Hi, David. +1 to cut main into a new 1.20 compatible release as flink-connector-jdbc for 1.20 is a demand frequently mentioned by users and is also prioritized for adaptation to Flink 2.0.
I’ve checked the issues[1] that were solved in 3.3, most of them are introduced by FLIP-377: Support fine-grained configuration to control filter push down for Table/SQL Sources and FLIP-449: Reorganization of flink-connector-jdbc, and only two bugfixs[4] that related jdbc-3.2.0 product codes, which I don't think is a blocking that must create a release. [1]https://issues.apache.org/jira/browse/FLINK-33460?jql=project%20%3D%20FLINK%20AND%20status%20%3D%20Resolved%20AND%20fixVersion%20%3D%20jdbc-3.3.0 [2]https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=276105768 [3]https://cwiki.apache.org/confluence/display/FLINK/FLIP-449%3A+Reorganization+of+flink-connector-jdbc [4]https://issues.apache.org/jira/browse/FLINK-35542?jql=project%20%3D%20FLINK%20AND%20issuetype%20%3D%20Bug%20AND%20status%20%3D%20Resolved%20AND%20fixVersion%20%3D%20jdbc-3.3.0 > 2024年11月22日 01:15，David Radley <david_rad...@uk.ibm.com> 写道： > > Hi, > Is there a plan to release the JDBC for Flink 1.20? Apologies if this is > already in hand, I could not find anything: > > * I notice that the last JDBC connector content corresponding to Flink > 1.19 contained minimal content. > * The main branch – now has a pom with 1.20 and contains a massive amount > of content compared to 1.19, which has minimal cherry picked fixes in. > * So the worry is that we have a large amount of unshipped code in main, > amplified by the fact that there is introduces support for many new databases. > * The latest code appears to have open lineage support with > https://github.com/apache/flink-connector-jdbc/pull/137 so I assume the main > code is now not compatible with Flink 1.19 > > > Thinking about how to move this forward. Some options: > > 1. I am wondering is there an appetite to cut main into a new 1.20 > compatible release pretty much as-is > 2. We could do a 1.19 and 1.20 compatible release like the Kafka connector > 3.3 and a 1.20 open lineage release like the Kafka connector 3.4, with > minimal critical backported fixes. > > I am keen to do option 1), as the main code will otherwise continue to > diverse from what we release. I wonder what manual per database testing would > be required; to reassure us all the new database support actually works. Or > is the existing automated testing enough? > > If we get some consensus on approach, I can help drive this, as this is going > to be an inhibitor for many of us in the community to migrate to Flink 1.20 > and subsequently to Flink v2. > > Kind regards, David. > > Unless otherwise stated above: > > IBM United Kingdom Limited > Registered in England and Wales with number 741598 > Registered office: Building C, IBM Hursley Office, Hursley Park Road, > Winchester, Hampshire SO21 2JN
