blob: 01955c5672b48ac482fa2bcd7160a1dd484de877 [file] [log] [blame]
/*
* 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.distributed.upgrade;
import org.junit.Test;
import org.apache.cassandra.distributed.api.ConsistencyLevel;
import static org.apache.cassandra.distributed.shared.AssertUtils.assertRows;
import static org.apache.cassandra.distributed.shared.AssertUtils.row;
/**
* Test read repair after partial deletions when the cluster nodes are on different versions.
*
* This test and {@link MixedModeReadRepairWriteTest} are separated to avoid OOM errors in CI (see CASSANDRA-16237).
*/
public class MixedModeReadRepairDeleteTest extends UpgradeTestBase
{
/**
* Test that queries repair rows that exist in both replicas but have been deleted only in one replica.
* The row deletion can be either in the upgraded or in the not-upgraded node.
*/
@Test
public void mixedModeReadRepairDeleteRow() throws Throwable
{
// rows for columns (k, c, v, s)
Object[] row1 = row(0, 1, 10, 8);
Object[] row2 = row(0, 2, 20, 8);
allUpgrades(2, 1)
.setup(cluster -> {
cluster.schemaChange(withKeyspace("CREATE TABLE %s.t (k int, c int, v int, s int static, PRIMARY KEY (k, c))"));
// insert the rows in all the nodes
String insert = withKeyspace("INSERT INTO %s.t (k, c, v, s) VALUES (?, ?, ?, ?)");
cluster.coordinator(1).execute(insert, ConsistencyLevel.ALL, row1);
cluster.coordinator(2).execute(insert, ConsistencyLevel.ALL, row2);
})
.runAfterClusterUpgrade(cluster -> {
// internally delete one row per replica
String delete = withKeyspace("DELETE FROM %s.t WHERE k=? AND c=?");
cluster.get(1).executeInternal(delete, 0, 1);
cluster.get(2).executeInternal(delete, 0, 2);
// query to trigger read repair
String query = withKeyspace("SELECT k, c, v, s FROM %s.t");
assertRows(cluster.get(1).executeInternal(query), row2);
assertRows(cluster.get(2).executeInternal(query), row1);
Object[] emptyPartition = row(0, null, null, 8);
assertRows(cluster.coordinator(2).execute(query, ConsistencyLevel.ALL), emptyPartition);
assertRows(cluster.get(1).executeInternal(query), emptyPartition);
assertRows(cluster.get(2).executeInternal(query), emptyPartition);
})
.run();
}
/**
* Test that queries repair partitions that exist in both replicas but have been deleted only in one replica.
* The partition deletion can be either in the upgraded or in the not-upgraded node.
*/
@Test
public void mixedModeReadRepairDeletePartition() throws Throwable
{
// partitions for columns (k, c, v, s)
Object[][] partition1 = new Object[][]{ row(1, 1, 11, 10), row(1, 2, 12, 10) };
Object[][] partition2 = new Object[][]{ row(2, 1, 21, 20), row(2, 2, 22, 20) };
allUpgrades(2, 1)
.setup(cluster -> {
cluster.schemaChange(withKeyspace("CREATE TABLE %s.t (k int, c int, v int, s int static, PRIMARY KEY (k, c))"));
// insert half partition in each node
String insert = withKeyspace("INSERT INTO %s.t (k, c, v, s) VALUES (?, ?, ?, ?)");
cluster.coordinator(1).execute(insert, ConsistencyLevel.ALL, partition1[0]);
cluster.coordinator(1).execute(insert, ConsistencyLevel.ALL, partition1[1]);
cluster.coordinator(1).execute(insert, ConsistencyLevel.ALL, partition2[0]);
cluster.coordinator(1).execute(insert, ConsistencyLevel.ALL, partition2[1]);
})
.runAfterClusterUpgrade(cluster -> {
// internally delete one partition per replica
String delete = withKeyspace("DELETE FROM %s.t WHERE k=?");
cluster.get(1).executeInternal(delete, 1);
cluster.get(2).executeInternal(delete, 2);
// query to trigger read repair
String query = withKeyspace("SELECT k, c, v, s FROM %s.t");
assertRows(cluster.get(1).executeInternal(query), partition2);
assertRows(cluster.get(2).executeInternal(query), partition1);
assertRows(cluster.coordinator(2).execute(query, ConsistencyLevel.ALL));
assertRows(cluster.get(1).executeInternal(query));
assertRows(cluster.get(2).executeInternal(query));
})
.run();
}
}