Community
Participate
Working Groups
With Linux (Ubuntu 10.04, 64bit) we get IOExceptions quite soon (after a few thousand records) because of "too many open files" (ulimit -n is 16384). It seems that Lucene uses 1000s of file handles. The situation is a bit better when ForceFlush="false" in the DataDictionary: I can add my 16000 test documents then, but there are still about 8000 open file handles on Lucene files. It seems to be better on Windows machines.
Is this issue still relevant since we now use Solr?
guess not
Ok. Closing this one.
Closing this