Exceeded the limit for exceeding the upper limit of Neo4j

I ran the test and I had the error "exceeding the limit on GC". I realized that this was because I loaded too many primitives into the cache. I'm wrong?

Then my question is, how can we prevent this? For example, can we estimate the size of the required memory based on the number of primitives? Is there a clue to know about this approximately?

My boss wants to know how many primitives we can manage at the same time. I assume this is due to the JVM settings, but it is not possible to find which settings.

Sorry if these are stupid questions, I'm not used to JVM settings and performance, and I have a pretty big lack of knowledge about this atm. Try and desire to understand though!

Jimmy.

+6
source share
2 answers

Understanding the details of Java garbage collectors is far from a trivial thing. Since you doubt that this is non-specific, I can simply give a rather non-specific answer. There, in the section of the Neo4j reference manual about JVM settings, http://docs.neo4j.org/chunked/stable/configuration-jvm.html .

Another idea, depending on your schedule and heap size, is to change the type of implementation for the object cache. There is an implementation of soft (by default in the community version), weak and strong . In addition, the enterprise version comes with an implementation of hpc (high-performance cache), which significantly reduces the number of complete garbage collections by dynamically adjusting the cache size. Read more at http://docs.neo4j.org/chunked/stable/configuration-caches.html#_object_cache .

+1
source

For me, this helped add / uncomment the following lines in neo4j-wrapper.conf :

 wrapper.java.initmemory=4096 wrapper.java.maxmemory=4096 

The problem is the amount of RAM.

You can find this file in "C: \ Program Files \ Neo4j Community \ neo4j-community-2.1.3 \ conf" on a Windows system, for example.

PS: another good article on what else can help you ( http://jexp.de/blog/2014/06/load-csv-into-neo4j-quickly-and-successfully/ ).

0
source

All Articles