Hi there, I've got a pretty simple question regarding the DIH full-import command. I have a SOLR server running that has a full index with lots of documents in it. Once a day, a full-import is run, which uses the default parameters (clean=true, because it's not an incremental index). When I run a full-import, the first step is cleaning up the whole index:
Feb 7, 2009 2:12:01 AM org.apache.solr.update.DirectUpdateHandler2 deleteAll INFO: [] REMOVING ALL DOCUMENTS FROM INDEX After that, suppose the import suddenly fails for one reason or another (ie. SQL error), which initiates a rollback: Feb 7, 2009 2:12:02 AM org.apache.solr.handler.dataimport.DataImporter doFullImport SEVERE: Full Import failed [...] Feb 7, 2009 2:12:02 AM org.apache.solr.update.DirectUpdateHandler2 rollback INFO: start rollback Feb 7, 2009 2:12:02 AM org.apache.solr.update.DirectUpdateHandler2 rollback INFO: end_rollback Feb 7, 2009 2:12:02 AM org.apache.solr.update.DirectUpdateHandler2 commit INFO: start commit(optimize=false,waitFlush=false,waitSearcher=true) Unfortunately, this rollback does not "refill" the index with the old data, and neither keeps the old index from being overwritten with the new, erroneous index. Now my question is: is there anything I can do to keep Solr from trashing my index on a full-import when there is a problem with the database? Or should I use clean=false, even though 99% of the imported documents are not incremental but the same documents that already were in the index, only with new data? Any tips will be greatly appreciated! :) - Steffen -- View this message in context: http://www.nabble.com/DIH-full-import-with-clean%3Dtrue-fails-and-rollback-empties-index-tp22055065p22055065.html Sent from the Solr - User mailing list archive at Nabble.com.
