On Sat, Apr 11, 2009 at 7:26 PM, Chris Masters <roti...@yahoo.com> wrote:
> > I'm evaluating Solr to provide a single fast and effective searching > solution for a range of web applications (Java, .Net & PHP). That's good news for Solr :) > So a couple questions: > > 1) Can we configure Solr to auto reindex/import if we point it to a > datasource? I've done some searching on this and have only come up with > specific mailing list questions on DataImporter - there seems to be no docs > on that. If so, can we use this mechnism to point to multiple datasources > for different indexes i.e. use datasource A for index A and datasource B for > index B? There's a comprehensive wiki document at http://wiki.apache.org/solr/DataImportHandler > 2) Can we create multiple indexes of the same objects i.e. index of users > for customer A and another for users of customer B? If so, how would be > specify context/index to use? Looks like you need http://wiki.apache.org/solr/MultipleIndexes > 3) It seems that the built-in replication depends on rsync; we use a SAN > for our production databases, can we easily use the SAN for > replication/redundancy? Are there fiel system locking issues (as with > out-of-the-box lucene) with two Solr servers using the same index shared via > a SAN? The issues would be same as for Lucene. I've never used Solr with SAN so I do not know about if it will work or how well it will perform. > 4) We've extended one of our local lucene search implementations to store > the index in a database (not for speed but for redundancy & load balancing) > using a JDBCDirectoryProvider implementation - is it possible to use that > with Solr? Again, never tried. Though, support for configurable directory provider was added recently (in trunk). See https://issues.apache.org/jira/browse/SOLR-465 for more details. -- Regards, Shalin Shekhar Mangar.
