I have used Titan graph database with Cassandra and Solr in production environment. Currently our application is deployed and minor down-time is acceptable. Unfortunately due to some changes in schema we need to re-index the whole Titan graph instance. What is the best way of re-indexing process with considering these concerns:
Minor down-time would be acceptable.
Classic Solr schema is used and due to space efficiency all fields were considered with
store="false". (except the ID)We have some hard-coded id indexing in Solr. (for making some of the aggregation queries more efficiently by indexing some of the vertex-IDs in Solr.) Therefore, the cost of solutions that might change vertex-id is high.
