mroonga: prefer deferring full text on the same node for better cache efficiency

In this specific case, the number of grouped messages is bound,
and actually not greater than the number of grouped messages
being processed for normal indexing.

bt5/erp5_full_text_mroonga_catalog/CatalogMethodTemplateItem/portal_catalog/erp5_mysql_innodb/SQLCatalog_deferFullTextIndex.py

@@ -6,6 +6,7 @@ for document, root_document_path in zip(getPath, getRootDocumentPath):
     document,
     activity='SQLQueue',
     priority=4,
+    node='same',
     group_method_id=GROUP_METHOD_ID,
     serialization_tag='full_text_' + root_document_path,
   ).SQLCatalog_deferFullTextIndexActivity()