Hi, I have a very strange issue where I urgently need some help. I have developed a nutch plugin (own implementations of IndexingFilter and HtmlParseFilter) using Windows (+Cygwin) and everything works fine. Now I need to deploy it to a Linux machine having the same nutch, solr and java versions (1.2, 3.1.0 and Java 6). My plugin is registered correctly on both systems as seen by the following output in hadoop.log: 2011-07-22 14:13:53,853 INFO plugin.PluginRepository - Plugin Auto-activation mode: [true] 2011-07-22 14:13:53,853 INFO plugin.PluginRepository - Registered Plugins: 2011-07-22 14:13:53,853 INFO plugin.PluginRepository - myPlugin Parser/Filter (myplugin)
On Windows myParser (implementation of HtmlParseFilter) and myIndexer (implementation of IndexingFilter) classes are both executed during their stages. On Linux however only myIndexer is called, which makes no sense as myParser first needs to extract metadata used by myIndexer. There seems to be nothing relevant in the logs. The jar of my plugin is the same on both systems. Where should I start to search for the problem? On both systems the plugin is loaded but on Linux only part of my Plugin implementation (only the indexer) is called. Has anyone already experienced such a weird behaviour? I would appreciate a reply as I need to work out this issue in the next few hours. Thanks. -- View this message in context: http://lucene.472066.n3.nabble.com/Nutch-plugin-ignored-in-linux-works-on-windows-tp3191884p3191884.html Sent from the Nutch - User mailing list archive at Nabble.com.
