Re: Avoiding the system IO cache for compaction

2011-08-02 Thread Jason Rutherglen
Looks like there were issues with the native parts. Mike McCandless notes this in [1]. If we reuse the Lucene code, I think it'll work for Cassandra. 1. https://issues.apache.org/jira/browse/LUCENE-2500 On Tue, Aug 2, 2011 at 9:42 PM, Jonathan Ellis wrote: > Tried and failed. https://issues.ap

Re: Avoiding the system IO cache for compaction

2011-08-02 Thread Jonathan Ellis
Tried and failed. https://issues.apache.org/jira/browse/CASSANDRA-1902 On Tue, Aug 2, 2011 at 11:24 PM, Jason Rutherglen wrote: > Cassandra 'compacts' the way Lucene 'merges' segments.  One > interesting new feature built into Lucene is [1] which avoids loading > the source files into the system