Navneet Verma

@reta I did some deep-dive and checked with CCR team(@ankitkala) and got to know recovery_source is not getting used in CCR flow. One flow which I am still confirming where...

@reta So I did some more deep-dive and validated few scenarios where whole source and recovery source is removed(with doc replication and seg replication). 1. The recovery source/source is only...

> > In my opinion removing recovery source helps and there are implementation of Opensearch where segments are stored remotely and don't rely on Lucene docs for peer recovery. This...

> > Its upto the service and specific implementation to deicide. If as a user after every bulk request call I am calling flush then there will be no issues...

> > I don't see how this is different. > > @navneet1v `_source` was and is user visible (mapping, search, etc), `_recovery_source` was never. It is not directly visible but...

> > It is not directly visible but in the docs it is mentioned. > > @navneet1v you asked the difference, I explained (I didn't know about `_recovery_source` fe), could...

@msfroh, @reta I have done some checking. Addition is recovery source is mentioned at few places and issues. Let me try to post them. 1. Main PR which added recovery...

@reta and @msfroh I have updated the code with comments and expert flag. Will do the same for Opensearch documentation.

> ❌ Gradle check result for [3cdba45](https://github.com/opensearch-project/OpenSearch/commit/3cdba4593eb8edb6bddb09d352a5bfa243407b0e): [FAILURE](https://build.ci.opensearch.org/job/gradle-check/39507/) > > Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure [a flaky...

@jackiehanyang do we know the impact on JVM and latency for using a JSON to JSON transform(using painless and Json path) on a search response(taking reference of search response as...