Performance insertion with and without index

I did a couple of tests.

Based on some great suggestions from Wes, etc., I set up some of the neo4j features without a cache that you need to embed on a large scale in a multi-threaded environment, and the performance is pretty good.

However, when I entered the index (on the nodes), performance was greatly degraded. The difference is 5 times. Are there any configuration settings to make them better?

Thanks in advance,

Sechin

Neo4j version - 1.8.1; JVM - 1.6

+5
source share
2 answers

Inserting nodes (or relationships) into the Lucene index is expensive. Lucene is a powerful yet sophisticated tool for finding full-text / keywords. Compared to an open database, it is rather slow.

, , Microsoft:

http://jexp.de/blog/2012/10/parallel-batch-inserter-with-neo4j/

:

http://blog.xebia.com/2012/11/13/combining-neo4j-and-hadoop-part-i/

SSD- . Neo4j ACID- , Lucene ( ), , .

+4

- , Neo4j ext4 Linux:

( USV/ ) 10-15 !

: http://structr.org/blog/neo4j-performance-on-ext4

0

All Articles