msokolov commented on issue #14431:
URL: https://github.com/apache/lucene/issues/14431#issuecomment-2781449497
We're operating in a setup where we have an initial phase that builds an
index while it is offline, not accepting query traffic. Once that is complete,
we enable the index to take
jpountz commented on issue #14431:
URL: https://github.com/apache/lucene/issues/14431#issuecomment-2780641860
> and some deletes being addressed is better than none.
This part of your message suggests that deletes get reclaimed progressively
over time, which is often not true. So wait
jpountz commented on issue #14431:
URL: https://github.com/apache/lucene/issues/14431#issuecomment-2778108878
For my understanding, what is the benefit of waiting until the timeout is
reached rather than not waiting at all?
--
This is an automated message from the Apache Git Service.
To r
houserjohn commented on issue #14431:
URL: https://github.com/apache/lucene/issues/14431#issuecomment-2779350159
Apologies if I am misunderstanding your question, but the example that it is
great for is right after the full indexing of your documents. The indexing
likely created many delete
houserjohn opened a new issue, #14431:
URL: https://github.com/apache/lucene/issues/14431
### Description
Using IndexWriter's `forceMergeDeletes` to eliminate merge debt is a very
useful feature -- especially during initial indexing. However, larger indexes
can require 20+ minutes to
