[ https://issues.jboss.org/browse/ISPN-3795?page=com.atlassian.jira.plugin.... ] Sanne Grinovero commented on ISPN-3795: --------------------------------------- {quote}[..] After that we add transparent batching to Hibernate Search SYNC backend, plus the async indexing backend, I reckon performance is vastly superior to do batch loading nowadays.{quote} Yes that is a solved problem, but I wonder if we'd still be fetching values from somewhere, or attempting to fetch from somewhere - even if there's no data - and block on such rpcs such as loading from a remote node or from a DB. Granted a "batch loading" is supposedly happening on an _empty_ data grid so there shouldn't be much of a payload to fetch for existing keys, but e.g. I remember before introducing SKIP_CACHESTORE even just looking into a cachestore for each value was a performance killer. In short I'm not worried about the indexing performance but the other costs which are typically a good reason to enable _IGNORE_RETURN_VALUE_ . I guess I might be too paranoid: if I'm wrong and the core team sees no such costs anymore then I'm happy to simply remove this. Let's just not forget to perf test for this. -- This message was sent by Atlassian JIRA (v7.2.3#72005)