kumarpritam863 opened a new pull request, #18858: URL: https://github.com/apache/iceberg/pull/18858
`TableUpdater` failed when someone else committed to the table between loading it and committing a schema or spec update: after a `CommitFailedException` the spec path evolved against a missing spec and threw a `NullPointerException`, and the schema path rethrew. Both now check the refreshed table and retry up to three times before rethrowing. A table that another writer created while this one was creating it is loaded instead of running the creation again, which also removes an unbounded recursion; if it cannot be loaded, for example because a view has the same name, the `AlreadyExistsException` is reported. This is Flink 2.3 only; backports follow. Test plan: `TestTableUpdater` has a test for the spec update, the schema update, a table created by another writer and a table that cannot be loaded. Without the change the first, second and last fail (`NullPointerException`, `CommitFailedException`, `StackOverflowError`). The `sink.dynamic` tests pass. --- **AI Disclosure** - Model: Claude Sonnet 5.5 - Platform/Tool: Claude Code -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
