[GitHub] [cassandra] frankgh commented on a change in pull request #1301: Cassandra 17116

[GitBox <gi...@apache.org>]

frankgh commented on a change in pull request #1301:
URL: https://github.com/apache/cassandra/pull/1301#discussion_r744965450



##########
File path: test/distributed/org/apache/cassandra/distributed/test/streaming/StreamCloseInMiddleTest.java
##########
@@ -0,0 +1,204 @@
+/*
+ * 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.test.streaming;
+
+import java.io.IOException;
+import java.nio.ByteBuffer;
+import java.util.Arrays;
+import java.util.List;
+import java.util.concurrent.Callable;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+import java.util.stream.Collectors;
+
+import org.junit.Test;
+
+import net.bytebuddy.ByteBuddy;
+import net.bytebuddy.dynamic.loading.ClassLoadingStrategy;
+import net.bytebuddy.implementation.MethodDelegation;
+import net.bytebuddy.implementation.bind.annotation.SuperCall;
+import org.apache.cassandra.db.rows.UnfilteredRowIterator;
+import org.apache.cassandra.db.streaming.CassandraIncomingFile;
+import org.apache.cassandra.distributed.Cluster;
+import org.apache.cassandra.distributed.api.Feature;
+import org.apache.cassandra.distributed.api.IInvokableInstance;
+import org.apache.cassandra.distributed.api.LogResult;
+import org.apache.cassandra.distributed.api.TokenSupplier;
+import org.apache.cassandra.distributed.shared.ClusterUtils;
+import org.apache.cassandra.distributed.test.TestBaseImpl;
+import org.apache.cassandra.io.sstable.format.RangeAwareSSTableWriter;
+import org.apache.cassandra.io.sstable.format.big.BigTableZeroCopyWriter;
+import org.apache.cassandra.io.util.SequentialWriter;
+import org.assertj.core.api.Assertions;
+
+import static net.bytebuddy.matcher.ElementMatchers.named;
+import static net.bytebuddy.matcher.ElementMatchers.takesArguments;
+
+public class StreamCloseInMiddleTest extends TestBaseImpl
+{
+    @Test
+    public void zeroCopy() throws IOException
+    {
+        streamClose(true);
+    }
+
+    @Test
+    public void notZeroCopy() throws IOException
+    {
+        streamClose(false);
+    }
+
+    private void streamClose(boolean zeroCopyStreaming) throws IOException
+    {
+        try (Cluster cluster = Cluster.build(2)
+                                      .withTokenSupplier(TokenSupplier.evenlyDistributedTokens(3))
+                                      .withInstanceInitializer(BBHelper::install)
+                                      .withConfig(c -> c.with(Feature.values())
+                                                        .set("stream_entire_sstables", zeroCopyStreaming)
+                                                        // when die, this will try to halt JVM, which is easier to validate in the test
+                                                        // other levels require checking state of the subsystems
+                                                        .set("disk_failure_policy", "die"))
+                                      .start())
+        {
+            init(cluster);
+
+            cluster.schemaChange(withKeyspace("CREATE TABLE %s.tbl (pk int PRIMARY KEY)"));
+
+            triggerStreaming(cluster, zeroCopyStreaming);
+            // make sure disk failure policy is not triggered
+            assertNoNodeShutdown(cluster);
+
+            // now bootstrap a new node; streaming will fail
+            IInvokableInstance node3 = ClusterUtils.addInstance(cluster, c -> c.set("auto_bootstrap", true));
+            node3.startup();
+            for (String line : Arrays.asList("Error while waiting on bootstrap to complete. Bootstrap will have to be restarted", // bootstrap failed
+                                             "Some data streaming failed. Use nodetool to check bootstrap state and resume")) // didn't join ring because bootstrap failed
+                Assertions.assertThat(node3.logs().grep(line).getResult())
+                          .hasSize(1);
+
+            assertNoNodeShutdown(cluster);
+        }
+    }
+
+    private void assertNoNodeShutdown(Cluster cluster)
+    {
+        AssertionError t = null;
+        for (IInvokableInstance i : cluster.stream().collect(Collectors.toList()))
+        {
+            try
+            {
+                Assertions.assertThat(i.isShutdown()).describedAs("%s was shutdown; this is not expected", i).isFalse();
+                Assertions.assertThat(i.killAttempts()).describedAs("%s saw kill attempts; this is not expected", i).isEqualTo(0);
+            }
+            catch (AssertionError t2)
+            {
+                if (t == null)
+                    t = t2;
+                else
+                    t.addSuppressed(t2);
+            }
+        }
+        if (t != null)
+            throw t;
+    }
+
+    private static void triggerStreaming(Cluster cluster, boolean expectedEntireSSTable)
+    {
+        // preemptive log search to make sure classes load... seems there is a bug where a downed instance can't return access to logs
+        cluster.forEach(n -> n.logs().grep("testing"));
+
+        IInvokableInstance node1 = cluster.get(1);
+        IInvokableInstance node2 = cluster.get(2);
+
+        // repair will do streaming IFF there is a mismatch; so cause one
+        for (int i = 0; i < 10; i++)
+            node1.executeInternal(withKeyspace("INSERT INTO %s.tbl (pk) VALUES (?)"), i);
+
+        // we do not have metrics for streaming impl; this is only done in debug level logging... so enable debug logging if not on
+        cluster.forEach(n -> n.nodetoolResult("setlogginglevel", "streaming", "DEBUG").asserts().success());
+
+        // trigger streaming; expected to fail as streaming socket closed in the middle (currently this is an unrecoverable event)
+        long mark = node2.logs().mark();
+        node2.nodetoolResult("repair", "-full", KEYSPACE, "tbl").asserts().failure();
+
+        assertStreamingFailed(node2, expectedEntireSSTable, mark);
+    }
+
+    private static void assertStreamingFailed(IInvokableInstance node, boolean expectedEntireSSTable, long mark)
+    {
+        // make sure zero-copy streaming was done; only way to know is by checking debug logs on the recieving end

Review comment:
       ```suggestion
           // make sure zero-copy streaming was done; only way to know is by checking debug logs on the receiving end
   ```




-- 
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: pr-unsubscribe@cassandra.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: pr-unsubscribe@cassandra.apache.org
For additional commands, e-mail: pr-help@cassandra.apache.org