Remove sort by primary term when reading soft-deletes (elastic#43845)

With Lucene rollback (elastic#33473), we should never have more than one
primary term for each sequence number. Therefore we don't have to sort
by the primary term when reading soft-deletes.

Author: dnhatn

server/src/main/java/org/elasticsearch/index/engine/LuceneChangesSnapshot.java

@@ -211,11 +211,8 @@ private void fillParallelArray(ScoreDoc[] scoreDocs, ParallelArray parallelArray
 
     private TopDocs searchOperations(ScoreDoc after) throws IOException {
         final Query rangeQuery = LongPoint.newRangeQuery(SeqNoFieldMapper.NAME, Math.max(fromSeqNo, lastSeenSeqNo), toSeqNo);
-        final Sort sortedBySeqNoThenByTerm = new Sort(
-            new SortField(SeqNoFieldMapper.NAME, SortField.Type.LONG),
-            new SortField(SeqNoFieldMapper.PRIMARY_TERM_NAME, SortField.Type.LONG, true)
-        );
-        return indexSearcher.searchAfter(after, rangeQuery, searchBatchSize, sortedBySeqNoThenByTerm);
+        final Sort sortedBySeqNo = new Sort(new SortField(SeqNoFieldMapper.NAME, SortField.Type.LONG));
+        return indexSearcher.searchAfter(after, rangeQuery, searchBatchSize, sortedBySeqNo);
     }
 
     private Translog.Operation readDocAsOp(int docIndex) throws IOException {