SEZ9 commented on issue #12983: URL: https://github.com/apache/seatunnel/issues/12983#issuecomment-6623484389
Thanks for the update, @goutamadwant — good to see this moving. A few points on the collector scope tracked here (`FlinkRowCollector` / `InternalRowCollector` `restoreSchema`): - Flink: you mention the restore checks passed locally. When the collector PR is up, please include the actual validation in it: a CDC schema-change checkpoint/savepoint restore through `FlinkRowCollector`, asserting columns, types and values of the *first* emitted row, plus the schema-change-disabled and legacy-checkpoint cases, and a check that refreshing local schema state does not replay downstream DDL. - Spark: thanks for reproducing the converter mismatch with restored schema state — that confirms the gap on `InternalRowCollector`. Since you note full Spark CDC restore depends on #13107 (and recovery beyond #12776), please keep that dependency explicit in the PR description and make clear which part of the Spark path is validated now versus blocked on that work. - Could you clarify how #13109 and #13110 relate to this issue — are they prerequisites only, or do they already carry any of the collector `restoreSchema` changes? If the collector changes land separately (as you indicate), a Flink PR and a Spark PR are fine if their state ownership differs; just state which adapter and recovery boundary each one actually validates. - Please link the collector PR(s) back here and report the relevant integration-suite result on the final head. Keep the Zeta-side evidence in #11932 separate from the Flink/Spark adapter evidence. Remaining asks: (1) collector PR link(s) with the restore test coverage above, (2) explicit dependency notes on #11932 and #13107, (3) integration-suite results on the final head for each PR. <!-- streview-comment:2276 --> -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
