rangareddy commented on issue #10587: URL: https://github.com/apache/hudi/issues/10587#issuecomment-4842794633
Two separate issues here. (1) The `Config conflict: hoodie.database.name` error is config immutability — your table was created by an older version without a database name, and `hoodie.database.name` is an immutable table property that can't be added/changed on later writes (note the blank "existing value"). Fix: remove `hoodie.database.name` from the write options for this table. (2) The upsert hang is because you're switching `hoodie.index.type` to `RECORD_INDEX` on a table that wasn't created with RLI — the `record_index` MDT partition doesn't exist, so the first upsert tries to bootstrap it over the whole dataset and stalls on 0.14.0 (the driver loops on broadcasts while executors idle). You can't flip to RLI in place: either rebuild a fresh table with RLI from the first write (bulk_insert → upsert), or build the partition first with `HoodieIndexer --index-types RECORD_INDEX`. Also upgrade off 0.14.0 (0.14.1 fixed several RLI/MDT bugs; 0.15.0/1.x recommended), and validate with a small quickstart bulk_insert→upsert before scaling. -- 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]
