Hi everyone,

I am using Janusgraph version 5.2 with cassandra backend.

When using

graph = GraphFactory.open(Resources.getResource("read-cql.properties").getFile());

g = inputGraph.traversal().withComputer(SparkGraphComputer.class);

g.V().connectedComponent().iterate();

 where "read-cql.properties" is the provided configuration file of janusgraph.

I get the following Logger output:

WARN  org.apache.spark.executor.Executor - Managed memory leak detected

This also happens in all sort of other Olap queries.

Is there something I can do about it? This does not necessarily sound so good.

Thanks for any suggestions!

Lilly

Hi Lilly,

I did some googling first and saw that this error message has a long history in the Apache Spark backlog and still shows up in recent versions:

https://issues.apache.org/jira/browse/SPARK-30443

So, I suppose the action lies with Apache Spark and in the mean time I hope you do not run into memory problems. I agree that it "does not necessarily sound so good", but just see it as an explicit bug among the invisible bugs :-).

Best wishes,    Marc

Op maandag 7 september 2020 om 17:22:03 UTC+2 schreef Lilly:

Hi everyone,

I am using Janusgraph version 5.2 with cassandra backend.

When using

graph = GraphFactory.open(Resources.getResource("read-cql.properties").getFile());

g = inputGraph.traversal().withComputer(SparkGraphComputer.class);

g.V().connectedComponent().iterate();

 where "read-cql.properties" is the provided configuration file of janusgraph.

I get the following Logger output:

WARN  org.apache.spark.executor.Executor - Managed memory leak detected

This also happens in all sort of other Olap queries.

Is there something I can do about it? This does not necessarily sound so good.

Thanks for any suggestions!

Lilly