dcapwell commented on code in PR #2256: URL: https://github.com/apache/cassandra/pull/2256#discussion_r1164475071
########## src/java/org/apache/cassandra/journal/Descriptor.java: ########## @@ -0,0 +1,187 @@ +/* + * 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.cassandra.journal; + +import java.io.IOException; +import java.util.Arrays; +import java.util.List; +import java.util.regex.Matcher; +import java.util.regex.Pattern; + +import org.apache.cassandra.io.util.File; + +import static java.lang.String.format; +import static java.util.stream.Collectors.toList; + +/** + * Timestamp and version encoded in the file name, e.g. + * log-1637159888484-2-1-1.data + * log-1637159888484-2-1-1.indx + * log-1637159888484-2-1-1.meta + * log-1637159888484-2-1-1.sync + */ +final class Descriptor implements Comparable<Descriptor> +{ + private static final String SEPARATOR = "-"; + private static final String PREFIX = "log" + SEPARATOR; + private static final String TMP_SUFFIX = "tmp"; + + private static final Pattern DATA_FILE_PATTERN = + Pattern.compile( PREFIX + "(\\d+)" // timestamp + + SEPARATOR + "(\\d+)" // generation + + SEPARATOR + "(\\d+)" // journal version + + SEPARATOR + "(\\d+)" // user version + + "\\." + Component.DATA.extension); + + private static final Pattern TMP_FILE_PATTERN = + Pattern.compile( PREFIX + "\\d+" // timestamp + + SEPARATOR + "\\d+" // generation + + SEPARATOR + "\\d+" // journal version + + SEPARATOR + "\\d+" // user version + + "\\." + "[a-z]+" // component extension + + "\\." + TMP_SUFFIX); + + + static final int JOURNAL_VERSION_1 = 1; + static final int CURRENT_JOURNAL_VERSION = JOURNAL_VERSION_1; + + final File directory; + final long timestamp; + final int generation; + + /** + * Serialization version for journal components; bumped as journal + * implementation evolves over time. + */ + final int journalVersion; + + /** + * Serialization version for user content - specifically journal keys + * and journal values; bumped when user logic evolves. + */ + final int userVersion; + + private Descriptor(File directory, long timestamp, int generation, int journalVersion, int userVersion) + { + this.directory = directory; + this.timestamp = timestamp; + this.generation = generation; + this.journalVersion = journalVersion; + this.userVersion = userVersion; + } + + static Descriptor create(File directory, long timestamp, int userVersion) + { + return new Descriptor(directory, timestamp, 1, CURRENT_JOURNAL_VERSION, userVersion); + } + + static Descriptor fromName(File directory, String name) + { + Matcher matcher = DATA_FILE_PATTERN.matcher(name); + if (!matcher.matches()) + throw new IllegalArgumentException("Provided filename is not valid for a data segment file"); Review Comment: ```suggestion throw new IllegalArgumentException("Provided filename (" + new File(directory, name) + ") is not valid for a data segment file"); ``` Can find in my feedback PR ########## src/java/org/apache/cassandra/journal/Descriptor.java: ########## @@ -0,0 +1,187 @@ +/* + * 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.cassandra.journal; + +import java.io.IOException; +import java.util.Arrays; +import java.util.List; +import java.util.regex.Matcher; +import java.util.regex.Pattern; + +import org.apache.cassandra.io.util.File; + +import static java.lang.String.format; +import static java.util.stream.Collectors.toList; + +/** + * Timestamp and version encoded in the file name, e.g. + * log-1637159888484-2-1-1.data + * log-1637159888484-2-1-1.indx + * log-1637159888484-2-1-1.meta + * log-1637159888484-2-1-1.sync + */ +final class Descriptor implements Comparable<Descriptor> +{ + private static final String SEPARATOR = "-"; + private static final String PREFIX = "log" + SEPARATOR; + private static final String TMP_SUFFIX = "tmp"; + + private static final Pattern DATA_FILE_PATTERN = + Pattern.compile( PREFIX + "(\\d+)" // timestamp + + SEPARATOR + "(\\d+)" // generation + + SEPARATOR + "(\\d+)" // journal version + + SEPARATOR + "(\\d+)" // user version + + "\\." + Component.DATA.extension); + + private static final Pattern TMP_FILE_PATTERN = + Pattern.compile( PREFIX + "\\d+" // timestamp + + SEPARATOR + "\\d+" // generation + + SEPARATOR + "\\d+" // journal version + + SEPARATOR + "\\d+" // user version + + "\\." + "[a-z]+" // component extension + + "\\." + TMP_SUFFIX); + + + static final int JOURNAL_VERSION_1 = 1; + static final int CURRENT_JOURNAL_VERSION = JOURNAL_VERSION_1; + + final File directory; + final long timestamp; + final int generation; + + /** + * Serialization version for journal components; bumped as journal + * implementation evolves over time. + */ + final int journalVersion; + + /** + * Serialization version for user content - specifically journal keys + * and journal values; bumped when user logic evolves. + */ + final int userVersion; + + private Descriptor(File directory, long timestamp, int generation, int journalVersion, int userVersion) + { + this.directory = directory; + this.timestamp = timestamp; + this.generation = generation; + this.journalVersion = journalVersion; + this.userVersion = userVersion; + } + + static Descriptor create(File directory, long timestamp, int userVersion) + { + return new Descriptor(directory, timestamp, 1, CURRENT_JOURNAL_VERSION, userVersion); + } + + static Descriptor fromName(File directory, String name) + { + Matcher matcher = DATA_FILE_PATTERN.matcher(name); + if (!matcher.matches()) + throw new IllegalArgumentException("Provided filename is not valid for a data segment file"); + + long timestamp = Long.parseLong(matcher.group(1)); + int generation = Integer.parseInt(matcher.group(2)); + int journalVersion = Integer.parseInt(matcher.group(3)); + int userVersion = Integer.parseInt(matcher.group(4)); + + return new Descriptor(directory, timestamp, generation, journalVersion, userVersion); + } + + Descriptor withIncrementedGeneration() + { + return new Descriptor(directory, timestamp, generation + 1, journalVersion, userVersion); + } + + File fileFor(Component component) + { + return new File(directory, formatFileName(component)); + } + + File tmpFileFor(Component component) + { + return new File(directory, formatFileName(component) + '.' + TMP_SUFFIX); + } + + static boolean isTmpFile(File file) + { + return TMP_FILE_PATTERN.matcher(file.name()).matches(); + } + + private String formatFileName(Component component) + { + return format("%s%d%s%d%s%d%s%d.%s", + PREFIX, timestamp, + SEPARATOR, generation, + SEPARATOR, journalVersion, + SEPARATOR, userVersion, + component.extension); + } + + static List<Descriptor> list(File directory) + { + try + { + return Arrays.stream(directory.listNames((file, name) -> DATA_FILE_PATTERN.matcher(name).matches())) + .map(name -> fromName(directory, name)) + .collect(toList()); + } + catch (IOException e) + { + throw new RuntimeException(e); + } + } + + @Override + public int compareTo(Descriptor other) + { + assert this.directory.equals(other.directory); Review Comment: ```suggestion assert this.directory.equals(other.directory) : String.format("Expected directory %s, but found %s", this.directory, other.directory); ``` found in my feedback pr ########## src/java/org/apache/cassandra/journal/Descriptor.java: ########## @@ -0,0 +1,187 @@ +/* + * 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.cassandra.journal; + +import java.io.IOException; +import java.util.Arrays; +import java.util.List; +import java.util.regex.Matcher; +import java.util.regex.Pattern; + +import org.apache.cassandra.io.util.File; + +import static java.lang.String.format; +import static java.util.stream.Collectors.toList; + +/** + * Timestamp and version encoded in the file name, e.g. + * log-1637159888484-2-1-1.data + * log-1637159888484-2-1-1.indx + * log-1637159888484-2-1-1.meta + * log-1637159888484-2-1-1.sync + */ +final class Descriptor implements Comparable<Descriptor> +{ + private static final String SEPARATOR = "-"; + private static final String PREFIX = "log" + SEPARATOR; + private static final String TMP_SUFFIX = "tmp"; + + private static final Pattern DATA_FILE_PATTERN = + Pattern.compile( PREFIX + "(\\d+)" // timestamp + + SEPARATOR + "(\\d+)" // generation + + SEPARATOR + "(\\d+)" // journal version + + SEPARATOR + "(\\d+)" // user version + + "\\." + Component.DATA.extension); + + private static final Pattern TMP_FILE_PATTERN = + Pattern.compile( PREFIX + "\\d+" // timestamp + + SEPARATOR + "\\d+" // generation + + SEPARATOR + "\\d+" // journal version + + SEPARATOR + "\\d+" // user version + + "\\." + "[a-z]+" // component extension + + "\\." + TMP_SUFFIX); + + + static final int JOURNAL_VERSION_1 = 1; + static final int CURRENT_JOURNAL_VERSION = JOURNAL_VERSION_1; + + final File directory; + final long timestamp; + final int generation; + + /** + * Serialization version for journal components; bumped as journal + * implementation evolves over time. + */ + final int journalVersion; + + /** + * Serialization version for user content - specifically journal keys + * and journal values; bumped when user logic evolves. + */ + final int userVersion; + + private Descriptor(File directory, long timestamp, int generation, int journalVersion, int userVersion) + { + this.directory = directory; + this.timestamp = timestamp; + this.generation = generation; + this.journalVersion = journalVersion; + this.userVersion = userVersion; + } + + static Descriptor create(File directory, long timestamp, int userVersion) + { + return new Descriptor(directory, timestamp, 1, CURRENT_JOURNAL_VERSION, userVersion); + } + + static Descriptor fromName(File directory, String name) + { + Matcher matcher = DATA_FILE_PATTERN.matcher(name); + if (!matcher.matches()) + throw new IllegalArgumentException("Provided filename is not valid for a data segment file"); + + long timestamp = Long.parseLong(matcher.group(1)); + int generation = Integer.parseInt(matcher.group(2)); + int journalVersion = Integer.parseInt(matcher.group(3)); + int userVersion = Integer.parseInt(matcher.group(4)); + + return new Descriptor(directory, timestamp, generation, journalVersion, userVersion); + } + + Descriptor withIncrementedGeneration() + { + return new Descriptor(directory, timestamp, generation + 1, journalVersion, userVersion); + } + + File fileFor(Component component) + { + return new File(directory, formatFileName(component)); + } + + File tmpFileFor(Component component) + { + return new File(directory, formatFileName(component) + '.' + TMP_SUFFIX); + } + + static boolean isTmpFile(File file) + { + return TMP_FILE_PATTERN.matcher(file.name()).matches(); + } + + private String formatFileName(Component component) + { + return format("%s%d%s%d%s%d%s%d.%s", + PREFIX, timestamp, + SEPARATOR, generation, + SEPARATOR, journalVersion, + SEPARATOR, userVersion, + component.extension); + } + + static List<Descriptor> list(File directory) + { + try + { + return Arrays.stream(directory.listNames((file, name) -> DATA_FILE_PATTERN.matcher(name).matches())) + .map(name -> fromName(directory, name)) + .collect(toList()); + } + catch (IOException e) + { + throw new RuntimeException(e); + } + } + + @Override + public int compareTo(Descriptor other) + { + assert this.directory.equals(other.directory); + + int cmp = Long.compare(this.timestamp, other.timestamp); + if (cmp == 0) cmp = Integer.compare(this.generation, other.generation); + if (cmp == 0) cmp = Integer.compare(this.journalVersion, other.journalVersion); + if (cmp == 0) cmp = Integer.compare(this.userVersion, other.userVersion); + return cmp; + } + + @Override + public boolean equals(Object other) + { + if (this == other) + return true; + return (other instanceof Descriptor) && equals((Descriptor) other); + } + + boolean equals(Descriptor other) + { + assert this.directory.equals(other.directory); Review Comment: ```suggestion assert this.directory.equals(other.directory) : String.format("Expected directory %s, but found %s", this.directory, other.directory); ``` found in my feedback pr ########## src/java/org/apache/cassandra/journal/Metadata.java: ########## @@ -0,0 +1,218 @@ +/* + * 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.cassandra.journal; + +import java.io.IOException; +import java.util.Map; +import java.util.Set; +import java.util.concurrent.ConcurrentHashMap; +import java.util.concurrent.atomic.AtomicIntegerFieldUpdater; +import java.util.zip.CRC32; + +import org.apache.commons.collections.set.UnmodifiableSet; + +import org.agrona.collections.Int2IntHashMap; +import org.agrona.collections.IntHashSet; +import org.apache.cassandra.io.util.*; +import org.apache.cassandra.utils.Crc; + +import static org.apache.cassandra.journal.Journal.validateCRC; +import static org.apache.cassandra.utils.FBUtilities.updateChecksumInt; + +/** + * Tracks and serializes the following information: + * - all the hosts with entries in the data segment and #of records each is tagged in; + * used for compaction prioritisation and to act in response to topology changes + * - total count of records in this segment file + * used for compaction prioritisation + */ +final class Metadata +{ + private final Set<Integer> unmodifiableHosts; + private final Map<Integer, Integer> recordsPerHost; Review Comment: to avoid new boxing, I wonder if its better to use `accord.local.Node.Id` instead of `Integer`, that way we don't add a brand new box w/e we touch the keys? ########## src/java/org/apache/cassandra/journal/SyncedOffsets.java: ########## @@ -0,0 +1,261 @@ +/* + * 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.cassandra.journal; + +import java.io.IOException; +import java.nio.MappedByteBuffer; +import java.nio.channels.FileChannel; +import java.nio.file.StandardOpenOption; +import java.util.zip.CRC32; + +import com.google.common.primitives.Ints; + +import org.apache.cassandra.io.util.File; +import org.apache.cassandra.io.util.FileUtils; +import org.apache.cassandra.io.util.RandomAccessReader; +import org.apache.cassandra.utils.Closeable; +import org.apache.cassandra.utils.Crc; +import org.apache.cassandra.utils.SyncUtil; + +import static org.apache.cassandra.utils.FBUtilities.updateChecksumInt; + +/** + * Keeps track of fsynced limits of a data file. Enables us to treat invalid + * records that are known to have been fsynced to disk differently from those + * that aren't. + * <p/> + * On disk representation is a sequence of 2-int tuples of: + * (synced offset, CRC32(tuple position in synced offsets file, synced offset)) + */ +interface SyncedOffsets extends Closeable +{ + /** + * @return furthest known synced offset + */ + int syncedOffset(); + + /** + * Record an offset as synced to disk. + * + * @param offset the offset into datafile, up to which contents have been fsynced (exclusive) + */ + void mark(int offset); + + @Override + default void close() + { + } + + /** + * @return a mutable MMAP-backed synced offset tracker for a new {@link ActiveSegment} + */ + static Active active(Descriptor descriptor, boolean syncOnMark) + { + return new Active(descriptor, syncOnMark); + } + + /** + * Load an existing log of synced offsets from disk into an immutable instance. + */ + static Static load(Descriptor descriptor) + { + return Static.load(descriptor); + } + + /** + * @return a placeholder instance in case this component is missing + */ + static Absent absent() + { + return Absent.INSTANCE; + } + + /** + * Single-threaded, MMAP-backed list of synced offsets. + */ + final class Active implements SyncedOffsets + { + private static final int INITIAL_CAPACITY = 4 << 10; + + private final Descriptor descriptor; + private final boolean syncOnMark; + + private final FileChannel channel; + private MappedByteBuffer buffer; + + private volatile int syncedOffset; + + private Active(Descriptor descriptor, boolean syncOnMark) + { + this.descriptor = descriptor; + this.syncOnMark = syncOnMark; + + File file = descriptor.fileFor(Component.SYNCED_OFFSETS); + if (file.exists()) + throw new IllegalArgumentException("Synced offsets file " + file + " already exists"); + + try + { + channel = FileChannel.open(file.toPath(), StandardOpenOption.WRITE, StandardOpenOption.READ, StandardOpenOption.CREATE); Review Comment: This causes issues with simulator ``` INFO [AsyncAppender-Worker-ASYNC] 2023-04-12 15:21:40,572 SubstituteLogger.java:169 - ERROR [AccordJournal-allocator:1] 2023-04-12 15:21:40,572 Journal.java:615 - Failed allocating journal segments. Journal AccordJournal failure policy is STOP; terminating thread. java.lang.UnsupportedOperationException: null at com.google.common.jimfs.JimfsFileChannel.map(JimfsFileChannel.java:598) at org.apache.cassandra.journal.SyncedOffsets$Active.map(SyncedOffsets.java:164) at org.apache.cassandra.journal.SyncedOffsets$Active.<init>(SyncedOffsets.java:120) at org.apache.cassandra.journal.SyncedOffsets$Active.<init>(SyncedOffsets.java:91) at org.apache.cassandra.journal.SyncedOffsets.active(SyncedOffsets.java:69) at org.apache.cassandra.journal.ActiveSegment.create(ActiveSegment.java:83) at org.apache.cassandra.journal.Journal.createSegment(Journal.java:450) at org.apache.cassandra.journal.Journal.access$200(Journal.java:75) at org.apache.cassandra.journal.Journal$AllocateRunnable.runNormal(Journal.java:392) at org.apache.cassandra.journal.Journal$AllocateRunnable.run(Journal.java:374) ``` -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
