/* * Copyright (c) 2014 Cisco Systems, Inc. and others. All rights reserved. * * This program and the accompanying materials are made available under the * terms of the Eclipse Public License v1.0 which accompanies this distribution, * and is available at http://www.eclipse.org/legal/epl-v10.html */ package org.opendaylight.controller.cluster.datastore; import static com.google.common.base.Verify.verify; import akka.actor.ActorRef; import akka.actor.ActorSelection; import akka.actor.Cancellable; import akka.actor.ExtendedActorSystem; import akka.actor.Props; import akka.actor.Status; import akka.actor.Status.Failure; import akka.serialization.JavaSerializer; import akka.serialization.Serialization; import com.google.common.annotations.VisibleForTesting; import com.google.common.base.Preconditions; import com.google.common.base.Ticker; import com.google.common.base.Verify; import com.google.common.collect.ImmutableList; import com.google.common.collect.ImmutableMap; import com.google.common.collect.Range; import java.io.IOException; import java.util.Arrays; import java.util.Collection; import java.util.Collections; import java.util.HashMap; import java.util.Map; import java.util.Optional; import java.util.OptionalLong; import java.util.concurrent.TimeUnit; import org.eclipse.jdt.annotation.NonNull; import org.eclipse.jdt.annotation.Nullable; import org.opendaylight.controller.cluster.access.ABIVersion; import org.opendaylight.controller.cluster.access.commands.ConnectClientRequest; import org.opendaylight.controller.cluster.access.commands.ConnectClientSuccess; import org.opendaylight.controller.cluster.access.commands.LocalHistoryRequest; import org.opendaylight.controller.cluster.access.commands.NotLeaderException; import org.opendaylight.controller.cluster.access.commands.OutOfSequenceEnvelopeException; import org.opendaylight.controller.cluster.access.commands.TransactionRequest; import org.opendaylight.controller.cluster.access.concepts.ClientIdentifier; import org.opendaylight.controller.cluster.access.concepts.FrontendIdentifier; import org.opendaylight.controller.cluster.access.concepts.LocalHistoryIdentifier; import org.opendaylight.controller.cluster.access.concepts.Request; import org.opendaylight.controller.cluster.access.concepts.RequestEnvelope; import org.opendaylight.controller.cluster.access.concepts.RequestException; import org.opendaylight.controller.cluster.access.concepts.RequestSuccess; import org.opendaylight.controller.cluster.access.concepts.RetiredGenerationException; import org.opendaylight.controller.cluster.access.concepts.RuntimeRequestException; import org.opendaylight.controller.cluster.access.concepts.SliceableMessage; import org.opendaylight.controller.cluster.access.concepts.TransactionIdentifier; import org.opendaylight.controller.cluster.access.concepts.UnsupportedRequestException; import org.opendaylight.controller.cluster.common.actor.CommonConfig; import org.opendaylight.controller.cluster.common.actor.Dispatchers; import org.opendaylight.controller.cluster.common.actor.Dispatchers.DispatcherType; import org.opendaylight.controller.cluster.common.actor.MessageTracker; import org.opendaylight.controller.cluster.common.actor.MessageTracker.Error; import org.opendaylight.controller.cluster.common.actor.MeteringBehavior; import org.opendaylight.controller.cluster.datastore.exceptions.NoShardLeaderException; import org.opendaylight.controller.cluster.datastore.identifiers.ShardIdentifier; import org.opendaylight.controller.cluster.datastore.jmx.mbeans.shard.ShardDataTreeListenerInfoMXBeanImpl; import org.opendaylight.controller.cluster.datastore.jmx.mbeans.shard.ShardMBeanFactory; import org.opendaylight.controller.cluster.datastore.jmx.mbeans.shard.ShardStats; import org.opendaylight.controller.cluster.datastore.messages.AbortTransaction; import org.opendaylight.controller.cluster.datastore.messages.ActorInitialized; import org.opendaylight.controller.cluster.datastore.messages.BatchedModifications; import org.opendaylight.controller.cluster.datastore.messages.CanCommitTransaction; import org.opendaylight.controller.cluster.datastore.messages.CloseTransactionChain; import org.opendaylight.controller.cluster.datastore.messages.CommitTransaction; import org.opendaylight.controller.cluster.datastore.messages.CreateTransaction; import org.opendaylight.controller.cluster.datastore.messages.CreateTransactionReply; import org.opendaylight.controller.cluster.datastore.messages.ForwardedReadyTransaction; import org.opendaylight.controller.cluster.datastore.messages.GetShardDataTree; import org.opendaylight.controller.cluster.datastore.messages.MakeLeaderLocal; import org.opendaylight.controller.cluster.datastore.messages.OnDemandShardState; import org.opendaylight.controller.cluster.datastore.messages.PeerAddressResolved; import org.opendaylight.controller.cluster.datastore.messages.ReadyLocalTransaction; import org.opendaylight.controller.cluster.datastore.messages.RegisterDataTreeChangeListener; import org.opendaylight.controller.cluster.datastore.messages.ShardLeaderStateChanged; import org.opendaylight.controller.cluster.datastore.messages.UpdateSchemaContext; import org.opendaylight.controller.cluster.datastore.persisted.DatastoreSnapshot; import org.opendaylight.controller.cluster.datastore.persisted.DatastoreSnapshot.ShardSnapshot; import org.opendaylight.controller.cluster.datastore.persisted.DisableTrackingPayload; import org.opendaylight.controller.cluster.messaging.MessageAssembler; import org.opendaylight.controller.cluster.messaging.MessageSlicer; import org.opendaylight.controller.cluster.messaging.SliceOptions; import org.opendaylight.controller.cluster.notifications.LeaderStateChanged; import org.opendaylight.controller.cluster.notifications.RegisterRoleChangeListener; import org.opendaylight.controller.cluster.notifications.RoleChangeNotifier; import org.opendaylight.controller.cluster.raft.LeadershipTransferFailedException; import org.opendaylight.controller.cluster.raft.RaftActor; import org.opendaylight.controller.cluster.raft.RaftActorRecoveryCohort; import org.opendaylight.controller.cluster.raft.RaftActorSnapshotCohort; import org.opendaylight.controller.cluster.raft.RaftState; import org.opendaylight.controller.cluster.raft.base.messages.FollowerInitialSyncUpStatus; import org.opendaylight.controller.cluster.raft.client.messages.OnDemandRaftState; import org.opendaylight.controller.cluster.raft.messages.AppendEntriesReply; import org.opendaylight.controller.cluster.raft.messages.RequestLeadership; import org.opendaylight.controller.cluster.raft.messages.ServerRemoved; import org.opendaylight.controller.cluster.raft.protobuff.client.messages.Payload; import org.opendaylight.yangtools.concepts.Identifier; import org.opendaylight.yangtools.yang.data.api.schema.tree.DataTree; import org.opendaylight.yangtools.yang.data.api.schema.tree.DataValidationFailedException; import org.opendaylight.yangtools.yang.data.api.schema.tree.TreeType; import org.opendaylight.yangtools.yang.model.api.SchemaContext; import org.opendaylight.yangtools.yang.model.api.SchemaContextProvider; import scala.concurrent.duration.FiniteDuration; /** * A Shard represents a portion of the logical data tree. * *

* Our Shard uses InMemoryDataTree as it's internal representation and delegates all requests it */ public class Shard extends RaftActor { @VisibleForTesting static final Object TX_COMMIT_TIMEOUT_CHECK_MESSAGE = new Object() { @Override public String toString() { return "txCommitTimeoutCheck"; } }; @VisibleForTesting static final Object GET_SHARD_MBEAN_MESSAGE = new Object() { @Override public String toString() { return "getShardMBeanMessage"; } }; static final Object RESUME_NEXT_PENDING_TRANSACTION = new Object() { @Override public String toString() { return "resumeNextPendingTransaction"; } }; // FIXME: shard names should be encapsulated in their own class and this should be exposed as a constant. public static final String DEFAULT_NAME = "default"; private static final Collection SUPPORTED_ABIVERSIONS; static { final ABIVersion[] values = ABIVersion.values(); final ABIVersion[] real = Arrays.copyOfRange(values, 1, values.length - 1); SUPPORTED_ABIVERSIONS = ImmutableList.copyOf(real).reverse(); } // FIXME: make this a dynamic property based on mailbox size and maximum number of clients private static final int CLIENT_MAX_MESSAGES = 1000; // The state of this Shard private final ShardDataTree store; /// The name of this shard private final String name; private final String shardName; private final ShardStats shardMBean; private final ShardDataTreeListenerInfoMXBeanImpl listenerInfoMXBean; private DatastoreContext datastoreContext; private final ShardCommitCoordinator commitCoordinator; private long transactionCommitTimeout; private Cancellable txCommitTimeoutCheckSchedule; private final Optional roleChangeNotifier; private final MessageTracker appendEntriesReplyTracker; private final ShardTransactionActorFactory transactionActorFactory; private final ShardSnapshotCohort snapshotCohort; private final DataTreeChangeListenerSupport treeChangeSupport = new DataTreeChangeListenerSupport(this); private ShardSnapshot restoreFromSnapshot; private final ShardTransactionMessageRetrySupport messageRetrySupport; private final FrontendMetadata frontendMetadata; private Map knownFrontends = ImmutableMap.of(); private boolean paused; private final MessageSlicer responseMessageSlicer; private final Dispatchers dispatchers; private final MessageAssembler requestMessageAssembler; protected Shard(final AbstractBuilder builder) { super(builder.getId().toString(), builder.getPeerAddresses(), Optional.of(builder.getDatastoreContext().getShardRaftConfig()), DataStoreVersions.CURRENT_VERSION); this.name = builder.getId().toString(); this.shardName = builder.getId().getShardName(); this.datastoreContext = builder.getDatastoreContext(); this.restoreFromSnapshot = builder.getRestoreFromSnapshot(); this.frontendMetadata = new FrontendMetadata(name); setPersistence(datastoreContext.isPersistent()); LOG.info("Shard created : {}, persistent : {}", name, datastoreContext.isPersistent()); ShardDataTreeChangeListenerPublisherActorProxy treeChangeListenerPublisher = new ShardDataTreeChangeListenerPublisherActorProxy(getContext(), name + "-DTCL-publisher", name); if (builder.getDataTree() != null) { store = new ShardDataTree(this, builder.getSchemaContext(), builder.getDataTree(), treeChangeListenerPublisher, name, frontendMetadata); } else { store = new ShardDataTree(this, builder.getSchemaContext(), builder.getTreeType(), builder.getDatastoreContext().getStoreRoot(), treeChangeListenerPublisher, name, frontendMetadata); } shardMBean = ShardMBeanFactory.getShardStatsMBean(name, datastoreContext.getDataStoreMXBeanType(), this); if (isMetricsCaptureEnabled()) { getContext().become(new MeteringBehavior(this)); } commitCoordinator = new ShardCommitCoordinator(store, LOG, this.name); setTransactionCommitTimeout(); // create a notifier actor for each cluster member roleChangeNotifier = createRoleChangeNotifier(name); appendEntriesReplyTracker = new MessageTracker(AppendEntriesReply.class, getRaftActorContext().getConfigParams().getIsolatedCheckIntervalInMillis()); dispatchers = new Dispatchers(context().system().dispatchers()); transactionActorFactory = new ShardTransactionActorFactory(store, datastoreContext, dispatchers.getDispatcherPath(Dispatchers.DispatcherType.Transaction), self(), getContext(), shardMBean, builder.getId().getShardName()); snapshotCohort = ShardSnapshotCohort.create(getContext(), builder.getId().getMemberName(), store, LOG, this.name); messageRetrySupport = new ShardTransactionMessageRetrySupport(this); responseMessageSlicer = MessageSlicer.builder().logContext(this.name) .messageSliceSize(datastoreContext.getMaximumMessageSliceSize()) .fileBackedStreamFactory(getRaftActorContext().getFileBackedOutputStreamFactory()) .expireStateAfterInactivity(2, TimeUnit.MINUTES).build(); requestMessageAssembler = MessageAssembler.builder().logContext(this.name) .fileBackedStreamFactory(getRaftActorContext().getFileBackedOutputStreamFactory()) .assembledMessageCallback((message, sender) -> self().tell(message, sender)) .expireStateAfterInactivity(datastoreContext.getRequestTimeout(), TimeUnit.NANOSECONDS).build(); listenerInfoMXBean = new ShardDataTreeListenerInfoMXBeanImpl(name, datastoreContext.getDataStoreMXBeanType(), self()); listenerInfoMXBean.register(); } private void setTransactionCommitTimeout() { transactionCommitTimeout = TimeUnit.MILLISECONDS.convert( datastoreContext.getShardTransactionCommitTimeoutInSeconds(), TimeUnit.SECONDS) / 2; } private Optional createRoleChangeNotifier(final String shardId) { ActorRef shardRoleChangeNotifier = this.getContext().actorOf( RoleChangeNotifier.getProps(shardId), shardId + "-notifier"); return Optional.of(shardRoleChangeNotifier); } @Override public void postStop() { LOG.info("Stopping Shard {}", persistenceId()); super.postStop(); messageRetrySupport.close(); if (txCommitTimeoutCheckSchedule != null) { txCommitTimeoutCheckSchedule.cancel(); } commitCoordinator.abortPendingTransactions("Transaction aborted due to shutdown.", this); shardMBean.unregisterMBean(); listenerInfoMXBean.unregister(); } @Override protected void handleRecover(final Object message) { LOG.debug("{}: onReceiveRecover: Received message {} from {}", persistenceId(), message.getClass(), getSender()); super.handleRecover(message); if (LOG.isTraceEnabled()) { appendEntriesReplyTracker.begin(); } } @Override protected void handleNonRaftCommand(final Object message) { try (MessageTracker.Context context = appendEntriesReplyTracker.received(message)) { final Optional maybeError = context.error(); if (maybeError.isPresent()) { LOG.trace("{} : AppendEntriesReply failed to arrive at the expected interval {}", persistenceId(), maybeError.get()); } store.resetTransactionBatch(); if (message instanceof RequestEnvelope) { handleRequestEnvelope((RequestEnvelope)message); } else if (MessageAssembler.isHandledMessage(message)) { handleRequestAssemblerMessage(message); } else if (message instanceof ConnectClientRequest) { handleConnectClient((ConnectClientRequest)message); } else if (CreateTransaction.isSerializedType(message)) { handleCreateTransaction(message); } else if (message instanceof BatchedModifications) { handleBatchedModifications((BatchedModifications)message); } else if (message instanceof ForwardedReadyTransaction) { handleForwardedReadyTransaction((ForwardedReadyTransaction) message); } else if (message instanceof ReadyLocalTransaction) { handleReadyLocalTransaction((ReadyLocalTransaction)message); } else if (CanCommitTransaction.isSerializedType(message)) { handleCanCommitTransaction(CanCommitTransaction.fromSerializable(message)); } else if (CommitTransaction.isSerializedType(message)) { handleCommitTransaction(CommitTransaction.fromSerializable(message)); } else if (AbortTransaction.isSerializedType(message)) { handleAbortTransaction(AbortTransaction.fromSerializable(message)); } else if (CloseTransactionChain.isSerializedType(message)) { closeTransactionChain(CloseTransactionChain.fromSerializable(message)); } else if (message instanceof RegisterDataTreeChangeListener) { treeChangeSupport.onMessage((RegisterDataTreeChangeListener) message, isLeader(), hasLeader()); } else if (message instanceof UpdateSchemaContext) { updateSchemaContext((UpdateSchemaContext) message); } else if (message instanceof PeerAddressResolved) { PeerAddressResolved resolved = (PeerAddressResolved) message; setPeerAddress(resolved.getPeerId(), resolved.getPeerAddress()); } else if (TX_COMMIT_TIMEOUT_CHECK_MESSAGE.equals(message)) { commitTimeoutCheck(); } else if (message instanceof DatastoreContext) { onDatastoreContext((DatastoreContext)message); } else if (message instanceof RegisterRoleChangeListener) { roleChangeNotifier.get().forward(message, context()); } else if (message instanceof FollowerInitialSyncUpStatus) { shardMBean.setFollowerInitialSyncStatus(((FollowerInitialSyncUpStatus) message).isInitialSyncDone()); context().parent().tell(message, self()); } else if (GET_SHARD_MBEAN_MESSAGE.equals(message)) { sender().tell(getShardMBean(), self()); } else if (message instanceof GetShardDataTree) { sender().tell(store.getDataTree(), self()); } else if (message instanceof ServerRemoved) { context().parent().forward(message, context()); } else if (ShardTransactionMessageRetrySupport.TIMER_MESSAGE_CLASS.isInstance(message)) { messageRetrySupport.onTimerMessage(message); } else if (message instanceof DataTreeCohortActorRegistry.CohortRegistryCommand) { store.processCohortRegistryCommand(getSender(), (DataTreeCohortActorRegistry.CohortRegistryCommand) message); } else if (message instanceof MakeLeaderLocal) { onMakeLeaderLocal(); } else if (RESUME_NEXT_PENDING_TRANSACTION.equals(message)) { store.resumeNextPendingTransaction(); } else if (!responseMessageSlicer.handleMessage(message)) { super.handleNonRaftCommand(message); } } } private void handleRequestAssemblerMessage(final Object message) { dispatchers.getDispatcher(DispatcherType.Serialization).execute(() -> { JavaSerializer.currentSystem().value_$eq((ExtendedActorSystem) context().system()); requestMessageAssembler.handleMessage(message, self()); }); } @SuppressWarnings("checkstyle:IllegalCatch") private void handleRequestEnvelope(final RequestEnvelope envelope) { final long now = ticker().read(); try { final RequestSuccess success = handleRequest(envelope, now); if (success != null) { final long executionTimeNanos = ticker().read() - now; if (success instanceof SliceableMessage) { dispatchers.getDispatcher(DispatcherType.Serialization).execute(() -> responseMessageSlicer.slice(SliceOptions.builder().identifier(success.getTarget()) .message(envelope.newSuccessEnvelope(success, executionTimeNanos)) .sendTo(envelope.getMessage().getReplyTo()).replyTo(self()) .onFailureCallback(t -> LOG.warn("Error slicing response {}", success, t)).build())); } else { envelope.sendSuccess(success, executionTimeNanos); } } } catch (RequestException e) { LOG.debug("{}: request {} failed", persistenceId(), envelope, e); envelope.sendFailure(e, ticker().read() - now); } catch (Exception e) { LOG.debug("{}: request {} caused failure", persistenceId(), envelope, e); envelope.sendFailure(new RuntimeRequestException("Request failed to process", e), ticker().read() - now); } } private void commitTimeoutCheck() { store.checkForExpiredTransactions(transactionCommitTimeout, this::updateAccess); commitCoordinator.checkForExpiredTransactions(transactionCommitTimeout, this); requestMessageAssembler.checkExpiredAssembledMessageState(); } private OptionalLong updateAccess(final SimpleShardDataTreeCohort cohort) { final FrontendIdentifier frontend = cohort.getIdentifier().getHistoryId().getClientId().getFrontendId(); final LeaderFrontendState state = knownFrontends.get(frontend); if (state == null) { // Not tell-based protocol, do nothing return OptionalLong.empty(); } if (isIsolatedLeader()) { // We are isolated and no new request can come through until we emerge from it. We are still updating // liveness of frontend when we see it attempting to communicate. Use the last access timer. return OptionalLong.of(state.getLastSeenTicks()); } // If this frontend has freshly connected, give it some time to catch up before killing its transactions. return OptionalLong.of(state.getLastConnectTicks()); } private void disableTracking(final DisableTrackingPayload payload) { final ClientIdentifier clientId = payload.getIdentifier(); LOG.debug("{}: disabling tracking of {}", persistenceId(), clientId); frontendMetadata.disableTracking(clientId); if (isLeader()) { final FrontendIdentifier frontendId = clientId.getFrontendId(); final LeaderFrontendState frontend = knownFrontends.get(frontendId); if (frontend != null) { if (clientId.equals(frontend.getIdentifier())) { if (!(frontend instanceof LeaderFrontendState.Disabled)) { verify(knownFrontends.replace(frontendId, frontend, new LeaderFrontendState.Disabled(persistenceId(), clientId, store))); LOG.debug("{}: leader state for {} disabled", persistenceId(), clientId); } else { LOG.debug("{}: leader state {} is already disabled", persistenceId(), frontend); } } else { LOG.debug("{}: leader state {} does not match {}", persistenceId(), frontend, clientId); } } else { LOG.debug("{}: leader state for {} not found", persistenceId(), clientId); knownFrontends.put(frontendId, new LeaderFrontendState.Disabled(persistenceId(), clientId, getDataStore())); } } } private void onMakeLeaderLocal() { LOG.debug("{}: onMakeLeaderLocal received", persistenceId()); if (isLeader()) { getSender().tell(new Status.Success(null), getSelf()); return; } final ActorSelection leader = getLeader(); if (leader == null) { // Leader is not present. The cluster is most likely trying to // elect a leader and we should let that run its normal course // TODO we can wait for the election to complete and retry the // request. We can also let the caller retry by sending a flag // in the response indicating the request is "reTryable". getSender().tell(new Failure( new LeadershipTransferFailedException("We cannot initiate leadership transfer to local node. " + "Currently there is no leader for " + persistenceId())), getSelf()); return; } leader.tell(new RequestLeadership(getId(), getSender()), getSelf()); } // Acquire our frontend tracking handle and verify generation matches private @Nullable LeaderFrontendState findFrontend(final ClientIdentifier clientId) throws RequestException { final LeaderFrontendState existing = knownFrontends.get(clientId.getFrontendId()); if (existing != null) { final int cmp = Long.compareUnsigned(existing.getIdentifier().getGeneration(), clientId.getGeneration()); if (cmp == 0) { existing.touch(); return existing; } if (cmp > 0) { LOG.debug("{}: rejecting request from outdated client {}", persistenceId(), clientId); throw new RetiredGenerationException(clientId.getGeneration(), existing.getIdentifier().getGeneration()); } LOG.info("{}: retiring state {}, outdated by request from client {}", persistenceId(), existing, clientId); existing.retire(); knownFrontends.remove(clientId.getFrontendId()); } else { LOG.debug("{}: client {} is not yet known", persistenceId(), clientId); } return null; } private LeaderFrontendState getFrontend(final ClientIdentifier clientId) throws RequestException { final LeaderFrontendState ret = findFrontend(clientId); if (ret != null) { return ret; } // TODO: a dedicated exception would be better, but this is technically true, too throw new OutOfSequenceEnvelopeException(0); } private static @NonNull ABIVersion selectVersion(final ConnectClientRequest message) { final Range clientRange = Range.closed(message.getMinVersion(), message.getMaxVersion()); for (ABIVersion v : SUPPORTED_ABIVERSIONS) { if (clientRange.contains(v)) { return v; } } throw new IllegalArgumentException(String.format( "No common version between backend versions %s and client versions %s", SUPPORTED_ABIVERSIONS, clientRange)); } @SuppressWarnings("checkstyle:IllegalCatch") private void handleConnectClient(final ConnectClientRequest message) { try { final ClientIdentifier clientId = message.getTarget(); final LeaderFrontendState existing = findFrontend(clientId); if (existing != null) { existing.touch(); } if (!isLeader() || !isLeaderActive()) { LOG.info("{}: not currently leader, rejecting request {}. isLeader: {}, isLeaderActive: {}," + "isLeadershipTransferInProgress: {}.", persistenceId(), message, isLeader(), isLeaderActive(), isLeadershipTransferInProgress()); throw new NotLeaderException(getSelf()); } final ABIVersion selectedVersion = selectVersion(message); final LeaderFrontendState frontend; if (existing == null) { frontend = new LeaderFrontendState.Enabled(persistenceId(), clientId, store); knownFrontends.put(clientId.getFrontendId(), frontend); LOG.debug("{}: created state {} for client {}", persistenceId(), frontend, clientId); } else { frontend = existing; } frontend.reconnect(); message.getReplyTo().tell(new ConnectClientSuccess(message.getTarget(), message.getSequence(), getSelf(), ImmutableList.of(), store.getDataTree(), CLIENT_MAX_MESSAGES).toVersion(selectedVersion), ActorRef.noSender()); } catch (RequestException | RuntimeException e) { message.getReplyTo().tell(new Failure(e), ActorRef.noSender()); } } private @Nullable RequestSuccess handleRequest(final RequestEnvelope envelope, final long now) throws RequestException { // We are not the leader, hence we want to fail-fast. if (!isLeader() || paused || !isLeaderActive()) { LOG.debug("{}: not currently active leader, rejecting request {}. isLeader: {}, isLeaderActive: {}," + "isLeadershipTransferInProgress: {}, paused: {}", persistenceId(), envelope, isLeader(), isLeaderActive(), isLeadershipTransferInProgress(), paused); throw new NotLeaderException(getSelf()); } final Request request = envelope.getMessage(); if (request instanceof TransactionRequest) { final TransactionRequest txReq = (TransactionRequest)request; final ClientIdentifier clientId = txReq.getTarget().getHistoryId().getClientId(); return getFrontend(clientId).handleTransactionRequest(txReq, envelope, now); } else if (request instanceof LocalHistoryRequest) { final LocalHistoryRequest lhReq = (LocalHistoryRequest)request; final ClientIdentifier clientId = lhReq.getTarget().getClientId(); return getFrontend(clientId).handleLocalHistoryRequest(lhReq, envelope, now); } else { LOG.warn("{}: rejecting unsupported request {}", persistenceId(), request); throw new UnsupportedRequestException(request); } } private boolean hasLeader() { return getLeaderId() != null; } public int getPendingTxCommitQueueSize() { return store.getQueueSize(); } public int getCohortCacheSize() { return commitCoordinator.getCohortCacheSize(); } @Override protected Optional getRoleChangeNotifier() { return roleChangeNotifier; } String getShardName() { return shardName; } @Override protected LeaderStateChanged newLeaderStateChanged(final String memberId, final String leaderId, final short leaderPayloadVersion) { return isLeader() ? new ShardLeaderStateChanged(memberId, leaderId, store.getDataTree(), leaderPayloadVersion) : new ShardLeaderStateChanged(memberId, leaderId, leaderPayloadVersion); } protected void onDatastoreContext(final DatastoreContext context) { datastoreContext = context; setTransactionCommitTimeout(); setPersistence(datastoreContext.isPersistent()); updateConfigParams(datastoreContext.getShardRaftConfig()); } // applyState() will be invoked once consensus is reached on the payload void persistPayload(final Identifier id, final Payload payload, final boolean batchHint) { boolean canSkipPayload = !hasFollowers() && !persistence().isRecoveryApplicable(); if (canSkipPayload) { applyState(self(), id, payload); } else { // We are faking the sender persistData(self(), id, payload, batchHint); } } private void handleCommitTransaction(final CommitTransaction commit) { final TransactionIdentifier txId = commit.getTransactionId(); if (isLeader()) { askProtocolEncountered(txId); commitCoordinator.handleCommit(txId, getSender(), this); } else { ActorSelection leader = getLeader(); if (leader == null) { messageRetrySupport.addMessageToRetry(commit, getSender(), "Could not commit transaction " + txId); } else { LOG.debug("{}: Forwarding CommitTransaction to leader {}", persistenceId(), leader); leader.forward(commit, getContext()); } } } private void handleCanCommitTransaction(final CanCommitTransaction canCommit) { final TransactionIdentifier txId = canCommit.getTransactionId(); LOG.debug("{}: Can committing transaction {}", persistenceId(), txId); if (isLeader()) { askProtocolEncountered(txId); commitCoordinator.handleCanCommit(txId, getSender(), this); } else { ActorSelection leader = getLeader(); if (leader == null) { messageRetrySupport.addMessageToRetry(canCommit, getSender(), "Could not canCommit transaction " + txId); } else { LOG.debug("{}: Forwarding CanCommitTransaction to leader {}", persistenceId(), leader); leader.forward(canCommit, getContext()); } } } @SuppressWarnings("checkstyle:IllegalCatch") protected void handleBatchedModificationsLocal(final BatchedModifications batched, final ActorRef sender) { askProtocolEncountered(batched.getTransactionId()); try { commitCoordinator.handleBatchedModifications(batched, sender, this); } catch (Exception e) { LOG.error("{}: Error handling BatchedModifications for Tx {}", persistenceId(), batched.getTransactionId(), e); sender.tell(new Failure(e), getSelf()); } } private void handleBatchedModifications(final BatchedModifications batched) { // This message is sent to prepare the modifications transaction directly on the Shard as an // optimization to avoid the extra overhead of a separate ShardTransaction actor. On the last // BatchedModifications message, the caller sets the ready flag in the message indicating // modifications are complete. The reply contains the cohort actor path (this actor) for the caller // to initiate the 3-phase commit. This also avoids the overhead of sending an additional // ReadyTransaction message. // If we're not the leader then forward to the leader. This is a safety measure - we shouldn't // normally get here if we're not the leader as the front-end (TransactionProxy) should determine // the primary/leader shard. However with timing and caching on the front-end, there's a small // window where it could have a stale leader during leadership transitions. // boolean isLeaderActive = isLeaderActive(); if (isLeader() && isLeaderActive) { handleBatchedModificationsLocal(batched, getSender()); } else { ActorSelection leader = getLeader(); if (!isLeaderActive || leader == null) { messageRetrySupport.addMessageToRetry(batched, getSender(), "Could not process BatchedModifications " + batched.getTransactionId()); } else { // If this is not the first batch and leadership changed in between batched messages, // we need to reconstruct previous BatchedModifications from the transaction // DataTreeModification, honoring the max batched modification count, and forward all the // previous BatchedModifications to the new leader. Collection newModifications = commitCoordinator .createForwardedBatchedModifications(batched, datastoreContext.getShardBatchedModificationCount()); LOG.debug("{}: Forwarding {} BatchedModifications to leader {}", persistenceId(), newModifications.size(), leader); for (BatchedModifications bm : newModifications) { leader.forward(bm, getContext()); } } } } private boolean failIfIsolatedLeader(final ActorRef sender) { if (isIsolatedLeader()) { sender.tell(new Failure(new NoShardLeaderException(String.format( "Shard %s was the leader but has lost contact with all of its followers. Either all" + " other follower nodes are down or this node is isolated by a network partition.", persistenceId()))), getSelf()); return true; } return false; } protected boolean isIsolatedLeader() { return getRaftState() == RaftState.IsolatedLeader; } @SuppressWarnings("checkstyle:IllegalCatch") private void handleReadyLocalTransaction(final ReadyLocalTransaction message) { LOG.debug("{}: handleReadyLocalTransaction for {}", persistenceId(), message.getTransactionId()); boolean isLeaderActive = isLeaderActive(); if (isLeader() && isLeaderActive) { try { commitCoordinator.handleReadyLocalTransaction(message, getSender(), this); } catch (Exception e) { LOG.error("{}: Error handling ReadyLocalTransaction for Tx {}", persistenceId(), message.getTransactionId(), e); getSender().tell(new Failure(e), getSelf()); } } else { ActorSelection leader = getLeader(); if (!isLeaderActive || leader == null) { messageRetrySupport.addMessageToRetry(message, getSender(), "Could not process ready local transaction " + message.getTransactionId()); } else { LOG.debug("{}: Forwarding ReadyLocalTransaction to leader {}", persistenceId(), leader); message.setRemoteVersion(getCurrentBehavior().getLeaderPayloadVersion()); leader.forward(message, getContext()); } } } private void handleForwardedReadyTransaction(final ForwardedReadyTransaction forwardedReady) { LOG.debug("{}: handleForwardedReadyTransaction for {}", persistenceId(), forwardedReady.getTransactionId()); boolean isLeaderActive = isLeaderActive(); if (isLeader() && isLeaderActive) { askProtocolEncountered(forwardedReady.getTransactionId()); commitCoordinator.handleForwardedReadyTransaction(forwardedReady, getSender(), this); } else { ActorSelection leader = getLeader(); if (!isLeaderActive || leader == null) { messageRetrySupport.addMessageToRetry(forwardedReady, getSender(), "Could not process forwarded ready transaction " + forwardedReady.getTransactionId()); } else { LOG.debug("{}: Forwarding ForwardedReadyTransaction to leader {}", persistenceId(), leader); ReadyLocalTransaction readyLocal = new ReadyLocalTransaction(forwardedReady.getTransactionId(), forwardedReady.getTransaction().getSnapshot(), forwardedReady.isDoImmediateCommit(), forwardedReady.getParticipatingShardNames()); readyLocal.setRemoteVersion(getCurrentBehavior().getLeaderPayloadVersion()); leader.forward(readyLocal, getContext()); } } } private void handleAbortTransaction(final AbortTransaction abort) { final TransactionIdentifier transactionId = abort.getTransactionId(); askProtocolEncountered(transactionId); doAbortTransaction(transactionId, getSender()); } void doAbortTransaction(final Identifier transactionID, final ActorRef sender) { commitCoordinator.handleAbort(transactionID, sender, this); } private void handleCreateTransaction(final Object message) { if (isLeader()) { createTransaction(CreateTransaction.fromSerializable(message)); } else if (getLeader() != null) { getLeader().forward(message, getContext()); } else { getSender().tell(new Failure(new NoShardLeaderException( "Could not create a shard transaction", persistenceId())), getSelf()); } } private void closeTransactionChain(final CloseTransactionChain closeTransactionChain) { if (isLeader()) { final LocalHistoryIdentifier id = closeTransactionChain.getIdentifier(); askProtocolEncountered(id.getClientId()); store.closeTransactionChain(id); } else if (getLeader() != null) { getLeader().forward(closeTransactionChain, getContext()); } else { LOG.warn("{}: Could not close transaction {}", persistenceId(), closeTransactionChain.getIdentifier()); } } @SuppressWarnings("checkstyle:IllegalCatch") private void createTransaction(final CreateTransaction createTransaction) { askProtocolEncountered(createTransaction.getTransactionId()); try { if (TransactionType.fromInt(createTransaction.getTransactionType()) != TransactionType.READ_ONLY && failIfIsolatedLeader(getSender())) { return; } ActorRef transactionActor = createTransaction(createTransaction.getTransactionType(), createTransaction.getTransactionId()); getSender().tell(new CreateTransactionReply(Serialization.serializedActorPath(transactionActor), createTransaction.getTransactionId(), createTransaction.getVersion()).toSerializable(), getSelf()); } catch (Exception e) { getSender().tell(new Failure(e), getSelf()); } } private ActorRef createTransaction(final int transactionType, final TransactionIdentifier transactionId) { LOG.debug("{}: Creating transaction : {} ", persistenceId(), transactionId); return transactionActorFactory.newShardTransaction(TransactionType.fromInt(transactionType), transactionId); } // Called on leader only private void askProtocolEncountered(final TransactionIdentifier transactionId) { askProtocolEncountered(transactionId.getHistoryId().getClientId()); } // Called on leader only private void askProtocolEncountered(final ClientIdentifier clientId) { final FrontendIdentifier frontend = clientId.getFrontendId(); final LeaderFrontendState state = knownFrontends.get(frontend); if (!(state instanceof LeaderFrontendState.Disabled)) { LOG.debug("{}: encountered ask-based client {}, disabling transaction tracking", persistenceId(), clientId); if (knownFrontends.isEmpty()) { knownFrontends = new HashMap<>(); } knownFrontends.put(frontend, new LeaderFrontendState.Disabled(persistenceId(), clientId, getDataStore())); persistPayload(clientId, DisableTrackingPayload.create(clientId, datastoreContext.getInitialPayloadSerializedBufferCapacity()), false); } } private void updateSchemaContext(final UpdateSchemaContext message) { updateSchemaContext(message.getSchemaContext()); } @VisibleForTesting void updateSchemaContext(final SchemaContext schemaContext) { store.updateSchemaContext(schemaContext); } private boolean isMetricsCaptureEnabled() { CommonConfig config = new CommonConfig(getContext().system().settings().config()); return config.isMetricCaptureEnabled(); } @Override @VisibleForTesting public RaftActorSnapshotCohort getRaftActorSnapshotCohort() { return snapshotCohort; } @Override protected RaftActorRecoveryCohort getRaftActorRecoveryCohort() { if (restoreFromSnapshot == null) { return ShardRecoveryCoordinator.create(store, persistenceId(), LOG); } return ShardRecoveryCoordinator.forSnapshot(store, persistenceId(), LOG, restoreFromSnapshot.getSnapshot()); } @Override protected void onRecoveryComplete() { restoreFromSnapshot = null; //notify shard manager getContext().parent().tell(new ActorInitialized(), getSelf()); // Being paranoid here - this method should only be called once but just in case... if (txCommitTimeoutCheckSchedule == null) { // Schedule a message to be periodically sent to check if the current in-progress // transaction should be expired and aborted. FiniteDuration period = FiniteDuration.create(transactionCommitTimeout / 3, TimeUnit.MILLISECONDS); txCommitTimeoutCheckSchedule = getContext().system().scheduler().schedule( period, period, getSelf(), TX_COMMIT_TIMEOUT_CHECK_MESSAGE, getContext().dispatcher(), ActorRef.noSender()); } } @Override protected void applyState(final ActorRef clientActor, final Identifier identifier, final Object data) { if (data instanceof Payload) { if (data instanceof DisableTrackingPayload) { disableTracking((DisableTrackingPayload) data); return; } try { store.applyReplicatedPayload(identifier, (Payload)data); } catch (DataValidationFailedException | IOException e) { LOG.error("{}: Error applying replica {}", persistenceId(), identifier, e); } } else { LOG.error("{}: Unknown state for {} received {}", persistenceId(), identifier, data); } } @Override protected void onStateChanged() { boolean isLeader = isLeader(); boolean hasLeader = hasLeader(); treeChangeSupport.onLeadershipChange(isLeader, hasLeader); // If this actor is no longer the leader close all the transaction chains if (!isLeader) { if (LOG.isDebugEnabled()) { LOG.debug( "{}: onStateChanged: Closing all transaction chains because shard {} is no longer the leader", persistenceId(), getId()); } paused = false; store.purgeLeaderState(); } if (hasLeader && !isIsolatedLeader()) { messageRetrySupport.retryMessages(); } } @Override protected void onLeaderChanged(final String oldLeader, final String newLeader) { shardMBean.incrementLeadershipChangeCount(); paused = false; if (!isLeader()) { if (!knownFrontends.isEmpty()) { LOG.debug("{}: removing frontend state for {}", persistenceId(), knownFrontends.keySet()); knownFrontends = ImmutableMap.of(); } requestMessageAssembler.close(); if (!hasLeader()) { // No leader anywhere, nothing else to do return; } // Another leader was elected. If we were the previous leader and had pending transactions, convert // them to transaction messages and send to the new leader. ActorSelection leader = getLeader(); if (leader != null) { Collection messagesToForward = convertPendingTransactionsToMessages(); if (!messagesToForward.isEmpty()) { LOG.debug("{}: Forwarding {} pending transaction messages to leader {}", persistenceId(), messagesToForward.size(), leader); for (Object message : messagesToForward) { LOG.debug("{}: Forwarding pending transaction message {}", persistenceId(), message); leader.tell(message, self()); } } } else { commitCoordinator.abortPendingTransactions("The transacton was aborted due to inflight leadership " + "change and the leader address isn't available.", this); } } else { // We have become the leader, we need to reconstruct frontend state knownFrontends = Verify.verifyNotNull(frontendMetadata.toLeaderState(this)); LOG.debug("{}: became leader with frontend state for {}", persistenceId(), knownFrontends.keySet()); } if (!isIsolatedLeader()) { messageRetrySupport.retryMessages(); } } /** * Clears all pending transactions and converts them to messages to be forwarded to a new leader. * * @return the converted messages */ public Collection convertPendingTransactionsToMessages() { return commitCoordinator.convertPendingTransactionsToMessages( datastoreContext.getShardBatchedModificationCount()); } @Override protected void pauseLeader(final Runnable operation) { LOG.debug("{}: In pauseLeader, operation: {}", persistenceId(), operation); paused = true; // Tell-based protocol can replay transaction state, so it is safe to blow it up when we are paused. if (datastoreContext.isUseTellBasedProtocol()) { knownFrontends.values().forEach(LeaderFrontendState::retire); knownFrontends = ImmutableMap.of(); } store.setRunOnPendingTransactionsComplete(operation); } @Override protected void unpauseLeader() { LOG.debug("{}: In unpauseLeader", persistenceId()); paused = false; store.setRunOnPendingTransactionsComplete(null); // Restore tell-based protocol state as if we were becoming the leader knownFrontends = Verify.verifyNotNull(frontendMetadata.toLeaderState(this)); } @Override protected OnDemandRaftState.AbstractBuilder newOnDemandRaftStateBuilder() { return OnDemandShardState.newBuilder().treeChangeListenerActors(treeChangeSupport.getListenerActors()) .commitCohortActors(store.getCohortActors()); } @Override public String persistenceId() { return this.name; } @VisibleForTesting ShardCommitCoordinator getCommitCoordinator() { return commitCoordinator; } public DatastoreContext getDatastoreContext() { return datastoreContext; } @VisibleForTesting public ShardDataTree getDataStore() { return store; } @VisibleForTesting ShardStats getShardMBean() { return shardMBean; } public static Builder builder() { return new Builder(); } public abstract static class AbstractBuilder, S extends Shard> { private final Class shardClass; private ShardIdentifier id; private Map peerAddresses = Collections.emptyMap(); private DatastoreContext datastoreContext; private SchemaContextProvider schemaContextProvider; private DatastoreSnapshot.ShardSnapshot restoreFromSnapshot; private DataTree dataTree; private volatile boolean sealed; protected AbstractBuilder(final Class shardClass) { this.shardClass = shardClass; } protected void checkSealed() { Preconditions.checkState(!sealed, "Builder isalready sealed - further modifications are not allowed"); } @SuppressWarnings("unchecked") private T self() { return (T) this; } public T id(final ShardIdentifier newId) { checkSealed(); this.id = newId; return self(); } public T peerAddresses(final Map newPeerAddresses) { checkSealed(); this.peerAddresses = newPeerAddresses; return self(); } public T datastoreContext(final DatastoreContext newDatastoreContext) { checkSealed(); this.datastoreContext = newDatastoreContext; return self(); } public T schemaContextProvider(final SchemaContextProvider newSchemaContextProvider) { checkSealed(); this.schemaContextProvider = Preconditions.checkNotNull(newSchemaContextProvider); return self(); } public T restoreFromSnapshot(final DatastoreSnapshot.ShardSnapshot newRestoreFromSnapshot) { checkSealed(); this.restoreFromSnapshot = newRestoreFromSnapshot; return self(); } public T dataTree(final DataTree newDataTree) { checkSealed(); this.dataTree = newDataTree; return self(); } public ShardIdentifier getId() { return id; } public Map getPeerAddresses() { return peerAddresses; } public DatastoreContext getDatastoreContext() { return datastoreContext; } public SchemaContext getSchemaContext() { return Verify.verifyNotNull(schemaContextProvider.getSchemaContext()); } public DatastoreSnapshot.ShardSnapshot getRestoreFromSnapshot() { return restoreFromSnapshot; } public DataTree getDataTree() { return dataTree; } public TreeType getTreeType() { switch (datastoreContext.getLogicalStoreType()) { case CONFIGURATION: return TreeType.CONFIGURATION; case OPERATIONAL: return TreeType.OPERATIONAL; default: throw new IllegalStateException("Unhandled logical store type " + datastoreContext.getLogicalStoreType()); } } protected void verify() { Preconditions.checkNotNull(id, "id should not be null"); Preconditions.checkNotNull(peerAddresses, "peerAddresses should not be null"); Preconditions.checkNotNull(datastoreContext, "dataStoreContext should not be null"); Preconditions.checkNotNull(schemaContextProvider, "schemaContextProvider should not be null"); } public Props props() { sealed = true; verify(); return Props.create(shardClass, this); } } public static class Builder extends AbstractBuilder { Builder() { super(Shard.class); } } Ticker ticker() { return Ticker.systemTicker(); } void scheduleNextPendingTransaction() { self().tell(RESUME_NEXT_PENDING_TRANSACTION, ActorRef.noSender()); } }