Compress and cache cluster state during validate join request (opense…

…arch-project#7321) * Compress and cache cluster state during validate join request Signed-off-by: Aman Khare <amkhar@amazon.com> * Add changelog and license Signed-off-by: Aman Khare <amkhar@amazon.com> * Add javadoc and correct styling Signed-off-by: Aman Khare <amkhar@amazon.com> * Add new handler for sending compressed cluster state in validate join flow and refactor code Signed-off-by: Aman Khare <amkhar@amazon.com> * Refactor util method Signed-off-by: Aman Khare <amkhar@amazon.com> * optimize imports Signed-off-by: Aman Khare <amkhar@amazon.com> * Use cluster state version based cache instead of time based cache Signed-off-by: Aman Khare <amkhar@amazon.com> * style fix Signed-off-by: Aman Khare <amkhar@amazon.com> * fix styling 2 Signed-off-by: Aman Khare <amkhar@amazon.com> * Use concurrent hashmap instead of cache, add UT class for ClusterStateUtils Signed-off-by: Aman Khare <amkhar@amazon.com> * style fix Signed-off-by: Aman Khare <amkhar@amazon.com> * Use AtomicReference instead of ConcurrentHashMap Signed-off-by: Aman Khare <amkhar@amazon.com> * Use method overloading to simplify the caller code Signed-off-by: Aman Khare <amkhar@amazon.com> * Resolve conflicts Signed-off-by: Aman Khare <amkhar@amazon.com> * Change code structure to separate the flow for JoinHelper and PublicationTransportHelper Signed-off-by: Aman Khare <amkhar@amazon.com> * Remove unnecessary input.setVersion line Co-authored-by: Andrew Ross <andrross@amazon.com> Signed-off-by: Aman Khare <85096200+amkhar@users.noreply.github.com> --------- Signed-off-by: Aman Khare <amkhar@amazon.com> Signed-off-by: Aman Khare <85096200+amkhar@users.noreply.github.com> Co-authored-by: Aman Khare <amkhar@amazon.com> Co-authored-by: Andrew Ross <andrross@amazon.com>
gaiksaya · Jun 6, 2023 · b17c88c · b17c88c
1 parent 4592d7c
commit b17c88c
Show file tree

Hide file tree

Showing 9 changed files with 485 additions and 101 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -59,6 +59,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - Add ZSTD compression for snapshotting ([#2996](https://github.com/opensearch-project/OpenSearch/pull/2996))
 - Change `com.amazonaws.sdk.ec2MetadataServiceEndpointOverride` to `aws.ec2MetadataServiceEndpoint` ([7372](https://github.com/opensearch-project/OpenSearch/pull/7372/))
 - Change `com.amazonaws.sdk.stsEndpointOverride` to `aws.stsEndpointOverride` ([7372](https://github.com/opensearch-project/OpenSearch/pull/7372/))
+- Compress and cache cluster state during validate join request ([#7321](https://github.com/opensearch-project/OpenSearch/pull/7321))
 
 ### Deprecated
 

diff --git a/server/src/main/java/org/opensearch/cluster/coordination/CompressedStreamUtils.java b/server/src/main/java/org/opensearch/cluster/coordination/CompressedStreamUtils.java
@@ -0,0 +1,61 @@
+/*
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * The OpenSearch Contributors require contributions made to
+ * this file be licensed under the Apache-2.0 license or a
+ * compatible open source license.
+ */
+
+package org.opensearch.cluster.coordination;
+
+import org.apache.logging.log4j.LogManager;
+import org.apache.logging.log4j.Logger;
+import org.opensearch.Version;
+import org.opensearch.common.CheckedConsumer;
+import org.opensearch.common.bytes.BytesReference;
+import org.opensearch.common.compress.Compressor;
+import org.opensearch.common.compress.CompressorFactory;
+import org.opensearch.common.io.stream.BytesStreamOutput;
+import org.opensearch.common.io.stream.InputStreamStreamInput;
+import org.opensearch.common.io.stream.NamedWriteableAwareStreamInput;
+import org.opensearch.common.io.stream.NamedWriteableRegistry;
+import org.opensearch.common.io.stream.OutputStreamStreamOutput;
+import org.opensearch.common.io.stream.StreamInput;
+import org.opensearch.common.io.stream.StreamOutput;
+import org.opensearch.transport.BytesTransportRequest;
+
+import java.io.IOException;
+
+/**
+ * A helper class to utilize the compressed stream.
+ *
+ * @opensearch.internal
+ */
+public final class CompressedStreamUtils {
+    private static final Logger logger = LogManager.getLogger(CompressedStreamUtils.class);
+
+    public static BytesReference createCompressedStream(Version version, CheckedConsumer<StreamOutput, IOException> outputConsumer)
+        throws IOException {
+        final BytesStreamOutput bStream = new BytesStreamOutput();
+        try (StreamOutput stream = new OutputStreamStreamOutput(CompressorFactory.COMPRESSOR.threadLocalOutputStream(bStream))) {
+            stream.setVersion(version);
+            outputConsumer.accept(stream);
+        }
+        final BytesReference serializedByteRef = bStream.bytes();
+        logger.trace("serialized writable object for node version [{}] with size [{}]", version, serializedByteRef.length());
+        return serializedByteRef;
+    }
+
+    public static StreamInput decompressBytes(BytesTransportRequest request, NamedWriteableRegistry namedWriteableRegistry)
+        throws IOException {
+        final Compressor compressor = CompressorFactory.compressor(request.bytes());
+        final StreamInput in;
+        if (compressor != null) {
+            in = new InputStreamStreamInput(compressor.threadLocalInputStream(request.bytes().streamInput()));
+        } else {
+            in = request.bytes().streamInput();
+        }
+        in.setVersion(request.version());
+        return new NamedWriteableAwareStreamInput(in, namedWriteableRegistry);
+    }
+}
diff --git a/server/src/main/java/org/opensearch/cluster/coordination/Coordinator.java b/server/src/main/java/org/opensearch/cluster/coordination/Coordinator.java
@@ -222,7 +222,8 @@ public Coordinator(
             this.onJoinValidators,
             rerouteService,
             nodeHealthService,
-            this::onNodeCommissionStatusChange
+            this::onNodeCommissionStatusChange,
+            namedWriteableRegistry
         );
         this.persistedStateSupplier = persistedStateSupplier;
         this.noClusterManagerBlockService = new NoClusterManagerBlockService(settings, clusterSettings);

diff --git a/server/src/main/java/org/opensearch/cluster/coordination/JoinHelper.java b/server/src/main/java/org/opensearch/cluster/coordination/JoinHelper.java
@@ -35,6 +35,7 @@
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
 import org.apache.logging.log4j.message.ParameterizedMessage;
+import org.opensearch.Version;
 import org.opensearch.action.ActionListener;
 import org.opensearch.action.ActionListenerResponseHandler;
 import org.opensearch.cluster.ClusterState;
@@ -49,7 +50,9 @@
 import org.opensearch.cluster.routing.allocation.AllocationService;
 import org.opensearch.cluster.service.ClusterManagerService;
 import org.opensearch.common.Priority;
+import org.opensearch.common.bytes.BytesReference;
 import org.opensearch.common.collect.Tuple;
+import org.opensearch.common.io.stream.NamedWriteableRegistry;
 import org.opensearch.common.io.stream.StreamInput;
 import org.opensearch.common.settings.Setting;
 import org.opensearch.common.settings.Settings;
@@ -58,6 +61,7 @@
 import org.opensearch.monitor.StatusInfo;
 import org.opensearch.threadpool.ThreadPool;
 import org.opensearch.threadpool.ThreadPool.Names;
+import org.opensearch.transport.BytesTransportRequest;
 import org.opensearch.transport.RemoteTransportException;
 import org.opensearch.transport.TransportChannel;
 import org.opensearch.transport.TransportException;
@@ -98,6 +102,7 @@ public class JoinHelper {
 
     public static final String JOIN_ACTION_NAME = "internal:cluster/coordination/join";
     public static final String VALIDATE_JOIN_ACTION_NAME = "internal:cluster/coordination/join/validate";
+    public static final String VALIDATE_COMPRESSED_JOIN_ACTION_NAME = JOIN_ACTION_NAME + "/validate_compressed";
     public static final String START_JOIN_ACTION_NAME = "internal:cluster/coordination/start_join";
 
     // the timeout for Zen1 join attempts
@@ -122,6 +127,8 @@ public class JoinHelper {
 
     private final Supplier<JoinTaskExecutor> joinTaskExecutorGenerator;
     private final Consumer<Boolean> nodeCommissioned;
+    private final NamedWriteableRegistry namedWriteableRegistry;
+    private final AtomicReference<Tuple<Long, BytesReference>> serializedState = new AtomicReference<>();
 
     JoinHelper(
         Settings settings,
@@ -135,13 +142,16 @@ public class JoinHelper {
         Collection<BiConsumer<DiscoveryNode, ClusterState>> joinValidators,
         RerouteService rerouteService,
         NodeHealthService nodeHealthService,
-        Consumer<Boolean> nodeCommissioned
+        Consumer<Boolean> nodeCommissioned,
+        NamedWriteableRegistry namedWriteableRegistry
     ) {
         this.clusterManagerService = clusterManagerService;
         this.transportService = transportService;
         this.nodeHealthService = nodeHealthService;
         this.joinTimeout = JOIN_TIMEOUT_SETTING.get(settings);
         this.nodeCommissioned = nodeCommissioned;
+        this.namedWriteableRegistry = namedWriteableRegistry;
+
         this.joinTaskExecutorGenerator = () -> new JoinTaskExecutor(settings, allocationService, logger, rerouteService) {
 
             private final long term = currentTermSupplier.getAsLong();
@@ -208,22 +218,52 @@ public ClusterTasksResult<JoinTaskExecutor.Task> execute(ClusterState currentSta
             ThreadPool.Names.GENERIC,
             ValidateJoinRequest::new,
             (request, channel, task) -> {
-                final ClusterState localState = currentStateSupplier.get();
-                if (localState.metadata().clusterUUIDCommitted()
-                    && localState.metadata().clusterUUID().equals(request.getState().metadata().clusterUUID()) == false) {
-                    throw new CoordinationStateRejectedException(
-                        "join validation on cluster state"
-                            + " with a different cluster uuid "
-                            + request.getState().metadata().clusterUUID()
-                            + " than local cluster uuid "
-                            + localState.metadata().clusterUUID()
-                            + ", rejecting"
-                    );
-                }
-                joinValidators.forEach(action -> action.accept(transportService.getLocalNode(), request.getState()));
+                runJoinValidators(currentStateSupplier, request.getState(), joinValidators);
                 channel.sendResponse(Empty.INSTANCE);
             }
         );
+
+        transportService.registerRequestHandler(
+            VALIDATE_COMPRESSED_JOIN_ACTION_NAME,
+            ThreadPool.Names.GENERIC,
+            BytesTransportRequest::new,
+            (request, channel, task) -> {
+                handleCompressedValidateJoinRequest(currentStateSupplier, joinValidators, request);
+                channel.sendResponse(Empty.INSTANCE);
+            }
+        );
+
+    }
+
+    private void runJoinValidators(
+        Supplier<ClusterState> currentStateSupplier,
+        ClusterState incomingState,
+        Collection<BiConsumer<DiscoveryNode, ClusterState>> joinValidators
+    ) {
+        final ClusterState localState = currentStateSupplier.get();
+        if (localState.metadata().clusterUUIDCommitted()
+            && localState.metadata().clusterUUID().equals(incomingState.metadata().clusterUUID()) == false) {
+            throw new CoordinationStateRejectedException(
+                "join validation on cluster state"
+                    + " with a different cluster uuid "
+                    + incomingState.metadata().clusterUUID()
+                    + " than local cluster uuid "
+                    + localState.metadata().clusterUUID()
+                    + ", rejecting"
+            );
+        }
+        joinValidators.forEach(action -> action.accept(transportService.getLocalNode(), incomingState));
+    }
+
+    private void handleCompressedValidateJoinRequest(
+        Supplier<ClusterState> currentStateSupplier,
+        Collection<BiConsumer<DiscoveryNode, ClusterState>> joinValidators,
+        BytesTransportRequest request
+    ) throws IOException {
+        try (StreamInput input = CompressedStreamUtils.decompressBytes(request, namedWriteableRegistry)) {
+            ClusterState incomingState = ClusterState.readFrom(input, transportService.getLocalNode());
+            runJoinValidators(currentStateSupplier, incomingState, joinValidators);
+        }
     }
 
     private JoinCallback transportJoinCallback(TransportRequest request, TransportChannel channel) {
@@ -407,12 +447,42 @@ public String executor() {
     }
 
     public void sendValidateJoinRequest(DiscoveryNode node, ClusterState state, ActionListener<TransportResponse.Empty> listener) {
-        transportService.sendRequest(
-            node,
-            VALIDATE_JOIN_ACTION_NAME,
-            new ValidateJoinRequest(state),
-            new ActionListenerResponseHandler<>(listener, i -> Empty.INSTANCE, ThreadPool.Names.GENERIC)
-        );
+        if (node.getVersion().before(Version.V_3_0_0)) {
+            transportService.sendRequest(
+                node,
+                VALIDATE_JOIN_ACTION_NAME,
+                new ValidateJoinRequest(state),
+                new ActionListenerResponseHandler<>(listener, i -> Empty.INSTANCE, ThreadPool.Names.GENERIC)
+            );
+        } else {
+            try {
+                final BytesReference bytes = serializedState.updateAndGet(cachedState -> {
+                    if (cachedState == null || cachedState.v1() != state.version()) {
+                        try {
+                            return new Tuple<>(
+                                state.version(),
+                                CompressedStreamUtils.createCompressedStream(node.getVersion(), state::writeTo)
+                            );
+                        } catch (IOException e) {
+                            // mandatory as AtomicReference doesn't rethrow IOException.
+                            throw new RuntimeException(e);
+                        }
+                    } else {
+                        return cachedState;
+                    }
+                }).v2();
+                final BytesTransportRequest request = new BytesTransportRequest(bytes, node.getVersion());
+                transportService.sendRequest(
+                    node,
+                    VALIDATE_COMPRESSED_JOIN_ACTION_NAME,
+                    request,
+                    new ActionListenerResponseHandler<>(listener, i -> Empty.INSTANCE, ThreadPool.Names.GENERIC)
+                );
+            } catch (Exception e) {
+                logger.warn("error sending cluster state to {}", node);
+                listener.onFailure(e);
+            }
+        }
     }
 
     /**

diff --git a/server/src/main/java/org/opensearch/cluster/coordination/PublicationTransportHandler.java b/server/src/main/java/org/opensearch/cluster/coordination/PublicationTransportHandler.java
@@ -44,16 +44,8 @@
 import org.opensearch.cluster.node.DiscoveryNode;
 import org.opensearch.cluster.node.DiscoveryNodes;
 import org.opensearch.common.bytes.BytesReference;
-import org.opensearch.common.compress.Compressor;
-import org.opensearch.common.compress.CompressorFactory;
-import org.opensearch.common.io.stream.BytesStreamOutput;
-import org.opensearch.common.io.stream.InputStreamStreamInput;
-import org.opensearch.common.io.stream.NamedWriteableAwareStreamInput;
 import org.opensearch.common.io.stream.NamedWriteableRegistry;
-import org.opensearch.common.io.stream.OutputStreamStreamOutput;
 import org.opensearch.common.io.stream.StreamInput;
-import org.opensearch.common.io.stream.StreamOutput;
-import org.opensearch.common.util.io.IOUtils;
 import org.opensearch.threadpool.ThreadPool;
 import org.opensearch.transport.BytesTransportRequest;
 import org.opensearch.transport.TransportChannel;
@@ -168,17 +160,9 @@ public PublishClusterStateStats stats() {
     }
 
     private PublishWithJoinResponse handleIncomingPublishRequest(BytesTransportRequest request) throws IOException {
-        final Compressor compressor = CompressorFactory.compressor(request.bytes());
-        StreamInput in = request.bytes().streamInput();
-        try {
-            if (compressor != null) {
-                in = new InputStreamStreamInput(compressor.threadLocalInputStream(in));
-            }
-            in = new NamedWriteableAwareStreamInput(in, namedWriteableRegistry);
-            in.setVersion(request.version());
-            // If true we received full cluster state - otherwise diffs
+        try (StreamInput in = CompressedStreamUtils.decompressBytes(request, namedWriteableRegistry)) {
+            ClusterState incomingState;
             if (in.readBoolean()) {
-                final ClusterState incomingState;
                 // Close early to release resources used by the de-compression as early as possible
                 try (StreamInput input = in) {
                     incomingState = ClusterState.readFrom(input, transportService.getLocalNode());
@@ -198,7 +182,6 @@ private PublishWithJoinResponse handleIncomingPublishRequest(BytesTransportReque
                     incompatibleClusterStateDiffReceivedCount.incrementAndGet();
                     throw new IncompatibleClusterStateVersionException("have no local cluster state");
                 } else {
-                    ClusterState incomingState;
                     try {
                         final Diff<ClusterState> diff;
                         // Close stream early to release resources used by the de-compression as early as possible
@@ -225,8 +208,6 @@ private PublishWithJoinResponse handleIncomingPublishRequest(BytesTransportReque
                     return response;
                 }
             }
-        } finally {
-            IOUtils.close(in);
         }
     }
 
@@ -254,13 +235,10 @@ public PublicationContext newPublicationContext(ClusterChangedEvent clusterChang
     }
 
     private static BytesReference serializeFullClusterState(ClusterState clusterState, Version nodeVersion) throws IOException {
-        final BytesStreamOutput bStream = new BytesStreamOutput();
-        try (StreamOutput stream = new OutputStreamStreamOutput(CompressorFactory.COMPRESSOR.threadLocalOutputStream(bStream))) {
-            stream.setVersion(nodeVersion);
+        final BytesReference serializedState = CompressedStreamUtils.createCompressedStream(nodeVersion, stream -> {
             stream.writeBoolean(true);
             clusterState.writeTo(stream);
-        }
-        final BytesReference serializedState = bStream.bytes();
+        });
         logger.trace(
             "serialized full cluster state version [{}] for node version [{}] with size [{}]",
             clusterState.version(),
@@ -271,13 +249,10 @@ private static BytesReference serializeFullClusterState(ClusterState clusterStat
     }
 
     private static BytesReference serializeDiffClusterState(Diff<ClusterState> diff, Version nodeVersion) throws IOException {
-        final BytesStreamOutput bStream = new BytesStreamOutput();
-        try (StreamOutput stream = new OutputStreamStreamOutput(CompressorFactory.COMPRESSOR.threadLocalOutputStream(bStream))) {
-            stream.setVersion(nodeVersion);
+        return CompressedStreamUtils.createCompressedStream(nodeVersion, stream -> {
             stream.writeBoolean(false);
             diff.writeTo(stream);
-        }
-        return bStream.bytes();
+        });
     }
 
     /**

diff --git a/server/src/main/resources/org/opensearch/bootstrap/test-framework.policy b/server/src/main/resources/org/opensearch/bootstrap/test-framework.policy
@@ -153,4 +153,8 @@ grant codeBase "file:${gradle.worker.jar}" {
 grant {
   // since the gradle test worker jar is on the test classpath, our tests should be able to read it
   permission java.io.FilePermission "${gradle.worker.jar}", "read";
+  permission java.lang.RuntimePermission "accessDeclaredMembers";
+  permission java.lang.RuntimePermission "reflectionFactoryAccess";
+  permission java.lang.RuntimePermission "accessClassInPackage.sun.reflect";
+  permission java.lang.reflect.ReflectPermission "suppressAccessChecks";
 };