I mean this row https://github.com/apache/lucene-solr/blame/master/solr/core/src/java/org/apache/solr/handler/IndexFetcher.java#L458
Sure, "New index in Master. Deleting mine..." makes sense. But committing right after that is not really determined. Eg if this empty slave commit is delayed it's assigned with late timestamp, and ignores master commit as old. Don't you think it's a problem (at least for test stability)? The possible solutions are - pass earler timestamp to commit explicitly, I'm not sure if it's ever possible; - or always set skipCommitOnMasterVersionZero=true or by default at least. More details at https://issues.apache.org/jira/browse/SOLR-11673 -- Sincerely yours Mikhail Khludnev
