rmetzger commented on code in PR #6: URL: https://github.com/apache/flink-connector-redis-streams/pull/6#discussion_r3202336238
########## flink-connector-redis-streams/src/main/java/org/apache/flink/connector/redis/streams/source/enumerator/RedisStreamsSourceEnumerator.java: ########## @@ -0,0 +1,287 @@ +/* + * 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.flink.connector.redis.streams.source.enumerator; + +import org.apache.flink.annotation.Internal; +import org.apache.flink.annotation.VisibleForTesting; +import org.apache.flink.api.connector.source.SplitEnumerator; +import org.apache.flink.api.connector.source.SplitEnumeratorContext; +import org.apache.flink.api.connector.source.SplitsAssignment; +import org.apache.flink.connector.redis.streams.source.config.RedisStreamsSourceConfig; +import org.apache.flink.connector.redis.streams.source.split.RedisStreamsSourceSplit; +import org.apache.flink.util.FlinkRuntimeException; + +import io.lettuce.core.ClientOptions; +import io.lettuce.core.RedisClient; +import io.lettuce.core.RedisURI; +import io.lettuce.core.api.StatefulRedisConnection; +import io.lettuce.core.cluster.ClusterClientOptions; +import io.lettuce.core.cluster.RedisClusterClient; +import io.lettuce.core.cluster.api.StatefulRedisClusterConnection; +import io.lettuce.core.cluster.api.sync.RedisClusterCommands; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import javax.annotation.Nullable; + +import java.io.IOException; +import java.time.Duration; +import java.util.ArrayList; +import java.util.Collections; +import java.util.HashMap; +import java.util.HashSet; +import java.util.List; +import java.util.Map; +import java.util.Set; +import java.util.function.Function; + +/** + * Enumerator for the Redis Streams Source. One split per stream key, round-robin assignment. + * Bounded mode freezes a stopping entry ID per key via {@code XINFO STREAM} at startup. + */ +@Internal +public class RedisStreamsSourceEnumerator + implements SplitEnumerator<RedisStreamsSourceSplit, RedisStreamsSourceEnumeratorState> { + + private static final Logger LOG = LoggerFactory.getLogger(RedisStreamsSourceEnumerator.class); + + private final RedisStreamsSourceConfig sourceConfig; + private final SplitEnumeratorContext<RedisStreamsSourceSplit> context; + + // SplitEnumerator callbacks are serialized by the SourceCoordinator event loop, so no + // synchronization on this state is required. + private final Set<String> pendingSplitKeys; + private final Map<Integer, Set<String>> readerAssignments = new HashMap<>(); + private final Set<Integer> readersSignaledNoMoreSplits = new HashSet<>(); + private final Map<String, String> stoppingEntryIds = new HashMap<>(); + private final Function<String, String> lastGeneratedIdLookup; + + public RedisStreamsSourceEnumerator( + RedisStreamsSourceConfig sourceConfig, + SplitEnumeratorContext<RedisStreamsSourceSplit> context, + @Nullable RedisStreamsSourceEnumeratorState restoredState) { + this(sourceConfig, context, restoredState, defaultLookup(sourceConfig)); + } + + @VisibleForTesting + RedisStreamsSourceEnumerator( + RedisStreamsSourceConfig sourceConfig, + SplitEnumeratorContext<RedisStreamsSourceSplit> context, + @Nullable RedisStreamsSourceEnumeratorState restoredState, + Function<String, String> lastGeneratedIdLookup) { + this.sourceConfig = sourceConfig; + this.context = context; + this.lastGeneratedIdLookup = lastGeneratedIdLookup; + this.pendingSplitKeys = new HashSet<>(); + if (restoredState != null) { + this.pendingSplitKeys.addAll(restoredState.getPendingSplits()); + this.stoppingEntryIds.putAll(restoredState.getStoppingEntryIds()); + } else { + this.pendingSplitKeys.addAll(sourceConfig.getStreamKeys()); + } + } + + @Override + public void start() { + LOG.info("Starting Redis Streams Source Enumerator (bounded={})", sourceConfig.isBounded()); + if (sourceConfig.isBounded()) { + for (String streamKey : sourceConfig.getStreamKeys()) { + if (stoppingEntryIds.containsKey(streamKey)) { + continue; // restored from checkpoint; do not re-query XINFO. + } + String stoppingId = lastGeneratedIdLookup.apply(streamKey); + stoppingEntryIds.put(streamKey, stoppingId); + LOG.info("Bounded mode: stream {} stopping entry id = {}", streamKey, stoppingId); + } + } + dispatchAssignments(buildAssignments()); + } + + @Override + public void handleSplitRequest(int subtaskId, @Nullable String requesterHostname) { + LOG.debug("Received split request from reader {}", subtaskId); + dispatchAssignments(buildAssignments()); + } + + @Override + public void addSplitsBack(List<RedisStreamsSourceSplit> splits, int subtaskId) { + LOG.info("Adding {} splits back from reader {}", splits.size(), subtaskId); + Set<String> assigned = readerAssignments.get(subtaskId); + if (assigned != null) { + for (RedisStreamsSourceSplit split : splits) { + assigned.remove(split.splitId()); + } + if (assigned.isEmpty()) { + readerAssignments.remove(subtaskId); + } + } + readersSignaledNoMoreSplits.remove(subtaskId); + for (RedisStreamsSourceSplit split : splits) { + pendingSplitKeys.add(split.getStreamKey()); + } + dispatchAssignments(buildAssignments()); + } + + @Override + public void addReader(int subtaskId) { + LOG.info("Adding reader {}", subtaskId); + dispatchAssignments(buildAssignments()); + if (sourceConfig.isBounded() + && pendingSplitKeys.isEmpty() + && !readersSignaledNoMoreSplits.contains(subtaskId)) { + readersSignaledNoMoreSplits.add(subtaskId); + context.signalNoMoreSplits(subtaskId); + LOG.info("Signaled no more splits for reader {}", subtaskId); + } + } + + @Override + public RedisStreamsSourceEnumeratorState snapshotState(long checkpointId) { + LOG.debug("Snapshotting state for checkpoint {}", checkpointId); + return new RedisStreamsSourceEnumeratorState( + new HashSet<>(pendingSplitKeys), new HashMap<>(stoppingEntryIds)); + } + + @Override + public void close() throws IOException { + LOG.info("Closing Redis Streams Source Enumerator"); + } + + private Map<Integer, List<RedisStreamsSourceSplit>> buildAssignments() { + if (pendingSplitKeys.isEmpty()) { + return Collections.emptyMap(); + } + List<Integer> readers = new ArrayList<>(context.registeredReaders().keySet()); + if (readers.isEmpty()) { + LOG.debug("No readers available for split assignment"); + return Collections.emptyMap(); + } + Map<Integer, List<RedisStreamsSourceSplit>> assignments = new HashMap<>(); + int idx = 0; + for (String streamKey : new ArrayList<>(pendingSplitKeys)) { + int readerId = readers.get(idx % readers.size()); + String stoppingId = + sourceConfig.isBounded() ? stoppingEntryIds.get(streamKey) : null; + RedisStreamsSourceSplit split = + new RedisStreamsSourceSplit(streamKey, null, stoppingId); + assignments.computeIfAbsent(readerId, k -> new ArrayList<>()).add(split); + readerAssignments.computeIfAbsent(readerId, k -> new HashSet<>()).add(streamKey); + pendingSplitKeys.remove(streamKey); + idx++; + } + return assignments; + } + + private void dispatchAssignments(Map<Integer, List<RedisStreamsSourceSplit>> assignments) { + for (Map.Entry<Integer, List<RedisStreamsSourceSplit>> entry : assignments.entrySet()) { + int readerId = entry.getKey(); + List<RedisStreamsSourceSplit> splits = entry.getValue(); + LOG.info("Assigning {} splits to reader {}: {}", splits.size(), readerId, splits); + context.assignSplits(new SplitsAssignment<>(Map.of(readerId, splits))); + + if (sourceConfig.isBounded() + && pendingSplitKeys.isEmpty() + && !readersSignaledNoMoreSplits.contains(readerId)) { + readersSignaledNoMoreSplits.add(readerId); + context.signalNoMoreSplits(readerId); + LOG.info("Signaled no more splits for reader {}", readerId); + } + } + } + + /** Resolves {@code XINFO STREAM <key>}'s {@code last-generated-id}; "0-0" if the stream is missing. */ + private static Function<String, String> defaultLookup(RedisStreamsSourceConfig cfg) { + return streamKey -> { + try { + if (cfg.isClusterMode()) { + return lookupViaCluster(cfg, streamKey); + } + return lookupViaStandalone(cfg, streamKey); + } catch (Exception e) { + throw new FlinkRuntimeException( + "Failed to query XINFO STREAM for bounded mode setup of " + streamKey, e); + } + }; + } + + private static String lookupViaStandalone(RedisStreamsSourceConfig cfg, String streamKey) { + RedisURI.Builder uriBuilder = + RedisURI.builder() + .withHost(cfg.getHost()) + .withPort(cfg.getPort()) + .withDatabase(cfg.getDatabase()) + .withTimeout(Duration.ofSeconds(5)); + if (cfg.getPassword() != null && !cfg.getPassword().isEmpty()) { + uriBuilder.withPassword(cfg.getPassword().toCharArray()); + } + try (RedisClient client = RedisClient.create(uriBuilder.build())) { + client.setOptions(ClientOptions.builder().autoReconnect(false).build()); + try (StatefulRedisConnection<String, String> conn = client.connect()) { + return runXinfo(conn.sync(), streamKey); + } + } + } + + private static String lookupViaCluster(RedisStreamsSourceConfig cfg, String streamKey) { + List<RedisURI> seeds = new ArrayList<>(cfg.getClusterNodes().size()); + for (String node : cfg.getClusterNodes()) { + int colon = node.lastIndexOf(':'); + RedisURI.Builder b = + RedisURI.builder() + .withHost(node.substring(0, colon)) + .withPort(Integer.parseInt(node.substring(colon + 1))) + .withTimeout(Duration.ofSeconds(5)); + if (cfg.getPassword() != null && !cfg.getPassword().isEmpty()) { + b.withPassword(cfg.getPassword().toCharArray()); + } + seeds.add(b.build()); + } + try (RedisClusterClient client = RedisClusterClient.create(seeds)) { + client.setOptions(ClusterClientOptions.builder().autoReconnect(false).build()); + try (StatefulRedisClusterConnection<String, String> conn = client.connect()) { + return runXinfo(conn.sync(), streamKey); + } + } + } + + private static String runXinfo(RedisClusterCommands<String, String> cmds, String streamKey) { + try { + return extractLastGeneratedId(cmds.xinfoStream(streamKey)); + } catch (Exception e) { + LOG.warn("Could not read XINFO STREAM for {}; treating as empty: {}", streamKey, e.getMessage()); Review Comment: I wouldn't swallow the full stack trace here and only show the error message. My proposed approach will show all messages along the cause chain ```suggestion LOG.warn("Could not read XINFO STREAM for {}; treating as empty", streamKey, e); ``` -- 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]
