Virtuoso RDF Replication of Public Graphs Only

Virtuoso Universal Server Tips, Tricks, and HowTos
, , ,
1

What

Virtuoso RDF Graph Replication can be configured to only publish Graphs that are not part of the special Virtuoso private graph group, essentially enabling selective publishing and un-publishing of graphs to subscriber nodes.

Why

In some cases users do not what to publish all graphs in the RDF Quad Store with the RDF_RDF_REPL_GRAPH_INS() function special http://www.openlinksw.com/schemas/virtrdf#rdf_repl_all graph IRI name or may want to have the ability to selectively publish and un-publish graphs to subscriber nodes.

How

This can be done using the RDF_RDF_REPL_GRAPH_INS() functions special http://www.openlinksw.com/schemas/virtrdf#rdf_repl_world graph IRI name, run on the MASTER Publisher, which instructs the RDF Graph replication process to check the special graph group http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs for the list of graphs that should not be published ie be treated as private/embargoed.

If a graph in the http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs graph group needs to be published, simply delete is from the graph group. Likewise should a graph published need to be made private simply add it to the http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs graph group and it will automatically be deleted from the subscriber nodes.

The steps to setup are:

  1. Run the DB.DBA.RDF_REPL_GRAPH_INS ('http://www.openlinksw.com/schemas/virtrdf#rdf_repl_world') command to enable publishing of only those graph not in the http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs graph group.
  2. Create the http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs graph group with the command DB.DBA.RDF_GRAPH_GROUP_CREATE ('http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs',1);
  3. Add graph not to be published to the http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs graph group with the command `DB.DBA.RDF_GRAPH_GROUP_INS (‘http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs’, ‘graph-not-to-publish’);

Example

  1. Setup Graph Replication on a MASTER Publisher to replicate public graphs ONLY ie ‘http://www.openlinksw.com/schemas/virtrdf#rdf_repl_world’
DB.DBA.RDF_REPL_START();
DB.DBA.RDF_REPL_GRAPH_INS ('http://www.openlinksw.com/schemas/virtrdf#rdf_repl_world');
  1. Setup a Graph Replication SLAVE Subscriber (assuming publisher Replication server name “MASTER” and publisher ODBC DSN called “MASTER_DSN”):
repl_server ('MASTER', 'MASTER_DSN');
repl_subscribe ('MASTER', '__rdf_repl', 'dav', 'dav', 'dba', 'dba'); 
repl_sync_all ();
DB.DBA.SUB_SCHEDULE ('MASTER', '__rdf_repl', 1);
  1. Insert test data on the MASTER Publisher:
SPARQL INSERT INTO GRAPH <http://example.org> { <1> <2> <3> };
SPARQL SELECT * FROM <http://example.org> WHERE {?s ?p ?o};
repl_stat();
  1. Check on the SLAVE subscriber that data in http://example.org graph has been replicated:
SPARQL SELECT * FROM <http://example.org> WHERE {?s ?p ?o};
repl_stat();
  1. Add http://example.org graph to Private graph group on the MASTER Publisher:
DB.DBA.RDF_GRAPH_GROUP_CREATE ('http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs',1);
DB.DBA.RDF_GRAPH_GROUP_INS ('http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs', 'http://example.org');
SPARQL SELECT * FROM <http://example.org> WHERE {?s ?p ?o};
repl_stat();
  1. Check on the SLAVE subscriber that the data in the now private http://example.org graph has been removed:
SPARQL SELECT * FROM <http://example.org> WHERE {?s ?p ?o};
repl_stat();
  1. Make the private http://example.org graph public again on the MASTER Publisher:
SELECT id_to_iri (RGGM_GROUP_IID), id_to_iri(RGGM_MEMBER_IID) FROM DB.DBA.RDF_GRAPH_GROUP_MEMBER;
DB.DBA.RDF_GRAPH_GROUP_DEL ('http://www.openlinksw.com/schemas/virtrdf#PrivateGraphs', 'http://example.org');
SELECT id_to_iri (RGGM_GROUP_IID), id_to_iri(RGGM_MEMBER_IID) FROM DB.DBA.RDF_GRAPH_GROUP_MEMBER;
SPARQL SELECT * FROM <http://example.org> WHERE {?s ?p ?o};
repl_stat();
  1. Check on the SLAVE subscriber that the now public http://example.org graph’s data has been replicated to it:
SPARQL SELECT * FROM <http://example.org> WHERE {?s ?p ?o};
repl_stat();

Related