I am using 3.5 . ----- Original Message ----- From: Lance Norskog [mailto:goks...@gmail.com] Sent: Monday, May 14, 2012 11:08 AM To: solr-user@lucene.apache.org <solr-user@lucene.apache.org> Subject: Re: Index Corruption
"Index corruption" usually means data structure problems. There is a Lucene program 'org.apache.lucene.index.CheckIndex' in the lucene core jar. If there is a problem with the data structures, this program will find it: java -cp lucene-core-XX.jar org.apache.lucene.index.CheckIndex ..../index/data Do you use Solr 3.1, 3.2 or 3.3? There was an index flushing bug in this series of Solr releases. Solr 3.4, 3.5 and 3.6 don't have the problem, and the trunk never had the problem. You should not (to my knowledge) ever have duplicated documents if there is a crash while indexing. If this happens, but there is no Lucene index corruption, please file a bug. Lance On Sun, May 13, 2012 at 1:51 PM, shubham <shubha...@gmail.com> wrote: > We have a problem in last couple of days when a particular Solr master was > restarted while there was an import running . This led to the corruption of > some document entities where they had multiple doc's of same unique id etc. > > Is this kind of corruption possible , by now I expected that Solr indexing > works a way where either the data is completely imported/updated or nothing > has changed, But with this there exists a third possibility which is pretty > risky. Apart from writing queries to generate alerts when some kind of > corruption occurs is there a recommended way to do the same. However why the > same happened in terms of corruption still bother's me. > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Index-Corruption-tp3983579.html > Sent from the Solr - User mailing list archive at Nabble.com. -- Lance Norskog goks...@gmail.com
