pvary commented on code in PR #10308: URL: https://github.com/apache/iceberg/pull/10308#discussion_r1608734772
########## flink/v1.19/flink/src/main/java/org/apache/iceberg/flink/maintenance/operator/MonitorSource.java: ########## @@ -0,0 +1,200 @@ +/* + * 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.iceberg.flink.maintenance.operator; + +import java.io.IOException; +import java.util.Iterator; +import org.apache.flink.api.common.typeinfo.TypeInformation; +import org.apache.flink.api.connector.source.Boundedness; +import org.apache.flink.api.connector.source.SourceReader; +import org.apache.flink.api.connector.source.SourceReaderContext; +import org.apache.flink.api.connector.source.util.ratelimit.RateLimitedSourceReader; +import org.apache.flink.api.connector.source.util.ratelimit.RateLimiter; +import org.apache.flink.api.connector.source.util.ratelimit.RateLimiterStrategy; +import org.apache.flink.core.io.SimpleVersionedSerializer; +import org.apache.flink.core.memory.DataInputDeserializer; +import org.apache.flink.core.memory.DataOutputSerializer; +import org.apache.iceberg.Snapshot; +import org.apache.iceberg.Table; +import org.apache.iceberg.flink.TableLoader; +import org.apache.iceberg.relocated.com.google.common.annotations.VisibleForTesting; +import org.apache.iceberg.relocated.com.google.common.base.MoreObjects; +import org.apache.iceberg.relocated.com.google.common.base.Preconditions; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +/** Monitors an Iceberg table for changes */ +public class MonitorSource extends SingleThreadedIteratorSource<TableChange> { + private static final Logger LOG = LoggerFactory.getLogger(MonitorSource.class); + + private final TableLoader tableLoader; + private final RateLimiterStrategy rateLimiterStrategy; + private final long maxReadBack; + + /** + * Creates a {@link org.apache.flink.api.connector.source.Source} which monitors an Iceberg table + * for changes. + * + * @param tableLoader used for accessing the table + * @param rateLimiterStrategy limits the frequency the table is checked + * @param maxReadBack sets the number of snapshots read before stopping change collection + */ + public MonitorSource( + TableLoader tableLoader, RateLimiterStrategy rateLimiterStrategy, long maxReadBack) { + Preconditions.checkNotNull(tableLoader, "Table loader should no be null"); + Preconditions.checkNotNull(rateLimiterStrategy, "Rate limiter strategy should no be null"); + Preconditions.checkArgument(maxReadBack > 0, "Need to read at least 1 snapshot to work"); + + this.tableLoader = tableLoader; + this.rateLimiterStrategy = rateLimiterStrategy; + this.maxReadBack = maxReadBack; + } + + @Override + public Boundedness getBoundedness() { + return Boundedness.CONTINUOUS_UNBOUNDED; + } + + @Override + public TypeInformation<TableChange> getProducedType() { + return TypeInformation.of(TableChange.class); + } + + @Override + Iterator<TableChange> createIterator() { + return new SchedulerEventIterator(tableLoader, null, maxReadBack); + } + + @Override + SimpleVersionedSerializer<Iterator<TableChange>> getIteratorSerializer() { + return new SchedulerEventIteratorSerializer(tableLoader, maxReadBack); + } + + @Override + public SourceReader<TableChange, GlobalSplit<TableChange>> createReader( + SourceReaderContext readerContext) throws Exception { + RateLimiter rateLimiter = rateLimiterStrategy.createRateLimiter(1); + return new RateLimitedSourceReader<>(super.createReader(readerContext), rateLimiter); + } + + /** The Iterator which returns the latest changes on an Iceberg table. */ + @VisibleForTesting + static class SchedulerEventIterator implements Iterator<TableChange> { + private Long lastSnapshotId; + private final long maxReadBack; + private final Table table; + + SchedulerEventIterator(TableLoader tableLoader, Long lastSnapshotId, long maxReadBack) { + this.lastSnapshotId = lastSnapshotId; + this.maxReadBack = maxReadBack; + tableLoader.open(); + this.table = tableLoader.loadTable(); + } + + @Override + public boolean hasNext() { + return true; + } + + @Override + public TableChange next() { + try { + table.refresh(); + Snapshot currentSnapshot = table.currentSnapshot(); + Long current = currentSnapshot != null ? currentSnapshot.snapshotId() : null; + Long checking = current; + TableChange event = new TableChange(0, 0, 0L, 0L, 0); + long readBack = 0; + while (checking != null && !checking.equals(lastSnapshotId) && ++readBack <= maxReadBack) { Review Comment: > `maxReadBack` is another config that users need to know how to config. it may not make much difference to check from the last snapshot to the current snapshot, since the snapshots are loaded anyway as a list in `TableMetadata`. We need to have the added data files/added data size etc. for the snapshots - this means we need to read the manifest files. This could become costly after some size. I have faced a situation with a table which was never compacted, that it had ~2000 snapshots (my test table left for the weekend with 2 min checkpointing period - 60*24*3/2) and it took serious time to read all of the snapshots (more than 20 min) - In this case the parallel read might help if we want "correct" results, but as I have mentioned before, we don't really need exact data in this case. So having 100 as a default, and allow for override if necessary would be better. Normal users don't have to bother with it, advanced users can play around with it, if they want - also I would be fine with not allowing to change it from 100 if that is the community decision. -- 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: issues-unsubscr...@iceberg.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org For additional commands, e-mail: issues-h...@iceberg.apache.org
