[FLINK-25686] Support Pulsar Schema evolution in Pulsar Source Connector

• Type: New Feature
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: pulsar-4.0.0
• Component/s: Connectors / Pulsar
• Labels: Pulsar pull-request-available stale-assigned
• Assignee: Yufan Sheng
• Reporter: Yufei Zhang
• Created: 18/Jan/22 06:37
• Updated: 08/Jan/23 08:46
• Resolved: 08/Jan/23 08:46

Description

Currently the source connector reads all messages as Message<byte[]>, and implements a deserialization abstraction to deserialize the byte array message. This deserialization abstraction supports both flink schema and pulsar schema. However because of this extra layer of abstraction, even when using pulsar schema, the schema info is not provided to the consumer (client), so it is not using any pulsar’s schema validation mechanism. (If pulsar client has a schema, it will send the schema to broker on connect and broker will validate if it is a valid schema according to a preset compatibility rule :[ https://pulsar.apache.org/docs/en/schema-evolution-compatibility/|https://pulsar.apache.org/docs/en/schema-evolution-compatibility/]).  we need to support this schema evolution

Issue Links

• Blocked: FLINK-20732 Implement Pulsar Sink on extend Sink API (FLIP-191)
• is blocked by: FLINK-28352 [Umbrella] Make Pulsar connector stable
• links to: GitHub Pull Request #12
• links to: GitHub Pull Request #21207