Alexander, I'd stay away from NFS (slow). It sounds like you'd benefit from moving from vanilla Lucene to Solr that your app(s) could query from different boxes.
Otis -- Sematext -- http://sematext.com/ -- Lucene - Solr - Nutch ----- Original Message ---- From: Alexander Wallace <[EMAIL PROTECTED]> To: java-user@lucene.apache.org Sent: Wednesday, October 17, 2007 5:56:43 PM Subject: lucene index in a cluster. Hi All! We use lucene 2.0.0 as the search index of our portal (liferay based) and are trying to use it correctly in a cluster... I'm very foggy on wether i should share the index file using NFS (the servers run linux) or wether we should try the jdbc store... We currently have the index in an NFS mounted volume, however i'm pretty sure that the servers are not sharing locks, and don't know that the other node exists... Both servers read and write to the index freely currently. Can you give me advise as to what's best in this setup (share nfs vs jdbc) and how to make the nodes in the cluster know about each other so that they don't step on each others feet? Pointers to documentation are very welcome too! Thanks a lot in advance for your response! --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
