yifan-c commented on code in PR #4629: URL: https://github.com/apache/cassandra/pull/4629#discussion_r2850993539
########## test/unit/org/apache/cassandra/db/compression/CompressionDictionaryTrainingFrequencyTest.java: ########## @@ -0,0 +1,229 @@ +/* + * 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.db.compression; + +import java.time.Instant; +import java.time.temporal.ChronoUnit; +import java.util.Arrays; +import java.util.List; +import java.util.concurrent.TimeUnit; +import java.util.regex.Matcher; +import java.util.regex.Pattern; + +import com.google.common.util.concurrent.Uninterruptibles; + +import org.junit.Assert; +import org.junit.BeforeClass; +import org.junit.Test; + +import org.apache.cassandra.config.DurationSpec; +import org.apache.cassandra.cql3.CQLTester; +import org.apache.cassandra.db.ColumnFamilyStore; +import org.apache.cassandra.db.compression.CompressionDictionaryDetailsTabularData.CompressionDictionaryDataObject; +import org.apache.cassandra.io.compress.IDictionaryCompressor; +import org.apache.cassandra.io.util.File; +import org.apache.cassandra.io.util.FileUtils; +import org.apache.cassandra.schema.SystemDistributedKeyspace; +import org.apache.cassandra.tools.ToolRunner; +import org.apache.cassandra.utils.JsonUtils; +import org.apache.cassandra.utils.Pair; + +import static java.lang.String.format; +import static org.apache.cassandra.tools.ToolRunner.invokeNodetool; +import static org.assertj.core.api.Assertions.assertThat; +import static org.junit.Assert.assertEquals; +import static org.junit.Assert.assertFalse; +import static org.junit.Assert.assertNotNull; +import static org.junit.Assert.assertTrue; + +public class CompressionDictionaryTrainingFrequencyTest extends CQLTester +{ + private static final String tableName = "mytable"; + + @BeforeClass + public static void setup() throws Throwable + { + requireNetwork(); + startJMXServer(); + } + + @Test + public void testTrainingFrequency() throws Throwable + { + // we can train twice when no limit is imposed + String tableId = createDictTable(IDictionaryCompressor.DEFAULT_TRAINING_MIN_FREQUENCY); + trainDict(); + trainDict(); + + assertDicts(2, tableId); + + alterDictTable("5m"); + + assertDictTrainingFails("5m"); + + alterDictTable("1m"); + + // we can train again as 1 minute from the last training has passed + Uninterruptibles.sleepUninterruptibly(1, TimeUnit.MINUTES); + trainDict(); + assertDicts(3, tableId); + + // resetting back to 0, so we can train whenever we want + alterDictTable("0m"); + trainDict(); + assertDicts(4, tableId); + + alterDictTable("10m"); + + Pair<CompressionDictionaryDataObject, File> export = export(); + assertFailingImport(export.right); + + alterDictTable("0m"); + assertSuccessfulImport(export.right); + } + + private String getTableId() + { + ColumnFamilyStore cfs = ColumnFamilyStore.getIfExists(keyspace(), tableName); + Assert.assertNotNull(cfs); + return cfs.metadata.id.toLongString(); + } + + private String createDictTable(String frequency) + { + schemaChange(format("CREATE TABLE %s.%s (id int PRIMARY KEY, data text)" + + " WITH compression = {'class': 'ZstdDictionaryCompressor', '%s': %s}", + keyspace(), + tableName, + IDictionaryCompressor.TRAINING_MIN_FREQUENCY_PARAMETER_NAME, + frequency)); + + return getTableId(); + } + + private void alterDictTable(String trainingFrequency) + { + schemaChange(format("ALTER TABLE %s.%s WITH compression = {'class': 'ZstdDictionaryCompressor', '%s': %s}", + keyspace(), + tableName, + IDictionaryCompressor.TRAINING_MIN_FREQUENCY_PARAMETER_NAME, + trainingFrequency)); + } + + private void assertDicts(int expectedDicts, String tableId) + { + List<CompressionDictionary.LightweightCompressionDictionary> dicts = SystemDistributedKeyspace.retrieveLightweightCompressionDictionaries(); + assertNotNull(dicts); + assertEquals(expectedDicts, dicts.size()); + for (int i = 0; i < expectedDicts; i++) + assertEquals(tableId, dicts.get(i).tableId); + } + + private void trainDict() + { + createSSTables(); + + // Test training command with --force since we have limited test data + ToolRunner.ToolResult result = invokeNodetool("compressiondictionary", "train", "--force", keyspace(), tableName); + result.assertOnCleanExit(); + + assertThat(result.getStdout()) + .as("Should indicate training completed") + .contains("Training completed successfully") + .contains(keyspace()) + .contains(tableName); + } + + private void assertDictTrainingFails(String frequency) + { + createSSTables(); + + // Test training command with --force since we have limited test data + ToolRunner.ToolResult result = invokeNodetool("compressiondictionary", "train", "--force", keyspace(), tableName); + Assert.assertEquals(1, result.getExitCode()); Review Comment: Nit: can we use assertJ API for consistency ########## doc/modules/cassandra/pages/managing/operating/compression.adoc: ########## @@ -312,13 +313,13 @@ ALTER TABLE keyspace.table 'class': 'ZstdDictionaryCompressor', 'compression_level': '3', 'training_max_total_sample_size': '20MiB', - 'training_max_dictionary_size': '128KiB' + 'training_max_dictionary_size': '128KiB', + 'min_training_frequency': '1d' }; ---- -It is possible to override these training parameters by `nodetool compressiondictionary train` command as -explained in the section futher down below. If `train` subcommand do not override them, CQL parameters are -taken into account. +It is possible to override these training parameters by `nodetool compressiondictionary train` command (except `min_training_frequency` which is configurable via CQL only) as +explained in the section futher down below. If `train` subcommand do not override them, CQL parameters are taken into account. Review Comment: An pre-existing typo. `futher` -> `further` ########## test/unit/org/apache/cassandra/db/compression/CompressionDictionaryTrainingFrequencyTest.java: ########## @@ -0,0 +1,229 @@ +/* + * 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.db.compression; + +import java.time.Instant; +import java.time.temporal.ChronoUnit; +import java.util.Arrays; +import java.util.List; +import java.util.concurrent.TimeUnit; +import java.util.regex.Matcher; +import java.util.regex.Pattern; + +import com.google.common.util.concurrent.Uninterruptibles; + +import org.junit.Assert; +import org.junit.BeforeClass; +import org.junit.Test; + +import org.apache.cassandra.config.DurationSpec; +import org.apache.cassandra.cql3.CQLTester; +import org.apache.cassandra.db.ColumnFamilyStore; +import org.apache.cassandra.db.compression.CompressionDictionaryDetailsTabularData.CompressionDictionaryDataObject; +import org.apache.cassandra.io.compress.IDictionaryCompressor; +import org.apache.cassandra.io.util.File; +import org.apache.cassandra.io.util.FileUtils; +import org.apache.cassandra.schema.SystemDistributedKeyspace; +import org.apache.cassandra.tools.ToolRunner; +import org.apache.cassandra.utils.JsonUtils; +import org.apache.cassandra.utils.Pair; + +import static java.lang.String.format; +import static org.apache.cassandra.tools.ToolRunner.invokeNodetool; +import static org.assertj.core.api.Assertions.assertThat; +import static org.junit.Assert.assertEquals; +import static org.junit.Assert.assertFalse; +import static org.junit.Assert.assertNotNull; +import static org.junit.Assert.assertTrue; + +public class CompressionDictionaryTrainingFrequencyTest extends CQLTester +{ + private static final String tableName = "mytable"; + + @BeforeClass + public static void setup() throws Throwable + { + requireNetwork(); + startJMXServer(); + } + + @Test + public void testTrainingFrequency() throws Throwable + { + // we can train twice when no limit is imposed + String tableId = createDictTable(IDictionaryCompressor.DEFAULT_TRAINING_MIN_FREQUENCY); + trainDict(); + trainDict(); + + assertDicts(2, tableId); + + alterDictTable("5m"); + + assertDictTrainingFails("5m"); + + alterDictTable("1m"); + + // we can train again as 1 minute from the last training has passed + Uninterruptibles.sleepUninterruptibly(1, TimeUnit.MINUTES); Review Comment: nit: avoid sleep ideally. Maybe we can mock the time ########## doc/modules/cassandra/pages/managing/operating/compression.adoc: ########## @@ -302,6 +302,7 @@ These parameters are meant to be configured via CQL for each respective table if * `training_max_total_sample_size` (default: `10MiB`): Maximum total size of sample data to collect for training, approximately 10MB. This parameter is configured in the table's compression options for `ZstdDictionaryCompressor`. * `training_max_dictionary_size` (default: `64KiB`): Maximum size of trained dictionaries in bytes. Larger dictionaries can capture more patterns but increase memory overhead. This is a parameter of `ZstdDictionaryCompressor` of a table, in `compression` section. +* `min_training_frequency` (default: `0m`): Minimum time which needs to pass until we can train another compression dictionary. For example, if this property is set to `1h`, then we can train another dictionary no earlier than 1 hour after the last training was conducted. `0m`, default, means we can train as frequently as we want. The purpose of this parameter is to prevent excessive training which might not make sense from operational and performance perspective. If an operator wants to prevent training altogether, they can set this property to overly big value, like 36500d which means that next training can occur at least after 100 years from the last one, effectively banning any training. The minimum resolution of `min_training_frequency` is in minutes. Accepts minutes (`m`), hours (`h`) and days (`d`). Review Comment: nit: `banning` -> `disabling` -- 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]
