leaves12138 opened a new pull request, #7778:
URL: https://github.com/apache/paimon/pull/7778

   ### What changed
   
   - Parallelize manifest reads during manifest full compaction using the 
existing `ManifestReadThreadPool` and `scan.manifest.parallelism`.
   - Keep compaction output ordering stable by reading in parallel batches and 
writing results sequentially.
   - Add a regression test that blocks manifest reads until two readers enter 
concurrently.
   
   ### Why
   
   Commit full compaction previously read each manifest file serially in the 
second read pass. On object storage or Jindo-backed file systems, one slow 
manifest read could hold the global committer for a long time. This change lets 
the existing manifest read parallelism speed up that IO-bound phase.
   
   ### Validation
   
   - `mvn -pl paimon-core -Pfast-build -Dtest=ManifestFileMetaTest test`
   - `mvn -pl paimon-core -DskipTests compile`


-- 
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]

Reply via email to