {
// Validate that the header is what we expect and obtain the appropriate SerDe and Version information
final SerDeAndVersion serdeAndVersion = validateHeader(in);
final SerDe<T> serde = serdeAndVersion.getSerDe();
// Ensure that we get a valid transaction indicator
int transactionIndicator = in.read();
if (transactionIndicator != TRANSACTION_FOLLOWS && transactionIndicator != JOURNAL_COMPLETE && transactionIndicator != -1) {
throw new IOException("After reading " + byteCountingIn.getBytesConsumed() + " bytes from " + journalFile + ", encountered unexpected value of "
+ transactionIndicator + " for the Transaction Indicator. This journal may have been corrupted.");
}
long consumedAtLog = 0L;
// We don't want to apply the updates in a transaction until we've finished recovering the entire
// transaction. Otherwise, we could apply say 8 out of 10 updates and then hit an EOF. In such a case,
// we want to rollback the entire transaction. We handle this by not updating recordMap or swapLocations
// variables directly but instead keeping track of the things that occurred and then once we've read the
// entire transaction, we can apply those updates to the recordMap and swapLocations.
final Map<Object, T> transactionRecordMap = new HashMap<>();
final Set<Object> idsRemoved = new HashSet<>();
final Set<String> swapLocationsRemoved = new HashSet<>();
final Set<String> swapLocationsAdded = new HashSet<>();
int transactionUpdates = 0;
// While we have a transaction to recover, recover it
while (transactionIndicator == TRANSACTION_FOLLOWS) {
transactionRecordMap.clear();
idsRemoved.clear();
swapLocationsRemoved.clear();
swapLocationsAdded.clear();
transactionUpdates = 0;
// Format is <Transaction ID: 8 bytes> <Transaction Length: 4 bytes> <Transaction data: # of bytes indicated by Transaction Length Field>
final long transactionId = in.readLong();
maxTransactionId = Math.max(maxTransactionId, transactionId);
final int transactionLength = in.readInt();
// Use SerDe to deserialize the update. We use a LimitingInputStream to ensure that the SerDe is not able to read past its intended
// length, in case there is a bug in the SerDe. We then use a ByteCountingInputStream so that we can ensure that all of the data has
// been read and throw EOFException otherwise.
final InputStream transactionLimitingIn = new LimitingInputStream(in, transactionLength);
final ByteCountingInputStream transactionByteCountingIn = new ByteCountingInputStream(transactionLimitingIn);
final DataInputStream transactionDis = new DataInputStream(transactionByteCountingIn);
while (transactionByteCountingIn.getBytesConsumed() < transactionLength) {
final T record = serde.deserializeEdit(transactionDis, recordMap, serdeAndVersion.getVersion());
// Update our RecordMap so that we have the most up-to-date version of the Record.
final Object recordId = serde.getRecordIdentifier(record);
final UpdateType updateType = serde.getUpdateType(record);
switch (updateType) {
case DELETE: {
idsRemoved.add(recordId);
transactionRecordMap.remove(recordId);
break;
}
case SWAP_IN: {
final String location = serde.getLocation(record);
if (location == null) {
logger.error("Recovered SWAP_IN record from edit log, but it did not contain a Location; skipping record");
} else {
swapLocationsRemoved.add(location);
swapLocationsAdded.remove(location);
transactionRecordMap.put(recordId, record);
}
break;
}
case SWAP_OUT: {
final String location = serde.getLocation(record);
if (location == null) {
logger.error("Recovered SWAP_OUT record from edit log, but it did not contain a Location; skipping record");
} else {
swapLocationsRemoved.remove(location);
swapLocationsAdded.add(location);
idsRemoved.add(recordId);
transactionRecordMap.remove(recordId);
}
break;
}
default: {
transactionRecordMap.put(recordId, record);
idsRemoved.remove(recordId);
break;
}
}
transactionUpdates++;
}
// Apply the transaction
for (final Object id : idsRemoved) {
recordMap.remove(id);
}
recordMap.putAll(transactionRecordMap);
swapLocations.removeAll(swapLocationsRemoved);
swapLocations.addAll(swapLocationsAdded);
updateCount += transactionUpdates;
// Check if there is another transaction to read
transactionIndicator = in.read();
if (transactionIndicator != TRANSACTION_FOLLOWS && transactionIndicator != JOURNAL_COMPLETE && transactionIndicator != -1) {
throw new IOException("After reading " + byteCountingIn.getBytesConsumed() + " bytes from " + journalFile + ", encountered unexpected value of "
+ transactionIndicator + " for the Transaction Indicator. This journal may have been corrupted.");
}
// If we have a very large journal (for instance, if checkpoint is not called for a long time, or if there is a problem rolling over
// the journal), then we want to occasionally notify the user that we are, in fact, making progress, so that it doesn't appear that
// NiFi has become "stuck".
final long consumed = byteCountingIn.getBytesConsumed();
if (consumed - consumedAtLog > 50_000_000) {
final double percentage = consumed / journalLength * 100D;
final String pct = new DecimalFormat("#.00").format(percentage);
logger.info("{}% of the way finished recovering journal {}, having recovered {} updates", pct, journalFile, updateCount);
consumedAtLog = consumed;
}
}
} catch (final EOFException eof) {
eofException = true;
logger.warn("Encountered unexpected End-of-File when reading journal file {}; assuming that NiFi was shutdown unexpectedly and continuing recovery", journalFile);
} catch (final Exception e) {
// If the stream consists solely of NUL bytes, then we want to treat it
// the same as an EOF because we see this happen when we suddenly lose power
// while writing to a file. However, if that is not the case, then something else has gone wrong.
// In such a case, there is not much that we can do but to re-throw the Exception.
if (remainingBytesAllNul(in)) {
logger.warn("Failed to recover some of the data from Write-Ahead Log Journal because encountered trailing NUL bytes. "
+ "This will sometimes happen after a sudden power loss. The rest of this journal file will be skipped for recovery purposes."
+ "The following Exception was encountered while recovering the updates to the journal:", e);
} else {
throw e;
}
}
}
logger.info("Successfully recovered {} updates from journal {}", updateCount, journalFile);
return new StandardJournalRecovery(updateCount, maxTransactionId, eofException);
}
/**
* In the case of a sudden power loss, it is common - at least in a Linux journaling File System -
* that the partition file that is being written to will have many trailing "NUL bytes" (0's).
* If this happens, then on restart we want to treat this as an incomplete transaction, so we detect
* this case explicitly.
*
* @param in the input stream to scan
* @return <code>true</code> if the InputStream contains no data or contains only NUL bytes
* @throws IOException if unable to read from the given InputStream
*/
private boolean remainingBytesAllNul(final InputStream in) throws IOException {
int nextByte;
while ((nextByte = in.read()) != -1) {
if (nextByte != NUL_BYTE) {
return false;
}
}
return true;
}
@Override
public synchronized JournalSummary getSummary() {
if (transactionCount < 1) {
return INACTIVE_JOURNAL_SUMMARY;
}
return new StandardJournalSummary(initialTransactionId, currentTransactionId - 1, transactionCount);
}
private class SerDeAndVersion {
private final SerDe<T> serde;
private final int version;
public SerDeAndVersion(final SerDe<T> serde, final int version) {
this.serde = serde;
this.version = version;
}
public SerDe<T> getSerDe() {
return serde;
}
public int getVersion() {
return version;
}
}
}

25
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/ObjectPool.java Normal file
View File

@ -0,0 +1,25 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
public interface ObjectPool<T> {
T borrowObject();
void returnObject(T somethingBorrowed);
}

29
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/RecordLookup.java Normal file
View File

@ -0,0 +1,29 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
public interface RecordLookup<T> {
/**
* Returns the Record with the given identifier, or <code>null</code> if no such record exists
*
* @param identifier the identifier of the record to lookup
* @return the Record with the given identifier, or <code>null</code> if no such record exists
*/
T lookup(Object identifier);
}

320
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/SequentialAccessWriteAheadLog.java Normal file
View File

@ -0,0 +1,320 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import java.io.File;
import java.io.IOException;
import java.util.Arrays;
import java.util.Collection;
import java.util.Collections;
import java.util.Comparator;
import java.util.List;
import java.util.Map;
import java.util.Set;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.locks.Lock;
import java.util.concurrent.locks.ReadWriteLock;
import java.util.concurrent.locks.ReentrantReadWriteLock;
import java.util.regex.Pattern;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.wali.SerDeFactory;
import org.wali.SyncListener;
import org.wali.WriteAheadRepository;
/**
* <p>
* This implementation of WriteAheadRepository provides the ability to write all updates to the
* repository sequentially by writing to a single journal file. Serialization of data into bytes
* happens outside of any lock contention and is done so using recycled byte buffers. As such,
* we occur minimal garbage collection and the theoretical throughput of this repository is equal
* to the throughput of the underlying disk itself.
* </p>
*
* <p>
* This implementation makes the assumption that only a single thread will ever issue updates for
* a given Record at any one time. I.e., the implementation is thread-safe but cannot guarantee
* that records are recovered correctly if two threads simultaneously update the write-ahead log
* with updates for the same record.
* </p>
*/
public class SequentialAccessWriteAheadLog<T> implements WriteAheadRepository<T> {
private static final int PARTITION_INDEX = 0;
private static final Logger logger = LoggerFactory.getLogger(SequentialAccessWriteAheadLog.class);
private static final Pattern JOURNAL_FILENAME_PATTERN = Pattern.compile("\\d+\\.journal");
private static final int MAX_BUFFERS = 64;
private static final int BUFFER_SIZE = 256 * 1024;
private final File storageDirectory;
private final File journalsDirectory;
private final SerDeFactory<T> serdeFactory;
private final SyncListener syncListener;
private final ReadWriteLock journalRWLock = new ReentrantReadWriteLock();
private final Lock journalReadLock = journalRWLock.readLock();
private final Lock journalWriteLock = journalRWLock.writeLock();
private final ObjectPool<ByteArrayDataOutputStream> streamPool = new BlockingQueuePool<>(MAX_BUFFERS,
() -> new ByteArrayDataOutputStream(BUFFER_SIZE),
stream -> stream.getByteArrayOutputStream().size() < BUFFER_SIZE,
stream -> stream.getByteArrayOutputStream().reset());
private final WriteAheadSnapshot<T> snapshot;
private final RecordLookup<T> recordLookup;
private SnapshotRecovery<T> snapshotRecovery;
private volatile boolean recovered = false;
private WriteAheadJournal<T> journal;
private volatile long nextTransactionId = 0L;
public SequentialAccessWriteAheadLog(final File storageDirectory, final SerDeFactory<T> serdeFactory) throws IOException {
this(storageDirectory, serdeFactory, SyncListener.NOP_SYNC_LISTENER);
}
public SequentialAccessWriteAheadLog(final File storageDirectory, final SerDeFactory<T> serdeFactory, final SyncListener syncListener) throws IOException {
if (!storageDirectory.exists() && !storageDirectory.mkdirs()) {
throw new IOException("Directory " + storageDirectory + " does not exist and cannot be created");
}
if (!storageDirectory.isDirectory()) {
throw new IOException("File " + storageDirectory + " is a regular file and not a directory");
}
final HashMapSnapshot<T> hashMapSnapshot = new HashMapSnapshot<>(storageDirectory, serdeFactory);
this.snapshot = hashMapSnapshot;
this.recordLookup = hashMapSnapshot;
this.storageDirectory = storageDirectory;
this.journalsDirectory = new File(storageDirectory, "journals");
if (!journalsDirectory.exists() && !journalsDirectory.mkdirs()) {
throw new IOException("Directory " + journalsDirectory + " does not exist and cannot be created");
}
recovered = false;
this.serdeFactory = serdeFactory;
this.syncListener = (syncListener == null) ? SyncListener.NOP_SYNC_LISTENER : syncListener;
}
@Override
public int update(final Collection<T> records, final boolean forceSync) throws IOException {
if (!recovered) {
throw new IllegalStateException("Cannot update repository until record recovery has been performed");
}
journalReadLock.lock();
try {
journal.update(records, recordLookup);
if (forceSync) {
journal.fsync();
syncListener.onSync(PARTITION_INDEX);
}
snapshot.update(records);
} finally {
journalReadLock.unlock();
}
return PARTITION_INDEX;
}
@Override
public synchronized Collection<T> recoverRecords() throws IOException {
if (recovered) {
throw new IllegalStateException("Cannot recover records from repository because record recovery has already commenced");
}
logger.info("Recovering records from Write-Ahead Log at {}", storageDirectory);
final long recoverStart = System.nanoTime();
recovered = true;
snapshotRecovery = snapshot.recover();
final long snapshotRecoveryMillis = TimeUnit.NANOSECONDS.toMillis(System.nanoTime() - recoverStart);
final Map<Object, T> recoveredRecords = snapshotRecovery.getRecords();
final Set<String> swapLocations = snapshotRecovery.getRecoveredSwapLocations();
final File[] journalFiles = journalsDirectory.listFiles(this::isJournalFile);
if (journalFiles == null) {
throw new IOException("Cannot access the list of files in directory " + journalsDirectory + "; please ensure that appropriate file permissions are set.");
}
if (snapshotRecovery.getRecoveryFile() == null) {
logger.info("No Snapshot File to recover from at {}. Now recovering records from {} journal files", storageDirectory, journalFiles.length);
} else {
logger.info("Successfully recovered {} records and {} swap files from Snapshot at {} with Max Transaction ID of {} in {} milliseconds. Now recovering records from {} journal files",
recoveredRecords.size(), swapLocations.size(), snapshotRecovery.getRecoveryFile(), snapshotRecovery.getMaxTransactionId(),
snapshotRecoveryMillis, journalFiles.length);
}
final List<File> orderedJournalFiles = Arrays.asList(journalFiles);
Collections.sort(orderedJournalFiles, new Comparator<File>() {
@Override
public int compare(final File o1, final File o2) {
final long transactionId1 = getMinTransactionId(o1);
final long transactionId2 = getMinTransactionId(o2);
return Long.compare(transactionId1, transactionId2);
}
});
final long snapshotTransactionId = snapshotRecovery.getMaxTransactionId();
int totalUpdates = 0;
int journalFilesRecovered = 0;
int journalFilesSkipped = 0;
long maxTransactionId = snapshotTransactionId;
for (final File journalFile : orderedJournalFiles) {
final long journalMinTransactionId = getMinTransactionId(journalFile);
if (journalMinTransactionId < snapshotTransactionId) {
logger.debug("Will not recover records from journal file {} because the minimum Transaction ID for that journal is {} and the Transaction ID recovered from Snapshot was {}",
journalFile, journalMinTransactionId, snapshotTransactionId);
journalFilesSkipped++;
continue;
}
logger.debug("Min Transaction ID for journal {} is {}, so will recover records from journal", journalFile, journalMinTransactionId);
journalFilesRecovered++;
try (final WriteAheadJournal<T> journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, 0L)) {
final JournalRecovery journalRecovery = journal.recoverRecords(recoveredRecords, swapLocations);
final int updates = journalRecovery.getUpdateCount();
logger.debug("Recovered {} updates from journal {}", updates, journalFile);
totalUpdates += updates;
maxTransactionId = Math.max(maxTransactionId, journalRecovery.getMaxTransactionId());
}
}
logger.debug("Recovered {} updates from {} journal files and skipped {} journal files because their data was already encapsulated in the snapshot",
totalUpdates, journalFilesRecovered, journalFilesSkipped);
this.nextTransactionId = maxTransactionId + 1;
final long recoverNanos = System.nanoTime() - recoverStart;
final long recoveryMillis = TimeUnit.MILLISECONDS.convert(recoverNanos, TimeUnit.NANOSECONDS);
logger.info("Successfully recovered {} records in {} milliseconds. Now checkpointing to ensure that Write-Ahead Log is in a consistent state", recoveredRecords.size(), recoveryMillis);
checkpoint();
return recoveredRecords.values();
}
private long getMinTransactionId(final File journalFile) {
final String filename = journalFile.getName();
final String numeral = filename.substring(0, filename.indexOf("."));
return Long.parseLong(numeral);
}
private boolean isJournalFile(final File file) {
if (!file.isFile()) {
return false;
}
final String filename = file.getName();
return JOURNAL_FILENAME_PATTERN.matcher(filename).matches();
}
@Override
public synchronized Set<String> getRecoveredSwapLocations() throws IOException {
if (!recovered) {
throw new IllegalStateException("Cannot retrieve the Recovered Swap Locations until record recovery has been performed");
}
return snapshotRecovery.getRecoveredSwapLocations();
}
@Override
public int checkpoint() throws IOException {
final SnapshotCapture<T> snapshotCapture;
final long startNanos = System.nanoTime();
final File[] existingJournals;
journalWriteLock.lock();
try {
if (journal != null) {
final JournalSummary journalSummary = journal.getSummary();
if (journalSummary.getTransactionCount() == 0) {
logger.debug("Will not checkpoint Write-Ahead Log because no updates have occurred since last checkpoint");
return snapshot.getRecordCount();
}
journal.fsync();
journal.close();
nextTransactionId = Math.max(nextTransactionId, journalSummary.getLastTransactionId() + 1);
}
syncListener.onGlobalSync();
final File[] existingFiles = journalsDirectory.listFiles(this::isJournalFile);
existingJournals = (existingFiles == null) ? new File[0] : existingFiles;
snapshotCapture = snapshot.prepareSnapshot(nextTransactionId - 1);
// Create a new journal. We name the journal file <next transaction id>.journal but it is possible
// that we could have an empty journal file already created. If this happens, we don't want to create
// a new file on top of it because it would get deleted below when we clean up old journals. So we
// will simply increment our transaction ID and try again.
File journalFile = new File(journalsDirectory, String.valueOf(nextTransactionId) + ".journal");
while (journalFile.exists()) {
nextTransactionId++;
journalFile = new File(journalsDirectory, String.valueOf(nextTransactionId) + ".journal");
}
journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, nextTransactionId);
journal.writeHeader();
logger.debug("Created new Journal starting with Transaction ID {}", nextTransactionId);
} finally {
journalWriteLock.unlock();
}
final long stopTheWorldMillis = TimeUnit.NANOSECONDS.toMillis(System.nanoTime() - startNanos);
snapshot.writeSnapshot(snapshotCapture);
for (final File existingJournal : existingJournals) {
logger.debug("Deleting Journal {} because it is now encapsulated in the latest Snapshot", existingJournal.getName());
if (!existingJournal.delete() && existingJournal.exists()) {
logger.warn("Unable to delete expired journal file " + existingJournal + "; this file should be deleted manually.");
}
}
final long totalNanos = System.nanoTime() - startNanos;
final long millis = TimeUnit.NANOSECONDS.toMillis(totalNanos);
logger.info("Checkpointed Write-Ahead Log with {} Records and {} Swap Files in {} milliseconds (Stop-the-world time = {} milliseconds), max Transaction ID {}",
new Object[] {snapshotCapture.getRecords().size(), snapshotCapture.getSwapLocations().size(), millis, stopTheWorldMillis, snapshotCapture.getMaxTransactionId()});
return snapshotCapture.getRecords().size();
}
@Override
public void shutdown() throws IOException {
journalWriteLock.lock();
try {
if (journal != null) {
journal.close();
}
} finally {
journalWriteLock.unlock();
}
}
}

29
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/SnapshotCapture.java Normal file
View File

@ -0,0 +1,29 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import java.util.Map;
import java.util.Set;
public interface SnapshotCapture<T> {
Map<Object, T> getRecords();
long getMaxTransactionId();
Set<String> getSwapLocations();
}

58
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/SnapshotRecovery.java Normal file
View File

@ -0,0 +1,58 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import java.io.File;
import java.util.Collections;
import java.util.Map;
import java.util.Set;
public interface SnapshotRecovery<T> {
long getMaxTransactionId();
Map<Object, T> getRecords();
Set<String> getRecoveredSwapLocations();
File getRecoveryFile();
public static <T> SnapshotRecovery<T> emptyRecovery() {
return new SnapshotRecovery<T>() {
@Override
public long getMaxTransactionId() {
return -1L;
}
@Override
public Map<Object, T> getRecords() {
return Collections.emptyMap();
}
@Override
public Set<String> getRecoveredSwapLocations() {
return Collections.emptySet();
}
@Override
public File getRecoveryFile() {
return null;
}
};
}
}

45
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/StandardJournalRecovery.java Normal file
View File

@ -0,0 +1,45 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
public class StandardJournalRecovery implements JournalRecovery {
private final int updateCount;
private final long maxTransactionId;
private final boolean eofException;
public StandardJournalRecovery(final int updateCount, final long maxTransactionId, final boolean eofException) {
this.updateCount = updateCount;
this.maxTransactionId = maxTransactionId;
this.eofException = eofException;
}
@Override
public int getUpdateCount() {
return updateCount;
}
@Override
public long getMaxTransactionId() {
return maxTransactionId;
}
@Override
public boolean isEOFExceptionEncountered() {
return eofException;
}
}

46
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/StandardJournalSummary.java Normal file
View File

@ -0,0 +1,46 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
public class StandardJournalSummary implements JournalSummary {
private final long firstTransactionId;
private final long lastTransactionId;
private final int transactionCount;
public StandardJournalSummary(final long firstTransactionId, final long lastTransactionId, final int transactionCount) {
this.firstTransactionId = firstTransactionId;
this.lastTransactionId = lastTransactionId;
this.transactionCount = transactionCount;
}
@Override
public long getFirstTransactionId() {
return firstTransactionId;
}
@Override
public long getLastTransactionId() {
return lastTransactionId;
}
@Override
public int getTransactionCount() {
return transactionCount;
}
}

56
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/StandardSnapshotRecovery.java Normal file
View File

@ -0,0 +1,56 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import java.io.File;
import java.util.Map;
import java.util.Set;
public class StandardSnapshotRecovery<T> implements SnapshotRecovery<T> {
private final Map<Object, T> recordMap;
private final Set<String> recoveredSwapLocations;
private final File recoveryFile;
private final long maxTransactionId;
public StandardSnapshotRecovery(final Map<Object, T> recordMap, final Set<String> recoveredSwapLocations, final File recoveryFile, final long maxTransactionId) {
this.recordMap = recordMap;
this.recoveredSwapLocations = recoveredSwapLocations;
this.recoveryFile = recoveryFile;
this.maxTransactionId = maxTransactionId;
}
@Override
public long getMaxTransactionId() {
return maxTransactionId;
}
@Override
public Map<Object, T> getRecords() {
return recordMap;
}
@Override
public Set<String> getRecoveredSwapLocations() {
return recoveredSwapLocations;
}
@Override
public File getRecoveryFile() {
return recoveryFile;
}
}

51
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/WriteAheadJournal.java Normal file
View File

@ -0,0 +1,51 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import java.io.Closeable;
import java.io.IOException;
import java.util.Collection;
import java.util.Map;
import java.util.Set;
public interface WriteAheadJournal<T> extends Closeable {
JournalRecovery recoverRecords(Map<Object, T> recordMap, Set<String> swapLocations) throws IOException;
/**
* Updates the journal with the given set of records
*
* @param records the records to update
* @param recordLookup a lookup that can be used to access the current value of a record, given its ID
*
* @throws IOException if unable to write to the underlying storage mechanism
*/
void update(Collection<T> records, RecordLookup<T> recordLookup) throws IOException;
void writeHeader() throws IOException;
void fsync() throws IOException;
/**
* Returns information about what was written to the journal
*
* @return A JournalSummary indicating what was written to the journal
* @throws IOException if unable to write to the underlying storage mechanism.
*/
JournalSummary getSummary();
}

33
nifi-commons/nifi-write-ahead-log/src/main/java/org/apache/nifi/wali/WriteAheadSnapshot.java Normal file
View File

@ -0,0 +1,33 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import java.io.IOException;
import java.util.Collection;
public interface WriteAheadSnapshot<T> {
SnapshotCapture<T> prepareSnapshot(long maxTransactionId);
void writeSnapshot(SnapshotCapture<T> snapshot) throws IOException;
SnapshotRecovery<T> recover() throws IOException;
void update(Collection<T> records);
int getRecordCount();
}

13
nifi-commons/nifi-write-ahead-log/src/main/java/org/wali/MinimalLockingWriteAheadLog.java
View File

@ -61,6 +61,7 @@ import java.util.concurrent.locks.ReadWriteLock;
import java.util.concurrent.locks.ReentrantReadWriteLock;
import java.util.regex.Pattern;
import org.apache.nifi.wali.SequentialAccessWriteAheadLog;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
@ -73,7 +74,12 @@ import org.slf4j.LoggerFactory;
* </p>
*
* @param <T> type of record this WAL is for
*
* @deprecated This implementation is now deprecated in favor of {@link SequentialAccessWriteAheadLog}.
* This implementation, when given more than 1 partition, can have issues recovering after a sudden loss
* of power or an operating system crash.
*/
@Deprecated
public final class MinimalLockingWriteAheadLog<T> implements WriteAheadRepository<T> {
private final Path basePath;
@ -105,7 +111,7 @@ public final class MinimalLockingWriteAheadLog<T> implements WriteAheadRepositor
private volatile boolean recovered = false;
public MinimalLockingWriteAheadLog(final Path path, final int partitionCount, final SerDe<T> serde, final SyncListener syncListener) throws IOException {
this(new TreeSet<>(Collections.singleton(path)), partitionCount, new SingletonSerDeFactory<T>(serde), syncListener);
this(new TreeSet<>(Collections.singleton(path)), partitionCount, new SingletonSerDeFactory<>(serde), syncListener);
}
public MinimalLockingWriteAheadLog(final Path path, final int partitionCount, final SerDeFactory<T> serdeFactory, final SyncListener syncListener) throws IOException {
@ -113,7 +119,7 @@ public final class MinimalLockingWriteAheadLog<T> implements WriteAheadRepositor
}
public MinimalLockingWriteAheadLog(final SortedSet<Path> paths, final int partitionCount, final SerDe<T> serde, final SyncListener syncListener) throws IOException {
this(paths, partitionCount, new SingletonSerDeFactory<T>(serde), syncListener);
this(paths, partitionCount, new SingletonSerDeFactory<>(serde), syncListener);
}
/**
@ -645,6 +651,9 @@ public final class MinimalLockingWriteAheadLog<T> implements WriteAheadRepositor
} finally {
writeLock.unlock();
lockChannel.close();
final File lockFile = new File(basePath.toFile(), "wali.lock");
lockFile.delete();
}
}

10
nifi-commons/nifi-write-ahead-log/src/main/java/org/wali/SyncListener.java
View File

@ -59,4 +59,14 @@ public interface SyncListener {
* {@link WriteAheadRepository#sync()} method.
*/
void onGlobalSync();
public static final SyncListener NOP_SYNC_LISTENER = new SyncListener() {
@Override
public void onSync(int partitionIndex) {
}
@Override
public void onGlobalSync() {
}
};
}

115
nifi-commons/nifi-write-ahead-log/src/test/java/org/apache/nifi/wali/TestBlockingQueuePool.java Normal file
View File

@ -0,0 +1,115 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import static org.junit.Assert.assertFalse;
import static org.junit.Assert.assertNotNull;
import static org.junit.Assert.assertNotSame;
import static org.junit.Assert.assertSame;
import static org.junit.Assert.assertTrue;
import java.util.concurrent.atomic.AtomicBoolean;
import java.util.function.Consumer;
import org.junit.Test;
public class TestBlockingQueuePool {
private static final Consumer<AtomicBoolean> DO_NOTHING = ab -> {};
@Test
public void testReuse() {
final BlockingQueuePool<AtomicBoolean> pool = new BlockingQueuePool<>(10, AtomicBoolean::new, AtomicBoolean::get, DO_NOTHING);
final AtomicBoolean firstObject = pool.borrowObject();
firstObject.set(true);
pool.returnObject(firstObject);
for (int i = 0; i < 100; i++) {
final AtomicBoolean value = pool.borrowObject();
assertSame(firstObject, value);
pool.returnObject(value);
}
}
@Test
public void testCreateOnExhaustion() {
final BlockingQueuePool<AtomicBoolean> pool = new BlockingQueuePool<>(10, AtomicBoolean::new, AtomicBoolean::get, DO_NOTHING);
final AtomicBoolean firstObject = pool.borrowObject();
final AtomicBoolean secondObject = pool.borrowObject();
assertNotSame(firstObject, secondObject);
}
@Test
public void testCreateMoreThanMaxCapacity() {
final BlockingQueuePool<AtomicBoolean> pool = new BlockingQueuePool<>(10, AtomicBoolean::new, AtomicBoolean::get, DO_NOTHING);
for (int i = 0; i < 50; i++) {
final AtomicBoolean value = pool.borrowObject();
assertNotNull(value);
}
}
@Test
public void testDoesNotBufferMoreThanCapacity() {
final BlockingQueuePool<AtomicBoolean> pool = new BlockingQueuePool<>(10, AtomicBoolean::new, AtomicBoolean::get, DO_NOTHING);
final AtomicBoolean[] seen = new AtomicBoolean[50];
for (int i = 0; i < 50; i++) {
final AtomicBoolean value = pool.borrowObject();
assertNotNull(value);
value.set(true);
seen[i] = value;
}
for (final AtomicBoolean value : seen) {
pool.returnObject(value);
}
for (int i = 0; i < 10; i++) {
final AtomicBoolean value = pool.borrowObject();
// verify that the object exists in the 'seen' array
boolean found = false;
for (final AtomicBoolean seenBoolean : seen) {
if (value == seenBoolean) {
found = true;
break;
}
}
assertTrue(found);
}
for (int i = 0; i < 40; i++) {
final AtomicBoolean value = pool.borrowObject();
// verify that the object does not exist in the 'seen' array
boolean found = false;
for (final AtomicBoolean seenBoolean : seen) {
if (value == seenBoolean) {
found = true;
break;
}
}
assertFalse(found);
}
}
}

216
nifi-commons/nifi-write-ahead-log/src/test/java/org/apache/nifi/wali/TestHashMapSnapshot.java Normal file
View File

@ -0,0 +1,216 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertNotNull;
import static org.junit.Assert.assertTrue;
import java.io.File;
import java.io.IOException;
import java.nio.file.Files;
import java.util.Collections;
import java.util.HashMap;
import java.util.HashSet;
import java.util.Map;
import java.util.Set;
import org.junit.Assert;
import org.junit.Before;
import org.junit.Test;
import org.wali.DummyRecord;
import org.wali.DummyRecordSerde;
import org.wali.SerDeFactory;
import org.wali.SingletonSerDeFactory;
import org.wali.UpdateType;
public class TestHashMapSnapshot {
private final File storageDirectory = new File("target/test-hashmap-snapshot");
private DummyRecordSerde serde;
private SerDeFactory<DummyRecord> serdeFactory;
@Before
public void setup() throws IOException {
if (!storageDirectory.exists()) {
Files.createDirectories(storageDirectory.toPath());
}
final File[] childFiles = storageDirectory.listFiles();
for (final File childFile : childFiles) {
if (childFile.isFile()) {
Files.delete(childFile.toPath());
}
}
serde = new DummyRecordSerde();
serdeFactory = new SingletonSerDeFactory<>(serde);
}
@Test
public void testSuccessfulRoundTrip() throws IOException {
final HashMapSnapshot<DummyRecord> snapshot = new HashMapSnapshot<>(storageDirectory, serdeFactory);
final Map<String, String> props = new HashMap<>();
for (int i = 0; i < 10; i++) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.CREATE);
props.put("key", String.valueOf(i));
record.setProperties(props);
snapshot.update(Collections.singleton(record));
}
for (int i = 2; i < 10; i += 2) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.DELETE);
snapshot.update(Collections.singleton(record));
}
for (int i = 1; i < 10; i += 2) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.SWAP_OUT);
record.setSwapLocation("swapFile-" + i);
snapshot.update(Collections.singleton(record));
}
final DummyRecord swapIn7 = new DummyRecord("7", UpdateType.SWAP_IN);
swapIn7.setSwapLocation("swapFile-7");
snapshot.update(Collections.singleton(swapIn7));
final Set<String> swappedOutLocations = new HashSet<>();
swappedOutLocations.add("swapFile-1");
swappedOutLocations.add("swapFile-3");
swappedOutLocations.add("swapFile-5");
swappedOutLocations.add("swapFile-9");
final SnapshotCapture<DummyRecord> capture = snapshot.prepareSnapshot(180L);
assertEquals(180L, capture.getMaxTransactionId());
assertEquals(swappedOutLocations, capture.getSwapLocations());
final Map<Object, DummyRecord> records = capture.getRecords();
assertEquals(2, records.size());
assertTrue(records.containsKey("0"));
assertTrue(records.containsKey("7"));
snapshot.writeSnapshot(capture);
final SnapshotRecovery<DummyRecord> recovery = snapshot.recover();
assertEquals(180L, recovery.getMaxTransactionId());
assertEquals(swappedOutLocations, recovery.getRecoveredSwapLocations());
final Map<Object, DummyRecord> recoveredRecords = recovery.getRecords();
assertEquals(records, recoveredRecords);
}
@Test
public void testOOMEWhenWritingResultsInPreviousSnapshotStillRecoverable() throws IOException {
final HashMapSnapshot<DummyRecord> snapshot = new HashMapSnapshot<>(storageDirectory, serdeFactory);
final Map<String, String> props = new HashMap<>();
for (int i = 0; i < 11; i++) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.CREATE);
props.put("key", String.valueOf(i));
record.setProperties(props);
snapshot.update(Collections.singleton(record));
}
final DummyRecord swapOutRecord = new DummyRecord("10", UpdateType.SWAP_OUT);
swapOutRecord.setSwapLocation("SwapLocation-1");
snapshot.update(Collections.singleton(swapOutRecord));
snapshot.writeSnapshot(snapshot.prepareSnapshot(25L));
serde.setThrowOOMEAfterNSerializeEdits(3);
try {
snapshot.writeSnapshot(snapshot.prepareSnapshot(150L));
Assert.fail("Expected OOME");
} catch (final OutOfMemoryError oome) {
// expected
}
final SnapshotRecovery<DummyRecord> recovery = snapshot.recover();
assertEquals(25L, recovery.getMaxTransactionId());
final Map<Object, DummyRecord> recordMap = recovery.getRecords();
assertEquals(10, recordMap.size());
for (int i = 0; i < 10; i++) {
assertTrue(recordMap.containsKey(String.valueOf(i)));
}
for (final Map.Entry<Object, DummyRecord> entry : recordMap.entrySet()) {
final DummyRecord record = entry.getValue();
final Map<String, String> properties = record.getProperties();
assertNotNull(properties);
assertEquals(1, properties.size());
assertEquals(entry.getKey(), properties.get("key"));
}
final Set<String> swapLocations = recovery.getRecoveredSwapLocations();
assertEquals(1, swapLocations.size());
assertTrue(swapLocations.contains("SwapLocation-1"));
}
@Test
public void testIOExceptionWhenWritingResultsInPreviousSnapshotStillRecoverable() throws IOException {
final HashMapSnapshot<DummyRecord> snapshot = new HashMapSnapshot<>(storageDirectory, serdeFactory);
final Map<String, String> props = new HashMap<>();
for (int i = 0; i < 11; i++) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.CREATE);
props.put("key", String.valueOf(i));
record.setProperties(props);
snapshot.update(Collections.singleton(record));
}
final DummyRecord swapOutRecord = new DummyRecord("10", UpdateType.SWAP_OUT);
swapOutRecord.setSwapLocation("SwapLocation-1");
snapshot.update(Collections.singleton(swapOutRecord));
snapshot.writeSnapshot(snapshot.prepareSnapshot(25L));
serde.setThrowIOEAfterNSerializeEdits(3);
for (int i = 0; i < 5; i++) {
try {
snapshot.writeSnapshot(snapshot.prepareSnapshot(150L));
Assert.fail("Expected IOE");
} catch (final IOException ioe) {
// expected
}
}
final SnapshotRecovery<DummyRecord> recovery = snapshot.recover();
assertEquals(25L, recovery.getMaxTransactionId());
final Map<Object, DummyRecord> recordMap = recovery.getRecords();
assertEquals(10, recordMap.size());
for (int i = 0; i < 10; i++) {
assertTrue(recordMap.containsKey(String.valueOf(i)));
}
for (final Map.Entry<Object, DummyRecord> entry : recordMap.entrySet()) {
final DummyRecord record = entry.getValue();
final Map<String, String> properties = record.getProperties();
assertNotNull(properties);
assertEquals(1, properties.size());
assertEquals(entry.getKey(), properties.get("key"));
}
final Set<String> swapLocations = recovery.getRecoveredSwapLocations();
assertEquals(1, swapLocations.size());
assertTrue(swapLocations.contains("SwapLocation-1"));
}
}

353
nifi-commons/nifi-write-ahead-log/src/test/java/org/apache/nifi/wali/TestLengthDelimitedJournal.java Normal file
View File

@ -0,0 +1,353 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertFalse;
import static org.junit.Assert.assertNotNull;
import static org.junit.Assert.assertTrue;
import java.io.File;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.OutputStream;
import java.nio.file.Files;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collection;
import java.util.Collections;
import java.util.HashMap;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Set;
import org.junit.Assert;
import org.junit.Before;
import org.junit.Test;
import org.wali.DummyRecord;
import org.wali.DummyRecordSerde;
import org.wali.SerDeFactory;
import org.wali.SingletonSerDeFactory;
import org.wali.UpdateType;
public class TestLengthDelimitedJournal {
private final File journalFile = new File("target/testLengthDelimitedJournal/testJournal.journal");
private SerDeFactory<DummyRecord> serdeFactory;
private DummyRecordSerde serde;
private ObjectPool<ByteArrayDataOutputStream> streamPool;
private static final int BUFFER_SIZE = 4096;
@Before
public void setupJournal() throws IOException {
Files.deleteIfExists(journalFile.toPath());
if (!journalFile.getParentFile().exists()) {
Files.createDirectories(journalFile.getParentFile().toPath());
}
serde = new DummyRecordSerde();
serdeFactory = new SingletonSerDeFactory<>(serde);
streamPool = new BlockingQueuePool<>(1,
() -> new ByteArrayDataOutputStream(BUFFER_SIZE),
stream -> stream.getByteArrayOutputStream().size() < BUFFER_SIZE,
stream -> stream.getByteArrayOutputStream().reset());
}
@Test
public void testHandlingOfTrailingNulBytes() throws IOException {
try (final LengthDelimitedJournal<DummyRecord> journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, 0L)) {
journal.writeHeader();
final List<DummyRecord> firstTransaction = new ArrayList<>();
firstTransaction.add(new DummyRecord("1", UpdateType.CREATE));
firstTransaction.add(new DummyRecord("2", UpdateType.CREATE));
firstTransaction.add(new DummyRecord("3", UpdateType.CREATE));
final List<DummyRecord> secondTransaction = new ArrayList<>();
secondTransaction.add(new DummyRecord("1", UpdateType.UPDATE).setProperty("abc", "123"));
secondTransaction.add(new DummyRecord("2", UpdateType.UPDATE).setProperty("cba", "123"));
secondTransaction.add(new DummyRecord("3", UpdateType.UPDATE).setProperty("aaa", "123"));
final List<DummyRecord> thirdTransaction = new ArrayList<>();
thirdTransaction.add(new DummyRecord("1", UpdateType.DELETE));
thirdTransaction.add(new DummyRecord("2", UpdateType.DELETE));
journal.update(firstTransaction, id -> null);
journal.update(secondTransaction, id -> null);
journal.update(thirdTransaction, id -> null);
}
// Truncate the contents of the journal file by 8 bytes. Then replace with 28 trailing NUL bytes,
// as this is what we often see when we have a sudden power loss.
final byte[] contents = Files.readAllBytes(journalFile.toPath());
final byte[] truncated = Arrays.copyOfRange(contents, 0, contents.length - 8);
final byte[] withNuls = new byte[truncated.length + 28];
System.arraycopy(truncated, 0, withNuls, 0, truncated.length);
try (final OutputStream fos = new FileOutputStream(journalFile)) {
fos.write(withNuls);
}
try (final LengthDelimitedJournal<DummyRecord> journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, 0L)) {
final Map<Object, DummyRecord> recordMap = new HashMap<>();
final Set<String> swapLocations = new HashSet<>();
journal.recoverRecords(recordMap, swapLocations);
assertFalse(recordMap.isEmpty());
assertEquals(3, recordMap.size());
final DummyRecord record1 = recordMap.get("1");
assertNotNull(record1);
assertEquals(Collections.singletonMap("abc", "123"), record1.getProperties());
final DummyRecord record2 = recordMap.get("2");
assertNotNull(record2);
assertEquals(Collections.singletonMap("cba", "123"), record2.getProperties());
final DummyRecord record3 = recordMap.get("3");
assertNotNull(record3);
assertEquals(Collections.singletonMap("aaa", "123"), record3.getProperties());
}
}
@Test
public void testUpdateOnlyAppliedIfEntireTransactionApplied() throws IOException {
try (final LengthDelimitedJournal<DummyRecord> journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, 0L)) {
journal.writeHeader();
for (int i = 0; i < 3; i++) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.CREATE);
journal.update(Collections.singleton(record), key -> null);
}
final DummyRecord swapOut1Record = new DummyRecord("1", UpdateType.SWAP_OUT);
swapOut1Record.setSwapLocation("swap12");
journal.update(Collections.singleton(swapOut1Record), id -> null);
final DummyRecord swapOut2Record = new DummyRecord("2", UpdateType.SWAP_OUT);
swapOut2Record.setSwapLocation("swap12");
journal.update(Collections.singleton(swapOut2Record), id -> null);
final List<DummyRecord> records = new ArrayList<>();
for (int i = 0; i < 10; i++) {
final DummyRecord record = new DummyRecord("1" + i, UpdateType.CREATE);
records.add(record);
}
final DummyRecord swapIn1Record = new DummyRecord("1", UpdateType.SWAP_IN);
swapIn1Record.setSwapLocation("swap12");
records.add(swapIn1Record);
final DummyRecord swapOut1AgainRecord = new DummyRecord("1", UpdateType.SWAP_OUT);
swapOut1AgainRecord.setSwapLocation("swap12");
records.add(swapOut1AgainRecord);
final DummyRecord swapIn2Record = new DummyRecord("2", UpdateType.SWAP_IN);
swapIn2Record.setSwapLocation("swap12");
records.add(swapIn2Record);
final DummyRecord swapOut0Record = new DummyRecord("0", UpdateType.SWAP_OUT);
swapOut0Record.setSwapLocation("swap0");
records.add(swapOut0Record);
journal.update(records, id -> null);
}
// Truncate the last 8 bytes so that we will get an EOFException when reading the last transaction.
try (final FileOutputStream fos = new FileOutputStream(journalFile, true)) {
fos.getChannel().truncate(journalFile.length() - 8);
}
try (final LengthDelimitedJournal<DummyRecord> journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, 0L)) {
final Map<Object, DummyRecord> recordMap = new HashMap<>();
final Set<String> swapLocations = new HashSet<>();
final JournalRecovery recovery = journal.recoverRecords(recordMap, swapLocations);
assertEquals(5L, recovery.getMaxTransactionId());
assertEquals(5, recovery.getUpdateCount());
final Set<String> expectedSwap = Collections.singleton("swap12");
assertEquals(expectedSwap, swapLocations);
final Map<Object, DummyRecord> expectedRecordMap = new HashMap<>();
expectedRecordMap.put("0", new DummyRecord("0", UpdateType.CREATE));
assertEquals(expectedRecordMap, recordMap);
}
}
@Test
public void testPoisonedJournalNotWritableAfterIOE() throws IOException {
try (final LengthDelimitedJournal<DummyRecord> journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, 0L)) {
journal.writeHeader();
serde.setThrowIOEAfterNSerializeEdits(2);
final DummyRecord firstRecord = new DummyRecord("1", UpdateType.CREATE);
journal.update(Collections.singleton(firstRecord), key -> null);
final DummyRecord secondRecord = new DummyRecord("1", UpdateType.UPDATE);
journal.update(Collections.singleton(secondRecord), key -> firstRecord);
final DummyRecord thirdRecord = new DummyRecord("1", UpdateType.UPDATE);
final RecordLookup<DummyRecord> lookup = key -> secondRecord;
try {
journal.update(Collections.singleton(thirdRecord), lookup);
Assert.fail("Expected IOException");
} catch (final IOException ioe) {
// expected
}
serde.setThrowIOEAfterNSerializeEdits(-1);
final Collection<DummyRecord> records = Collections.singleton(thirdRecord);
for (int i = 0; i < 10; i++) {
try {
journal.update(records, lookup);
Assert.fail("Expected IOException");
} catch (final IOException expected) {
}
try {
journal.fsync();
Assert.fail("Expected IOException");
} catch (final IOException expected) {
}
}
}
}
@Test
public void testPoisonedJournalNotWritableAfterOOME() throws IOException {
try (final LengthDelimitedJournal<DummyRecord> journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, 0L)) {
journal.writeHeader();
serde.setThrowOOMEAfterNSerializeEdits(2);
final DummyRecord firstRecord = new DummyRecord("1", UpdateType.CREATE);
journal.update(Collections.singleton(firstRecord), key -> null);
final DummyRecord secondRecord = new DummyRecord("1", UpdateType.UPDATE);
journal.update(Collections.singleton(secondRecord), key -> firstRecord);
final DummyRecord thirdRecord = new DummyRecord("1", UpdateType.UPDATE);
final RecordLookup<DummyRecord> lookup = key -> secondRecord;
try {
journal.update(Collections.singleton(thirdRecord), lookup);
Assert.fail("Expected OOME");
} catch (final OutOfMemoryError oome) {
// expected
}
serde.setThrowOOMEAfterNSerializeEdits(-1);
final Collection<DummyRecord> records = Collections.singleton(thirdRecord);
for (int i = 0; i < 10; i++) {
try {
journal.update(records, lookup);
Assert.fail("Expected IOException");
} catch (final IOException expected) {
}
try {
journal.fsync();
Assert.fail("Expected IOException");
} catch (final IOException expected) {
}
}
}
}
@Test
public void testSuccessfulRoundTrip() throws IOException {
try (final LengthDelimitedJournal<DummyRecord> journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, 0L)) {
journal.writeHeader();
final DummyRecord firstRecord = new DummyRecord("1", UpdateType.CREATE);
journal.update(Collections.singleton(firstRecord), key -> null);
final DummyRecord secondRecord = new DummyRecord("1", UpdateType.UPDATE);
journal.update(Collections.singleton(secondRecord), key -> firstRecord);
final DummyRecord thirdRecord = new DummyRecord("1", UpdateType.UPDATE);
journal.update(Collections.singleton(thirdRecord), key -> secondRecord);
final Map<Object, DummyRecord> recordMap = new HashMap<>();
final Set<String> swapLocations = new HashSet<>();
final JournalRecovery recovery = journal.recoverRecords(recordMap, swapLocations);
assertFalse(recovery.isEOFExceptionEncountered());
assertEquals(2L, recovery.getMaxTransactionId());
assertEquals(3, recovery.getUpdateCount());
assertTrue(swapLocations.isEmpty());
assertEquals(1, recordMap.size());
final DummyRecord retrieved = recordMap.get("1");
assertNotNull(retrieved);
assertEquals(thirdRecord, retrieved);
}
}
@Test
public void testTruncatedJournalFile() throws IOException {
final DummyRecord firstRecord, secondRecord;
try (final LengthDelimitedJournal<DummyRecord> journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, 0L)) {
journal.writeHeader();
firstRecord = new DummyRecord("1", UpdateType.CREATE);
journal.update(Collections.singleton(firstRecord), key -> null);
secondRecord = new DummyRecord("2", UpdateType.CREATE);
journal.update(Collections.singleton(secondRecord), key -> firstRecord);
final DummyRecord thirdRecord = new DummyRecord("1", UpdateType.UPDATE);
journal.update(Collections.singleton(thirdRecord), key -> secondRecord);
}
// Truncate the file
try (final FileOutputStream fos = new FileOutputStream(journalFile, true)) {
fos.getChannel().truncate(journalFile.length() - 8);
}
// Ensure that we are able to recover the first two records without an issue but the third is lost.
try (final LengthDelimitedJournal<DummyRecord> journal = new LengthDelimitedJournal<>(journalFile, serdeFactory, streamPool, 0L)) {
final Map<Object, DummyRecord> recordMap = new HashMap<>();
final Set<String> swapLocations = new HashSet<>();
final JournalRecovery recovery = journal.recoverRecords(recordMap, swapLocations);
assertTrue(recovery.isEOFExceptionEncountered());
assertEquals(2L, recovery.getMaxTransactionId()); // transaction ID is still 2 because that's what was written to the journal
assertEquals(2, recovery.getUpdateCount()); // only 2 updates because the last update will incur an EOFException and be skipped
assertTrue(swapLocations.isEmpty());
assertEquals(2, recordMap.size());
final DummyRecord retrieved1 = recordMap.get("1");
assertNotNull(retrieved1);
assertEquals(firstRecord, retrieved1);
final DummyRecord retrieved2 = recordMap.get("2");
assertNotNull(retrieved2);
assertEquals(secondRecord, retrieved2);
}
}
}

345
nifi-commons/nifi-write-ahead-log/src/test/java/org/apache/nifi/wali/TestSequentialAccessWriteAheadLog.java Normal file
View File

@ -0,0 +1,345 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.nifi.wali;
import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertFalse;
import static org.junit.Assert.assertNotNull;
import static org.junit.Assert.assertTrue;
import java.io.File;
import java.io.IOException;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.text.NumberFormat;
import java.util.ArrayList;
import java.util.Collection;
import java.util.Collections;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Set;
import java.util.concurrent.TimeUnit;
import java.util.function.Function;
import java.util.stream.Collectors;
import org.junit.Assert;
import org.junit.Ignore;
import org.junit.Rule;
import org.junit.Test;
import org.junit.rules.TestName;
import org.wali.DummyRecord;
import org.wali.DummyRecordSerde;
import org.wali.SerDeFactory;
import org.wali.SingletonSerDeFactory;
import org.wali.UpdateType;
import org.wali.WriteAheadRepository;
public class TestSequentialAccessWriteAheadLog {
@Rule
public TestName testName = new TestName();
@Test
public void testRecoverWithNoCheckpoint() throws IOException {
final SequentialAccessWriteAheadLog<DummyRecord> repo = createWriteRepo();
final List<DummyRecord> records = new ArrayList<>();
for (int i = 0; i < 10; i++) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.CREATE);
records.add(record);
}
repo.update(records, false);
repo.shutdown();
final SequentialAccessWriteAheadLog<DummyRecord> recoveryRepo = createRecoveryRepo();
final Collection<DummyRecord> recovered = recoveryRepo.recoverRecords();
// ensure that we get the same records back, but the order may be different, so wrap both collections
// in a HashSet so that we can compare unordered collections of the same type.
assertEquals(new HashSet<>(records), new HashSet<>(recovered));
}
@Test
public void testRecoverWithNoJournalUpdates() throws IOException {
final SequentialAccessWriteAheadLog<DummyRecord> repo = createWriteRepo();
final List<DummyRecord> records = new ArrayList<>();
for (int i = 0; i < 10; i++) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.CREATE);
records.add(record);
}
repo.update(records, false);
repo.checkpoint();
repo.shutdown();
final SequentialAccessWriteAheadLog<DummyRecord> recoveryRepo = createRecoveryRepo();
final Collection<DummyRecord> recovered = recoveryRepo.recoverRecords();
// ensure that we get the same records back, but the order may be different, so wrap both collections
// in a HashSet so that we can compare unordered collections of the same type.
assertEquals(new HashSet<>(records), new HashSet<>(recovered));
}
@Test
public void testRecoverWithMultipleCheckpointsBetweenJournalUpdate() throws IOException {
final SequentialAccessWriteAheadLog<DummyRecord> repo = createWriteRepo();
final List<DummyRecord> records = new ArrayList<>();
for (int i = 0; i < 10; i++) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.CREATE);
records.add(record);
}
repo.update(records, false);
for (int i = 0; i < 8; i++) {
repo.checkpoint();
}
final DummyRecord updateRecord = new DummyRecord("4", UpdateType.UPDATE);
updateRecord.setProperties(Collections.singletonMap("updated", "true"));
repo.update(Collections.singleton(updateRecord), false);
repo.shutdown();
final SequentialAccessWriteAheadLog<DummyRecord> recoveryRepo = createRecoveryRepo();
final Collection<DummyRecord> recovered = recoveryRepo.recoverRecords();
// what we expect is the same as what we updated with, except we don't want the DummyRecord for CREATE 4
// because we will instead recover an UPDATE only for 4.
final Set<DummyRecord> expected = new HashSet<>(records);
expected.remove(new DummyRecord("4", UpdateType.CREATE));
expected.add(updateRecord);
// ensure that we get the same records back, but the order may be different, so wrap both collections
// in a HashSet so that we can compare unordered collections of the same type.
assertEquals(expected, new HashSet<>(recovered));
}
private SequentialAccessWriteAheadLog<DummyRecord> createRecoveryRepo() throws IOException {
final File targetDir = new File("target");
final File storageDir = new File(targetDir, testName.getMethodName());
final DummyRecordSerde serde = new DummyRecordSerde();
final SerDeFactory<DummyRecord> serdeFactory = new SingletonSerDeFactory<>(serde);
final SequentialAccessWriteAheadLog<DummyRecord> repo = new SequentialAccessWriteAheadLog<>(storageDir, serdeFactory);
return repo;
}
private SequentialAccessWriteAheadLog<DummyRecord> createWriteRepo() throws IOException {
final File targetDir = new File("target");
final File storageDir = new File(targetDir, testName.getMethodName());
deleteRecursively(storageDir);
assertTrue(storageDir.mkdirs());
final DummyRecordSerde serde = new DummyRecordSerde();
final SerDeFactory<DummyRecord> serdeFactory = new SingletonSerDeFactory<>(serde);
final SequentialAccessWriteAheadLog<DummyRecord> repo = new SequentialAccessWriteAheadLog<>(storageDir, serdeFactory);
final Collection<DummyRecord> recovered = repo.recoverRecords();
assertNotNull(recovered);
assertTrue(recovered.isEmpty());
return repo;
}
/**
* This test is designed to update the repository in several different wants, testing CREATE, UPDATE, SWAP IN, SWAP OUT, and DELETE
* update types, as well as testing updates with single records and with multiple records in a transaction. It also verifies that we
* are able to checkpoint, then update journals, and then recover updates to both the checkpoint and the journals.
*/
@Test
public void testUpdateThenRecover() throws IOException {
final SequentialAccessWriteAheadLog<DummyRecord> repo = createWriteRepo();
final DummyRecord firstCreate = new DummyRecord("0", UpdateType.CREATE);
repo.update(Collections.singleton(firstCreate), false);
final List<DummyRecord> creations = new ArrayList<>();
for (int i = 1; i < 11; i++) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.CREATE);
creations.add(record);
}
repo.update(creations, false);
final DummyRecord deleteRecord3 = new DummyRecord("3", UpdateType.DELETE);
repo.update(Collections.singleton(deleteRecord3), false);
final DummyRecord swapOutRecord4 = new DummyRecord("4", UpdateType.SWAP_OUT);
swapOutRecord4.setSwapLocation("swap");
final DummyRecord swapOutRecord5 = new DummyRecord("5", UpdateType.SWAP_OUT);
swapOutRecord5.setSwapLocation("swap");
final List<DummyRecord> swapOuts = new ArrayList<>();
swapOuts.add(swapOutRecord4);
swapOuts.add(swapOutRecord5);
repo.update(swapOuts, false);
final DummyRecord swapInRecord5 = new DummyRecord("5", UpdateType.SWAP_IN);
swapInRecord5.setSwapLocation("swap");
repo.update(Collections.singleton(swapInRecord5), false);
final int recordCount = repo.checkpoint();
assertEquals(9, recordCount);
final DummyRecord updateRecord6 = new DummyRecord("6", UpdateType.UPDATE);
updateRecord6.setProperties(Collections.singletonMap("greeting", "hello"));
repo.update(Collections.singleton(updateRecord6), false);
final List<DummyRecord> updateRecords = new ArrayList<>();
for (int i = 7; i < 11; i++) {
final DummyRecord updateRecord = new DummyRecord(String.valueOf(i), UpdateType.UPDATE);
updateRecord.setProperties(Collections.singletonMap("greeting", "hi"));
updateRecords.add(updateRecord);
}
final DummyRecord deleteRecord2 = new DummyRecord("2", UpdateType.DELETE);
updateRecords.add(deleteRecord2);
repo.update(updateRecords, false);
repo.shutdown();
final SequentialAccessWriteAheadLog<DummyRecord> recoveryRepo = createRecoveryRepo();
final Collection<DummyRecord> recoveredRecords = recoveryRepo.recoverRecords();
// We should now have records:
// 0-10 CREATED
// 2 & 3 deleted
// 4 & 5 swapped out
// 5 swapped back in
// 6 updated with greeting = hello
// 7-10 updated with greeting = hi
assertEquals(8, recoveredRecords.size());
final Map<String, DummyRecord> recordMap = recoveredRecords.stream()
.collect(Collectors.toMap(record -> record.getId(), Function.identity()));
assertFalse(recordMap.containsKey("2"));
assertFalse(recordMap.containsKey("3"));
assertFalse(recordMap.containsKey("4"));
assertTrue(recordMap.get("1").getProperties().isEmpty());
assertTrue(recordMap.get("5").getProperties().isEmpty());
assertEquals("hello", recordMap.get("6").getProperties().get("greeting"));
for (int i = 7; i < 11; i++) {
assertEquals("hi", recordMap.get(String.valueOf(i)).getProperties().get("greeting"));
}
recoveryRepo.shutdown();
}
@Test
@Ignore("For manual performance testing")
public void testUpdatePerformance() throws IOException, InterruptedException {
final Path path = Paths.get("target/sequential-access-repo");
deleteRecursively(path.toFile());
assertTrue(path.toFile().mkdirs());
final DummyRecordSerde serde = new DummyRecordSerde();
final SerDeFactory<DummyRecord> serdeFactory = new SingletonSerDeFactory<>(serde);
final WriteAheadRepository<DummyRecord> repo = new SequentialAccessWriteAheadLog<>(path.toFile(), serdeFactory);
final Collection<DummyRecord> initialRecs = repo.recoverRecords();
assertTrue(initialRecs.isEmpty());
final long updateCountPerThread = 1_000_000;
final int numThreads = 4;
final Thread[] threads = new Thread[numThreads];
final int batchSize = 1;
long previousBytes = 0L;
for (int j = 0; j < 2; j++) {
for (int i = 0; i < numThreads; i++) {
final Thread t = new Thread(new Runnable() {
@Override
public void run() {
final List<DummyRecord> batch = new ArrayList<>();
for (int i = 0; i < updateCountPerThread / batchSize; i++) {
batch.clear();
for (int j = 0; j < batchSize; j++) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.CREATE);
batch.add(record);
}
try {
repo.update(batch, false);
} catch (Throwable t) {
t.printStackTrace();
Assert.fail(t.toString());
}
}
}
});
threads[i] = t;
}
final long start = System.nanoTime();
for (final Thread t : threads) {
t.start();
}
for (final Thread t : threads) {
t.join();
}
long bytes = 0L;
for (final File journalFile : path.resolve("journals").toFile().listFiles()) {
bytes += journalFile.length();
}
bytes -= previousBytes;
previousBytes = bytes;
final long millis = TimeUnit.NANOSECONDS.toMillis(System.nanoTime() - start);
final long eventsPerSecond = (updateCountPerThread * numThreads * 1000) / millis;
final String eps = NumberFormat.getInstance().format(eventsPerSecond);
final long bytesPerSecond = bytes * 1000 / millis;
final String bps = NumberFormat.getInstance().format(bytesPerSecond);
if (j == 0) {
System.out.println(millis + " ms to insert " + updateCountPerThread * numThreads + " updates using " + numThreads
+ " threads, *as a warmup!* " + eps + " events per second, " + bps + " bytes per second");
} else {
System.out.println(millis + " ms to insert " + updateCountPerThread * numThreads + " updates using " + numThreads
+ " threads, " + eps + " events per second, " + bps + " bytes per second");
}
}
}
private void deleteRecursively(final File file) {
final File[] children = file.listFiles();
if (children != null) {
for (final File child : children) {
deleteRecursively(child);
}
}
file.delete();
}
}

31
nifi-commons/nifi-write-ahead-log/src/test/java/org/wali/DummyRecord.java
View File

@ -19,12 +19,14 @@ package org.wali;
import java.util.Collections;
import java.util.HashMap;
import java.util.Map;
import java.util.Objects;
public class DummyRecord {
private final String id;
private final Map<String, String> props;
private final UpdateType updateType;
private String swapLocation;
public DummyRecord(final String id, final UpdateType updateType) {
this.id = id;
@ -59,8 +61,37 @@ public class DummyRecord {
return props.get(name);
}
public String getSwapLocation() {
return swapLocation;
}
public void setSwapLocation(String swapLocation) {
this.swapLocation = swapLocation;
}
@Override
public String toString() {
return "DummyRecord [id=" + id + ", props=" + props + ", updateType=" + updateType + "]";
}
@Override
public int hashCode() {
return Objects.hash(this.id, this.props, this.updateType, this.swapLocation);
}
@Override
public boolean equals(final Object obj) {
if (obj == null) {
return false;
}
if (obj == this) {
return true;
}
if (!(obj instanceof DummyRecord)) {
return false;
}
final DummyRecord other = (DummyRecord) obj;
return Objects.equals(id, other.id) && Objects.equals(props, other.props) && Objects.equals(updateType, other.updateType) && Objects.equals(swapLocation, other.swapLocation);
}
}

37
nifi-commons/nifi-write-ahead-log/src/test/java/org/wali/DummyRecordSerde.java
View File

@ -27,6 +27,7 @@ public class DummyRecordSerde implements SerDe<DummyRecord> {
private int throwOOMEAfterNserializeEdits = -1;
private int serializeEditCount = 0;
@SuppressWarnings("fallthrough")
@Override
public void serializeEdit(final DummyRecord previousState, final DummyRecord record, final DataOutputStream out) throws IOException {
if (throwIOEAfterNserializeEdits >= 0 && (serializeEditCount++ >= throwIOEAfterNserializeEdits)) {
@ -39,7 +40,15 @@ public class DummyRecordSerde implements SerDe<DummyRecord> {
out.writeUTF(record.getUpdateType().name());
out.writeUTF(record.getId());
if (record.getUpdateType() != UpdateType.DELETE) {
switch (record.getUpdateType()) {
case DELETE:
break;
case SWAP_IN: {
out.writeUTF(record.getSwapLocation());
// intentionally fall through to CREATE/UPDATE block
}
case CREATE:
case UPDATE: {
final Map<String, String> props = record.getProperties();
out.writeInt(props.size());
for (final Map.Entry<String, String> entry : props.entrySet()) {
@ -47,6 +56,12 @@ public class DummyRecordSerde implements SerDe<DummyRecord> {
out.writeUTF(entry.getValue());
}
}
break;
case SWAP_OUT:
out.writeUTF(record.getSwapLocation());
break;
}
}
@Override
@ -55,20 +70,36 @@ public class DummyRecordSerde implements SerDe<DummyRecord> {
}
@Override
@SuppressWarnings("fallthrough")
public DummyRecord deserializeRecord(final DataInputStream in, final int version) throws IOException {
final String updateTypeName = in.readUTF();
final UpdateType updateType = UpdateType.valueOf(updateTypeName);
final String id = in.readUTF();
final DummyRecord record = new DummyRecord(id, updateType);
if (record.getUpdateType() != UpdateType.DELETE) {
switch (record.getUpdateType()) {
case DELETE:
break;
case SWAP_IN: {
final String swapLocation = in.readUTF();
record.setSwapLocation(swapLocation);
// intentionally fall through to the CREATE/UPDATE block
}
case CREATE:
case UPDATE:
final int numProps = in.readInt();
for (int i = 0; i < numProps; i++) {
final String key = in.readUTF();
final String value = in.readUTF();
record.setProperty(key, value);
}
break;
case SWAP_OUT:
final String swapLocation = in.readUTF();
record.setSwapLocation(swapLocation);
break;
}
return record;
}
@ -102,6 +133,6 @@ public class DummyRecordSerde implements SerDe<DummyRecord> {
@Override
public String getLocation(final DummyRecord record) {
return null;
return record.getSwapLocation();
}
}

53
nifi-commons/nifi-write-ahead-log/src/test/java/org/wali/TestMinimalLockingWriteAheadLog.java
View File

@ -34,6 +34,7 @@ import java.io.OutputStream;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.text.NumberFormat;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collection;
@ -139,9 +140,9 @@ public class TestMinimalLockingWriteAheadLog {
}
@Test
@Ignore("for local testing only")
@Ignore("For manual performance testing")
public void testUpdatePerformance() throws IOException, InterruptedException {
final int numPartitions = 4;
final int numPartitions = 16;
final Path path = Paths.get("target/minimal-locking-repo");
deleteRecursively(path.toFile());
@ -152,23 +153,34 @@ public class TestMinimalLockingWriteAheadLog {
final Collection<DummyRecord> initialRecs = repo.recoverRecords();
assertTrue(initialRecs.isEmpty());
final int updateCountPerThread = 1_000_000;
final int numThreads = 16;
final long updateCountPerThread = 1_000_000;
final int numThreads = 4;
final Thread[] threads = new Thread[numThreads];
final int batchSize = 1;
long previousBytes = 0;
for (int j = 0; j < 2; j++) {
for (int i = 0; i < numThreads; i++) {
final Thread t = new Thread(new Runnable() {
@Override
public void run() {
for (int i = 0; i < updateCountPerThread; i++) {
final List<DummyRecord> batch = new ArrayList<>();
for (int i = 0; i < updateCountPerThread / batchSize; i++) {
batch.clear();
for (int j = 0; j < batchSize; j++) {
final DummyRecord record = new DummyRecord(String.valueOf(i), UpdateType.CREATE);
batch.add(record);
}
try {
repo.update(Collections.singleton(record), false);
} catch (IOException e) {
e.printStackTrace();
Assert.fail(e.toString());
repo.update(batch, false);
} catch (Throwable t) {
t.printStackTrace();
Assert.fail(t.toString());
}
}
}
@ -185,11 +197,30 @@ public class TestMinimalLockingWriteAheadLog {
t.join();
}
long bytes = 0L;
for (final File file : path.toFile().listFiles()) {
if (file.getName().startsWith("partition-")) {
for (final File journalFile : file.listFiles()) {
bytes += journalFile.length();
}
}
}
bytes -= previousBytes;
previousBytes = bytes;
final long millis = TimeUnit.NANOSECONDS.toMillis(System.nanoTime() - start);
final long eventsPerSecond = (updateCountPerThread * numThreads * 1000) / millis;
final String eps = NumberFormat.getInstance().format(eventsPerSecond);
final long bytesPerSecond = bytes * 1000 / millis;
final String bps = NumberFormat.getInstance().format(bytesPerSecond);
if (j == 0) {
System.out.println(millis + " ms to insert " + updateCountPerThread * numThreads + " updates using " + numPartitions + " partitions and " + numThreads + " threads, *as a warmup!*");
System.out.println(millis + " ms to insert " + updateCountPerThread * numThreads + " updates using " + numThreads + " threads, *as a warmup!* "
+ eps + " events per second, " + bps + " bytes per second");
} else {
System.out.println(millis + " ms to insert " + updateCountPerThread * numThreads + " updates using " + numPartitions + " partitions and " + numThreads + " threads");
System.out.println(millis + " ms to insert " + updateCountPerThread * numThreads + " updates using " + numThreads + " threads, "
+ eps + " events per second, " + bps + " bytes per second");
}
}
}

114
nifi-nar-bundles/nifi-framework-bundle/nifi-framework/nifi-framework-core/src/main/java/org/apache/nifi/controller/repository/WriteAheadFlowFileRepository.java
View File

@ -16,10 +16,10 @@
*/
package org.apache.nifi.controller.repository;
import java.io.File;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.util.ArrayList;
import java.util.Collection;
import java.util.Collections;
@ -27,6 +27,7 @@ import java.util.HashMap;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Optional;
import java.util.Set;
import java.util.SortedSet;
import java.util.TreeSet;
@ -47,6 +48,7 @@ import org.apache.nifi.controller.repository.claim.ResourceClaim;
import org.apache.nifi.controller.repository.claim.ResourceClaimManager;
import org.apache.nifi.util.FormatUtils;
import org.apache.nifi.util.NiFiProperties;
import org.apache.nifi.wali.SequentialAccessWriteAheadLog;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.wali.MinimalLockingWriteAheadLog;
@ -86,7 +88,8 @@ public class WriteAheadFlowFileRepository implements FlowFileRepository, SyncLis
private volatile ScheduledFuture<?> checkpointFuture;
private final long checkpointDelayMillis;
private final SortedSet<Path> flowFileRepositoryPaths = new TreeSet<>();
private final File flowFileRepositoryPath;
private final List<File> recoveryFiles = new ArrayList<>();
private final int numPartitions;
private final ScheduledExecutorService checkpointExecutor;
@ -126,16 +129,23 @@ public class WriteAheadFlowFileRepository implements FlowFileRepository, SyncLis
checkpointDelayMillis = 0l;
numPartitions = 0;
checkpointExecutor = null;
flowFileRepositoryPath = null;
}
public WriteAheadFlowFileRepository(final NiFiProperties nifiProperties) {
alwaysSync = Boolean.parseBoolean(nifiProperties.getProperty(NiFiProperties.FLOWFILE_REPOSITORY_ALWAYS_SYNC, "false"));
// determine the database file path and ensure it exists
final String directoryName = nifiProperties.getProperty(FLOWFILE_REPOSITORY_DIRECTORY_PREFIX);
flowFileRepositoryPath = new File(directoryName);
// We used to use the MinimalLockingWriteAheadLog, but we now use the SequentialAccessWriteAheadLog. Since the
// MinimalLockingWriteAheadLog supports multiple partitions, we need to ensure that we recover records from all
// partitions, so we build up a List of Files for the recovery files.
for (final String propertyName : nifiProperties.getPropertyKeys()) {
if (propertyName.startsWith(FLOWFILE_REPOSITORY_DIRECTORY_PREFIX)) {
final String directoryName = nifiProperties.getProperty(propertyName);
flowFileRepositoryPaths.add(Paths.get(directoryName));
final String dirName = nifiProperties.getProperty(propertyName);
recoveryFiles.add(new File(dirName));
}
}
@ -149,16 +159,14 @@ public class WriteAheadFlowFileRepository implements FlowFileRepository, SyncLis
public void initialize(final ResourceClaimManager claimManager) throws IOException {
this.claimManager = claimManager;
for (final Path path : flowFileRepositoryPaths) {
Files.createDirectories(path);
}
Files.createDirectories(flowFileRepositoryPath.toPath());
// TODO: Should ensure that only 1 instance running and pointing at a particular path
// TODO: Allow for backup path that can be used if disk out of space?? Would allow a snapshot to be stored on
// backup and then the data deleted from the normal location; then can move backup to normal location and
// delete backup. On restore, if no files exist in partition's directory, would have to check backup directory
serdeFactory = new RepositoryRecordSerdeFactory(claimManager);
wal = new MinimalLockingWriteAheadLog<>(flowFileRepositoryPaths, numPartitions, serdeFactory, this);
wal = new SequentialAccessWriteAheadLog<>(flowFileRepositoryPath, serdeFactory, this);
logger.info("Initialized FlowFile Repository using {} partitions", numPartitions);
}
@ -179,22 +187,12 @@ public class WriteAheadFlowFileRepository implements FlowFileRepository, SyncLis
@Override
public long getStorageCapacity() throws IOException {
long capacity = 0L;
for (final Path path : flowFileRepositoryPaths) {
capacity += Files.getFileStore(path).getTotalSpace();
}
return capacity;
return Files.getFileStore(flowFileRepositoryPath.toPath()).getTotalSpace();
}
@Override
public long getUsableStorageSpace() throws IOException {
long usableSpace = 0L;
for (final Path path : flowFileRepositoryPaths) {
usableSpace += Files.getFileStore(path).getUsableSpace();
}
return usableSpace;
return Files.getFileStore(flowFileRepositoryPath.toPath()).getUsableSpace();
}
@Override
@ -371,6 +369,72 @@ public class WriteAheadFlowFileRepository implements FlowFileRepository, SyncLis
logger.info("Repository updated to reflect that {} FlowFiles were swapped in to {}", new Object[]{swapRecords.size(), queue});
}
@SuppressWarnings("deprecation")
private Optional<Collection<RepositoryRecord>> recoverFromOldWriteAheadLog() throws IOException {
final List<File> partitionDirs = new ArrayList<>();
for (final File recoveryFile : recoveryFiles) {
final File[] partitions = recoveryFile.listFiles(file -> file.getName().startsWith("partition-"));
for (final File partition : partitions) {
partitionDirs.add(partition);
}
}
if (partitionDirs == null || partitionDirs.isEmpty()) {
return Optional.empty();
}
logger.info("Encountered FlowFile Repository that was written using an old version of the Write-Ahead Log. "
+ "Will recover from this version and re-write the repository using the new version of the Write-Ahead Log.");
final SortedSet<Path> paths = recoveryFiles.stream()
.map(File::toPath)
.collect(Collectors.toCollection(TreeSet::new));
final Collection<RepositoryRecord> recordList;
final MinimalLockingWriteAheadLog<RepositoryRecord> minimalLockingWal = new MinimalLockingWriteAheadLog<>(paths, partitionDirs.size(), serdeFactory, null);
try {
recordList = minimalLockingWal.recoverRecords();
} finally {
minimalLockingWal.shutdown();
}
wal.update(recordList, true);
// Delete the old repository
logger.info("Successfully recovered files from existing Write-Ahead Log and transitioned to new implementation. Will now delete old files.");
for (final File partitionDir : partitionDirs) {
final File[] children = partitionDir.listFiles();
if (children != null) {
for (final File child : children) {
final boolean deleted = child.delete();
if (!deleted) {
logger.warn("Failed to delete old file {}; this file should be cleaned up manually", child);
}
}
}
if (!partitionDir.delete()) {
logger.warn("Failed to delete old directory {}; this directory should be cleaned up manually", partitionDir);
}
}
for (final File recoveryFile : recoveryFiles) {
final File snapshotFile = new File(recoveryFile, "snapshot");
if (!snapshotFile.delete() && snapshotFile.exists()) {
logger.warn("Failed to delete old file {}; this file should be cleaned up manually", snapshotFile);
}
final File partialFile = new File(recoveryFile, "snapshot.partial");
if (!partialFile.delete() && partialFile.exists()) {
logger.warn("Failed to delete old file {}; this file should be cleaned up manually", partialFile);
}
}
return Optional.of(recordList);
}
@Override
public long loadFlowFiles(final QueueProvider queueProvider, final long minimumSequenceNumber) throws IOException {
final Map<String, FlowFileQueue> queueMap = new HashMap<>();
@ -378,7 +442,15 @@ public class WriteAheadFlowFileRepository implements FlowFileRepository, SyncLis
queueMap.put(queue.getIdentifier(), queue);
}
serdeFactory.setQueueMap(queueMap);
final Collection<RepositoryRecord> recordList = wal.recoverRecords();
// Since we used to use the MinimalLockingWriteAheadRepository, we need to ensure that if the FlowFile
// Repo was written using that impl, that we properly recover from the implementation.
Collection<RepositoryRecord> recordList = wal.recoverRecords();
if (recordList == null || recordList.isEmpty()) {
recordList = recoverFromOldWriteAheadLog().orElse(new ArrayList<>());
}
serdeFactory.setQueueMap(null);
for (final RepositoryRecord record : recordList) {