Hi Devs,
Following the release of v3.3 (for Flink 1.20.x) and v4.0 (for Flink
2.0.x), the Flink JDBC Connector has not had a new release for quite some
time. During this period, Flink itself has released versions 2.1, 2.2, and
2.3 (upcoming).
I would like to initiate a discussion regarding the next group of Flink
JDBC Connector releases.
0. I’d like to hear more thoughts on the next release
Based on a quick review of the current repository status, I have identified
several topics that need to be discussed before the release plan:
1. Versioning Strategy for Flink 2.1, 2.2, and 2.3
Do we need a new branching/versioning scheme to maintain patches and
updates for these specific Flink versions?
Option A: Semantic versioning tied to Flink (e.g., v2.1-1.0)
Format: {Flink_Version}-{Connector_Version}(e.g., 2.1-1.0).
Pros: It is immediately clear which Flink version the connector supports.
Cons: This deviates from our historical naming convention and might cause
confusion for users and developers.
Option B: Continuing sequential versioning (e.g., v5.x, v6.x)
Mapping: v5.xfor Flink 2.1, v6.xfor Flink 2.2, etc.
Pros: Consistent with the previous naming scheme (v3.x, v4.x).
Cons: It is less intuitive to map connector versions to Flink versions.
Note on LTS: Regardless of the outcome for the 2.x series, continuing to
use the v3.xseries for the long-term support (LTS) Flink 1.20.x line may
remain a solid choice.
2. Support Policy for Flink 2.0
Do we still need to support the JDBC connector for Flink 2.0? Since the
latest Flink version is 2.3 (upcoming), Flink 2.0 is now three minor
versions behind.
BTW, I'd be willing to volunteer as a Release Manager (RM) for the next
release of Flink JDBC connector.
I also look forward to collaborating with other interested contributors to
drive this effort forward.
Thank you very much. And looking forward to your feedback.
Best regards,
Yuepeng Pan
