Near Realtime Search Replication

• Key: SOLR-1278
• Type: New Feature
• Status: Open
• Priority: Minor
• Assignee: Unassigned
• Reporter: Jason Rutherglen
• Component/s: replication (java), search, update
• Affects Version/s: 1.4
• Fix Version/s: 1.5

Description

Lucene 2.9 and later offer near realtime search (LUCENE-1516,
LUCENE-1313). For SOLR this means integrating
IndexWriter.getReader and adding a way to replicate newly
created segments that may not exist on the file system to other
SOLR servers in an efficient way.

I don't think replicating documents as is would be optimal as it
requires re-analyzing on the slaves which we should seek to
avoid.

Issues:

• Replicate using the existing Java based non-script system that
  uses HTTP or create a protocol that uses sockets?

• Lucene needs a more efficient way of adding these segments
  (LUCENE-1738)