0

We're trying to replace RDBMS (Oracle, Teradata) in our web application. Reasons are complicated selects with many joins, which also affect performance and option to embed database to our distribution package. Application contains basically many table data with complicated structure, dynamic relations and even dynamic inheritance. We need primary read performance over write. It is written in Java and we want to preserve multi-platformness so we want to stick with Java databases.

We're currently trying Solr for it's awesome search performance and because it's Java standalone application so we can embed it to our distribution or even embed it directly to our app.

My question is whether it is a good idea to use Solr as primary storage? What are the risks of using it in this way? Generally it's not recommended, but why? What is the blocker? Because we haven't find any yet.

Michal Vala
  • 41
  • 4
  • 2
    Look here: http://stackoverflow.com/questions/4258593/using-solr-search-index-as-a-database-is-this-wrong – ItayD May 23 '16 at 08:52
  • To expand on the duplicate question as well: Often the answer to updating structures in Solr is to reindex (for example to process content differently or index it differently). If Solr is your primary datastore, that gets more tricky. Single field updates are a "recent" feature - Solr was never intended as a key/value store to begin with. That being said - there's no problem using it as the _primary_ datastore for all operations, but I'd be wary of using it as the master data source. – MatsLindh May 23 '16 at 10:48

0 Answers0