rreddy-22 commented on code in PR #13443: URL: https://github.com/apache/kafka/pull/13443#discussion_r1149659044
########## group-coordinator/src/main/java/org/apache/kafka/coordinator/group/assignor/ServerSideStickyRangeAssignor.java: ########## @@ -0,0 +1,268 @@ +/* + * 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. + */ + +/** + * <p>The Server Side Sticky Range Assignor inherits properties of both the range assignor and the sticky assignor. + * Properties are :- + * <ul> + * <li> 1) Each consumer must get at least one partition per topic that it is subscribed to whenever the number of consumers is + * less than or equal to the number of partitions for that topic. (Range) </li> + * <li> 2) Partitions should be assigned to consumers in a way that facilitates join operations where required. (Range) </li> + * This can only be done if the topics are co-partitioned in the first place + * Co-partitioned:- + * Two streams are co-partitioned if the following conditions are met:- + * ->The keys must have the same schemas + * ->The topics involved must have the same number of partitions + * <li> 3) Consumers should retain as much as their previous assignment as possible. (Sticky) </li> + * </ul> + * </p> + * + * <p>The algorithm works mainly in 5 steps described below + * <ul> + * <li> 1) Get a map of the consumersPerTopic created using the member subscriptions.</li> + * <li> 2) Get a list of consumers (potentiallyUnfilled) that have not met the minimum required quota for assignment AND + * get a list of sticky partitions that we want to retain in the new assignment.</li> + * <li> 3) Add consumers from potentiallyUnfilled to Unfilled if they haven't met the total required quota = minQuota + (if necessary) extraPartition </li> + * <li> 4) Get a list of available partitions by calculating the difference between total partitions and assigned sticky partitions </li> + * <li> 5) Iterate through unfilled consumers and assign partitions from available partitions </li> + * </ul> + * </p> + * + */ +package org.apache.kafka.coordinator.group.assignor; + + +import org.apache.kafka.common.Uuid; + +import java.util.List; +import java.util.Map; +import java.util.Set; +import java.util.ArrayList; +import java.util.HashSet; +import java.util.HashMap; +import java.util.Collections; + +import static java.lang.Math.min; + +public class ServerSideStickyRangeAssignor implements PartitionAssignor { + + public static final String RANGE_ASSIGNOR_NAME = "range-sticky"; + + @Override + public String name() { + return RANGE_ASSIGNOR_NAME; + } + + protected static <K, V> void putList(Map<K, List<V>> map, K key, V value) { + List<V> list = map.computeIfAbsent(key, k -> new ArrayList<>()); + list.add(value); + } + + protected static <K, V> void putSet(Map<K, Set<V>> map, K key, V value) { + Set<V> set = map.computeIfAbsent(key, k -> new HashSet<>()); + set.add(value); + } + + static class Pair<T, U> { + private final T first; + private final U second; + + public Pair(T first, U second) { + this.first = first; + this.second = second; + } + + public T getFirst() { + return first; + } + + public U getSecond() { + return second; + } + + @Override + public String toString() { + return "(" + first + ", " + second + ")"; + } + } + + // Returns a map of the list of consumers per Topic (keyed by topicId) + private Map<Uuid, List<String>> consumersPerTopic(AssignmentSpec assignmentSpec) { + Map<Uuid, List<String>> mapTopicsToConsumers = new HashMap<>(); + Map<String, AssignmentMemberSpec> membersData = assignmentSpec.members; + + for (Map.Entry<String, AssignmentMemberSpec> memberEntry : membersData.entrySet()) { + String memberId = memberEntry.getKey(); + AssignmentMemberSpec memberMetadata = memberEntry.getValue(); + List<Uuid> topics = new ArrayList<>(memberMetadata.subscribedTopics); + for (Uuid topicId: topics) { + putList(mapTopicsToConsumers, topicId, memberId); + } + } + return mapTopicsToConsumers; + } + + private Map<Uuid, List<Integer>> getAvailablePartitionsPerTopic(AssignmentSpec assignmentSpec, Map<Uuid, Set<Integer>> assignedStickyPartitionsPerTopic) { + Map<Uuid, List<Integer>> availablePartitionsPerTopic = new HashMap<>(); + Map<Uuid, AssignmentTopicMetadata> topicsMetadata = assignmentSpec.topics; + // Iterate through the topics map provided in assignmentSpec + for (Map.Entry<Uuid, AssignmentTopicMetadata> topicMetadataEntry : topicsMetadata.entrySet()) { + Uuid topicId = topicMetadataEntry.getKey(); + availablePartitionsPerTopic.put(topicId, new ArrayList<>()); + int numPartitions = topicsMetadata.get(topicId).numPartitions; + // since the loop iterates from 0 to n, the partitions will be in ascending order within the list of available partitions per topic + Set<Integer> assignedStickyPartitionsForTopic = assignedStickyPartitionsPerTopic.get(topicId); + for (int i = 0; i < numPartitions; i++) { + if (assignedStickyPartitionsForTopic == null || !assignedStickyPartitionsForTopic.contains(i)) { + availablePartitionsPerTopic.get(topicId).add(i); + } + } + } + return availablePartitionsPerTopic; + } + + @Override + public GroupAssignment assign(AssignmentSpec assignmentSpec) throws PartitionAssignorException { + Map<String, Map<Uuid, Set<Integer>>> membersWithNewAssignmentPerTopic = new HashMap<>(); + // Step 1 + Map<Uuid, List<String>> consumersPerTopic = consumersPerTopic(assignmentSpec); + // Step 2 + Map<Uuid, List<Pair<String, Integer>>> unfilledConsumersPerTopic = new HashMap<>(); + Map<Uuid, Set<Integer>> assignedStickyPartitionsPerTopic = new HashMap<>(); + + for (Map.Entry<Uuid, List<String>> topicEntry : consumersPerTopic.entrySet()) { + Uuid topicId = topicEntry.getKey(); + // For each topic we have a temporary list of consumers stored in potentiallyUnfilledConsumers. + // The list is populated with consumers that satisfy one of the two conditions :- + // 1) Consumers that have the minimum required number of partitions .i.e numPartitionsPerConsumer BUT they could be assigned an extra partition later on. + // In this case we add the consumer to the unfilled consumers map iff an extra partition needs to be assigned to it. + // 2) Consumers that don't have the minimum required partitions, so irrespective of whether they get an extra partition or not they get added to the unfilled map later. + List<Pair<String, Integer>> potentiallyUnfilledConsumers = new ArrayList<>(); + List<String> consumersForTopic = topicEntry.getValue(); + + AssignmentTopicMetadata topicData = assignmentSpec.topics.get(topicId); + int numPartitionsForTopic = topicData.numPartitions; + int numPartitionsPerConsumer = numPartitionsForTopic / consumersForTopic.size(); + // Each consumer can get only one extra partition per topic after receiving the minimum quota = numPartitionsPerConsumer + int numConsumersWithExtraPartition = numPartitionsForTopic % consumersForTopic.size(); + + for (String memberId: consumersForTopic) { + // Size of the older assignment, this will be 0 when assign is called for the first time. + // The older assignment is required when a reassignment occurs to ensure stickiness. + int currentAssignmentSize = 0; + List<Integer> currentAssignmentListForTopic = new ArrayList<>(); + // Convert the set to a list first and sort the partitions in numeric order since we want the same partition numbers from each topic + // to go to the same consumer in case of co-partitioned topics. + Set<Integer> currentAssignmentSetForTopic = assignmentSpec.members.get(memberId).currentAssignmentPerTopic.get(topicId); + // We want to make sure that the currentAssignment is not null to avoid a null pointer exception + if (currentAssignmentSetForTopic != null) { + currentAssignmentListForTopic = new ArrayList<>(currentAssignmentSetForTopic); + currentAssignmentSize = currentAssignmentListForTopic.size(); + } + // Initially, minRequiredQuota is the minimum number of partitions that each consumer should get. + // The value is at least 1 unless numConsumers subscribed is greater than numPartitions. In such cases, all consumers get assigned "extra partitions". + int minRequiredQuota = numPartitionsPerConsumer; + // We need to make sure that the partitions we're keeping are still part of the topic metadata. + // Ex:- In case the number of partitions is reduced for a topic, the sticky partitions must be part of the new set of partitions available. + for (int i = 0; i < currentAssignmentListForTopic.size() && currentAssignmentSize > 0; i++) { + // numPartitionsForTopic - 1 is the max possible partition number + if (currentAssignmentListForTopic.get(i) >= numPartitionsForTopic) { + currentAssignmentSize--; + } + } + // If there are previously assigned partitions present, we want to retain + if (currentAssignmentSize > 0) { + // We either need to retain currentSize number of partitions when currentSize < required OR required number of partitions otherwise. + int retainedPartitionsCount = min(currentAssignmentSize, minRequiredQuota); + for (int i = 0; i < retainedPartitionsCount; i++) { + Collections.sort(currentAssignmentListForTopic); + putSet(assignedStickyPartitionsPerTopic, topicId, currentAssignmentListForTopic.get(i)); + membersWithNewAssignmentPerTopic.computeIfAbsent(memberId, k -> new HashMap<>()).computeIfAbsent(topicId, k -> new HashSet<>()).add(currentAssignmentListForTopic.get(i)); + } + } + // Number of partitions left to reach the minRequiredQuota + int remaining = minRequiredQuota - currentAssignmentSize; + + // There are 3 cases w.r.t value of remaining + // 1) remaining < 0 this means that the consumer has more than the min required amount. + // It could have an extra partition, so we check for that. + if (remaining < 0 && numConsumersWithExtraPartition > 0) { + // In order to remain as sticky as possible, since the order of members can be different, we want the consumers that already had extra + // partitions to retain them if it's still required, instead of assigning the extras to the first few consumers directly. + // Ex:- If two consumers out of 3 are supposed to get an extra partition and currently 1 of them already has the extra, we want this consumer + // to retain it first and later if we have partitions left they will be assigned to the first few consumers and the unfilled map is updated + numConsumersWithExtraPartition--; + // Since we already added the minimumRequiredQuota of partitions in the previous step (until minReq - 1), we just need to + // add the extra partition which will be present at the index right after min quota is satisfied. + putSet(assignedStickyPartitionsPerTopic, topicId, currentAssignmentListForTopic.get(minRequiredQuota)); + membersWithNewAssignmentPerTopic.computeIfAbsent(memberId, k -> new HashMap<>()).computeIfAbsent(topicId, k -> new HashSet<>()).add(currentAssignmentListForTopic.get(minRequiredQuota)); + } else { + // 3) If remaining = 0 it has min req partitions but there is scope for getting an extra partition later on, so it is a potentialUnfilledConsumer. + // 4) If remaining > 0 it doesn't even have the min required partitions, and it definitely is unfilled so it should be added to potentialUnfilledConsumers. + Pair<String, Integer> newPair = new Pair<>(memberId, remaining); + potentiallyUnfilledConsumers.add(newPair); + } + } + + // Step 3 + // Iterate through potentially unfilled consumers and if remaining > 0 after considering the extra partitions assignment, add to the unfilled list. + for (Pair<String, Integer> pair : potentiallyUnfilledConsumers) { + String memberId = pair.getFirst(); + Integer remaining = pair.getSecond(); + if (numConsumersWithExtraPartition > 0) { + remaining++; Review Comment: remaining = minRequiredQuota - completedQuota when we give it an extra partition, we're essentially increasing the quota by 1 -> requiredQuota = minReq + 1 newRemaining = requiredQuota - completedQuota = minReq + 1 - completedQuota = minReq - completed + 1 = remaining + 1 -- 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: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org