StevenLuMT commented on PR #3103: URL: https://github.com/apache/bookkeeper/pull/3103#issuecomment-1164039070
> I still cannot see a case where flushing the entry location index (ledgerId, entryId, offset) becomes the bottleneck, compared to: > > 1. Journal > 2. Entry log flush (where all the data has to be put on disk) > > If it is not the bottleneck, then why reduce the guarantees? (eg: if we miss the async update, we have effectively lost data in the bookie). yes, in the step : entrylog flush In order to speed up the flush, we only need to ensure that the EntryLocationIndex is successfully written under normal circumstances, regardless of machine downtime. Like LedgerMetadataIndex, it also uses an asynchronous method to write rocksdb @merlimat -- 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]
