mimaison commented on code in PR #14734: URL: https://github.com/apache/kafka/pull/14734#discussion_r1391134864
########## metadata/src/main/java/org/apache/kafka/metadata/placement/StripedReplicaPlacer.java: ########## @@ -33,89 +33,86 @@ /** * The striped replica placer. - * - * - * GOALS - * The design of this placer attempts to satisfy a few competing goals. Firstly, we want - * to spread the replicas as evenly as we can across racks. In the simple case where - * broker racks have not been configured, this goal is a no-op, of course. But it is the + * <p> + * <h3>Goals</h3> + * <p>The design of this placer attempts to satisfy a few competing goals. Firstly, we want + * to spread the replicas as evenly as we can across racks. In the simple case where + * broker racks have not been configured, this goal is a no-op, of course. But it is the * highest priority goal in multi-rack clusters. * - * Our second goal is to spread the replicas evenly across brokers. Since we are placing + * <p>Our second goal is to spread the replicas evenly across brokers. Since we are placing * multiple partitions, we try to avoid putting each partition on the same set of - * replicas, even if it does satisfy the rack placement goal. If any specific broker is + * replicas, even if it does satisfy the rack placement goal. If any specific broker is * fenced, we would like the new leaders to distributed evenly across the remaining * brokers. * - * However, we treat the rack placement goal as higher priority than this goal-- if you + * <p>However, we treat the rack placement goal as higher priority than this goal-- if you * configure 10 brokers in rack A and B, and 1 broker in rack C, you will end up with a * lot of partitions on that one broker in rack C. If you were to place a lot of * partitions with replication factor 3, each partition would try to get a replica there. * In general racks are supposed to be about the same size -- if they aren't, this is a * user error. * - * Finally, we would prefer to place replicas on unfenced brokers, rather than on fenced + * <p>Finally, we would prefer to place replicas on unfenced brokers, rather than on fenced * brokers. - * - * - * CONSTRAINTS - * In addition to these goals, we have two constraints. Unlike the goals, these are not - * optional -- they are mandatory. Placement will fail if a constraint cannot be - * satisfied. The first constraint is that we can't place more than one replica on the - * same broker. This imposes an upper limit on replication factor-- for example, a 3-node - * cluster can't have any topics with replication factor 4. This constraint comes from + * <p> Review Comment: Same for the other instance below. -- 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