Running multiple RDF4J servers against same Halyard/HBase

[peter.l...@inova8.com]

Maybe an off-the-wall question, but is it possible to run multiple RDF4J servers against the same Halyard HBase in a read only mode.

My question is motivated to see if, in the situation where there are multiple concurrent users issuing SPARQL queries (not updates or inserts), different users could be serviced by different RDF4J servers and hence improve concurrent SPARQL query performance.

Thanks

[Adam Sotona]

Yes, that is the primary benefit of the architecture. Just don't forget to watch the hbase region servers logs for the signs of overload and be ready to tweak the hbase in dependency of the load. It is even possible to use multiple servers in write mode, there is just no transactions locking so the results might be non-deterministic. However for mild parallel additive write load it is pretty safe. Heavy parallel write load would probably kill the hbase soon, so bulk-loads are then recommended.

Adam

12. 7. 2020 v 12:48, peter.l...@inova8.com <peter.l...@inova8.com>:

Maybe an off-the-wall question, but is it possible to run multiple RDF4J servers against the same Halyard HBase in a read only mode.

My question is motivated to see if, in the situation where there are multiple concurrent users issuing SPARQL queries (not updates or inserts), different users could be serviced by different RDF4J servers and hence improve concurrent SPARQL query performance.

Thanks

--
You received this message because you are subscribed to the Google Groups "Halyard Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to halyard-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/halyard-users/fe7aa392-2d32-44de-b6c2-e1a5adcb54e6n%40googlegroups.com.