msokolov commented on issue #580: LUCENE-8700: IndexWriter.yield() URL: https://github.com/apache/lucene-solr/pull/580#issuecomment-466045920 HI @s1monw there was some context in the mailing list, sorry I did not copy here. The use case here is batch indexing followed by a commit. When one has lots of RAM and no need to commit frequently for visibility (eg offline testing, or some nightly background process), it may be more performant to index without committing, accumulating fairly large in-memory buffers before flushing, and then at the end of a large batch, commit(). The problem with this case today is that we do not make use of any parallel I/O when flushing unless calls are being made to `updateDocument()` and the like. I considered calling that with an empty document, but I was concerned that it might have some undesirable side effects?
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
