Search before asking

• I searched in the issues and found nothing similar.

Motivation

With #640, we synchronously create a datalake table in Paimon/Iceberg when creating a lake enabled table in Fluss. It is implemented by 3-steps:

1. Sanity check for the table to create
2. Create the datalake table in Paimon/Iceberg
3. Create the table in Fluss

However, step#3 may fail if the table/db exists or for some reasons else and can result in in-consistent metadata between Paimon and Fluss. It's hard to support distributed transactions, but it's easy to support eventual consistency by idempotent writing. Therefore, we should tolerate if setp#2 the lake table exist but matches our schema and properties instead of throwing exceptions. Thus, we can have an eventual consistency for the metadata.

Solution

No response

Anything else?

No response

Willingness to contribute

• I'm willing to submit a PR!