You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The round robin balancer seems to not be balancing on a per topic basis and instead balances between partitions across all topics messages are enqueued for, ignoring different topics.
E.g. the partition is increased by 1 for each new message, no matter if it is send to the same topic as the previous one. In case of 4 messages written onto topic1, topic2, topic1, topic2 the partitions are chosen as follows:
Topic 1, Partition 0
Topic 2, Partition 1
Topic 1, Partition 2
Topic 2, Partition 3
Resulting in half of the partitions on each topic not receiving any messages. Instead I'd expect the following behavior:
Describe the bug
The round robin balancer seems to not be balancing on a per topic basis and instead balances between partitions across all topics messages are enqueued for, ignoring different topics.
E.g. the partition is increased by 1 for each new message, no matter if it is send to the same topic as the previous one. In case of 4 messages written onto topic1, topic2, topic1, topic2 the partitions are chosen as follows:
Resulting in half of the partitions on each topic not receiving any messages. Instead I'd expect the following behavior:
Kafka Version
github.com/segmentio/kafka-go v0.4.47
Kafka version is unknown.
To Reproduce
A minimal working example:
Expected Behavior
I'd expect for both topics 1 message on partition 0 and 1.
Observed Behavior
There are 2 messages on topic 1, partition 0 and 2 messages on topic 2 partition 1.
Additional Context
The text was updated successfully, but these errors were encountered: