I'm not quite done indexing yet, but I find that if I set both of these parameters to -1, I don't get an error (for now). It's a bit slower, because it closes the reader for every item, but it seems to be working.
-Jose On Tue, Jun 24, 2014 at 1:07 PM, Jose Blanco <[email protected]> wrote: > Perhaps my problem has to do with these two parameters in DSInderer.java: > > private static int indexFlushDelay = > ConfigurationManager.getIntProperty("search.index.delay", -1); > > private static int batchFlushAfterDocuments = > ConfigurationManager.getIntProperty("search.batch.documents", 20); > > I can't find them in dspace.cfg for version 4. Not sure what I should > set them to, but will experiment. > > Thank you! > Jose > > On Tue, Jun 24, 2014 at 10:44 AM, Jose Blanco <[email protected]> wrote: >> TIm, Presently in version 3 I have both discovery and lucene working >> together. I use lucene from this page where the user can enter a >> search term and press "Go" >> >> http://deepblue.lib.umich.edu/ >> >> I also use it in this page: >> >> http://deepblue.lib.umich.edu/advanced-search >> >> I'm also using SRW and it uses the lucene API to do it's searches.. >> >> I would like to do the same with version 4.1, especially because I >> want to keep the check book to only search peer reviewed documents >> from the Advanced search page, and to do this with discovery would be >> a big change. I don't want to add a filter. I just don't think this >> would be a good user experience with the filter. >> >> I changed launcher.xml to only index lucene: >> >> <command> >> <name>index-lucene-init</name> >> <description>Initialise the Lucene search and browse >> indexes</description> >> >> <step passuserargs="false"> >> <class>org.dspace.search.DSIndexer</class> >> <argument>-b</argument> >> </step> >> </command> >> >> and then I ran >> >> ./dspace index-lucene-init >> >> And everything was going just fine, and then I start getting Too many >> open files. >> >> 2014-06-24 10:29:29,971 INFO org.dspace.search.DSIndexer @ Wrote >> Item: TEMP-BOGUS/229576 to Index >> 2014-06-24 10:29:30,033 ERROR org.dspace.search.DSIndexer @ >> java.io.FileNotFoundException: >> /repository/dev/search/_5nto_Lucene41_0.tim (Too many open files) >> 2014-06-24 10:29:30,133 ERROR org.dspace.search.DSIndexer @ >> java.io.FileNotFoundException: >> /repository/dev/search/_5nto_Lucene41_0.tim (Too many open files) >> 2014-06-24 10:29:30,222 ERROR org.dspace.search.DSIndexer @ >> java.io.FileNotFoundException: /search/_5nto_Lucene41_0.tim (Too many >> open files) >> 2014-06-24 10:29:30,291 ERROR org.dspace.search.DSIndexer @ >> java.io.FileNotFoundException: >> /repository/dev/search/_5nto_Lucene41_0.tim (Too many open files) >> 2014-06-24 10:29:30,359 ERROR org.dspace.search.DSIndexer @ >> java.io.FileNotFoundException: >> /repository/dev/search/_5nto_Lucene41_0.tim (Too many open files) >> >> Any suggestions? >> >> Thank you! >> Jose >> >> On Thu, May 22, 2014 at 4:58 PM, Tim Donohue <[email protected]> wrote: >>> Hi Jose, >>> >>> >>> On 5/22/2014 1:37 PM, Jose Blanco wrote: >>>> >>>> I have a few questions, that I think are going to be simple. >>>> >>>> (1) I was looking at the cron jobs needed for 4.1 and notice that we now >>>> use: >>>> >>>> $DSPACE/bin/dspace index-discovery >>>> >>>> I've been running the following in 3 >>>> >>>> $DSPACE/bin/dspace update-discovery-index >>>> >>>> So I suppose index-discovery is taking the place of >>>> update-discovery-index? >>> >>> >>> That's correct, "update-discovery-index" was renamed to "index-discovery" in >>> DSpace 4.x >>> >>> >>>> >>>> (2) In previous upgrades, I've run >>>> >>>> ./dspace index-init >>>> >>>> But I see that is no longer in the documentation. I'm going to >>>> continue to use ArtifactBrowse because I have some customizations in >>>> there that I would like to keep. So would I need to run dspace >>>> index-init? >>> >>> >>> By default, in DSpace 4, everything runs through Discovery (Search & >>> Browse). So, by default, the old browse indexes (in the database) and the >>> old search indexes (in Lucene) are now disabled entirely. The "index-init" >>> script used to control those indexes, and it has been split into two >>> scripts: >>> * index-db-browse - to update DB browse indexes (if enabled) >>> * index-lucene-init - to update Lucene search indexes (if enabled) >>> >>> However, by default, you should only need to use "index-discovery" if you >>> have Discovery enabled, as the older search/browse options are no longer >>> necessary. >>> >>> If you wanted to re-enable the older search/browse options, there's >>> documentation on that at: >>> https://wiki.duraspace.org/display/DSDOC4x/Legacy+methods+for+re-indexing+content >>> (However, you probably should not run both Discovery and an older >>> search/browse, as Discovery now handles everything itself.) >>> >>> - Tim ------------------------------------------------------------------------------ Open source business process management suite built on Java and Eclipse Turn processes into business applications with Bonita BPM Community Edition Quickly connect people, data, and systems into organized workflows Winner of BOSSIE, CODIE, OW2 and Gartner awards http://p.sf.net/sfu/Bonitasoft _______________________________________________ DSpace-tech mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dspace-tech List Etiquette: https://wiki.duraspace.org/display/DSPACE/Mailing+List+Etiquette
