incubator-clerezza-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andy Seaborne <andy.seabo...@epimorphics.com>
Subject Re: leak but where after parsing rdf files?
Date Sun, 23 Jan 2011 20:29:27 GMT
Works for me (RIOT).

java version "1.6.0_20"
OpenJDK Runtime Environment (IcedTea6 1.9.2) (6b20-1.9.2-0ubuntu2)
OpenJDK 64-Bit Server VM (build 19.0-b09, mixed mode)

I've run to completion of all 100 iterations, and with the smaller -Xmx210m

There are about 6 iterations for the OOME below. This shows that cache 
maximum size isn't reached in a single pass.  Given the very large 
literals and the bNodes, this isn't surprising (the cache is a simple 
slot replacement cache).

The cache can be turned off with:

com.hp.hpl.jena.graph.Node.cache(false) ;

It then ran at -Xmx150.

As the files are about 21M bytes, that's 40Mbytes in java characters 
which is about right.

	Andy


On 22/01/11 11:38, Reto Bachmann-Gmuer wrote:
> To me this looks like a leak, it successfully parses the same bund of files
> a few time and then it has an OutOfMemoryError:
>
> $ MAVEN_OPTS="-Xmx215m -Xms128m"  mvn  exec:java -o -e
> + Error stacktraces are turned on.
> [INFO]
> NOTE: Maven is executing in offline mode. Any artifacts not already in your
> local
> repository will be inaccessible.
>
> [INFO] Scanning for projects...
> [INFO]
> ------------------------------------------------------------------------
> [INFO] Building Unnamed -
> org.apache.clerezza:turtlememory:bundle:0.10-incubating-SNAPSHOT
> [INFO]    task-segment: [exec:java]
> [INFO]
> ------------------------------------------------------------------------

Mime
View raw message