Bulk loading CPU busy


Edoardo N <e.n...@...>
 

Hi!

sorry for the question but i'm new with janusgraph. How can I import csv file into janus???


Il giorno martedì 23 maggio 2017 09:41:15 UTC+2, CGen ha scritto:
Hello!

I need to load CSV files into JanusGraph with a size of 1.7 gigabytes.

How can I increase the load speed? 

I added to the config:
storage.batch-loading=true
schema.default=none

I have almost no access to the disk, but the CPU is always busy.

In the profiler, the hot spot is: org.janusgraph.graphdb.internal.AbstractElement equals (8%).


Rafael Fernandes <luizr...@...>
 

Can u try setting this property and compare?

query.fast-property=true

Thanks
Rafa


CGen <new...@...>
 

I put Vertices in Map. This was the reason for frequent equals calls.
I removed this.

Now hot spots is:

java.lang.Object.hashCode[native] () (6,1%)
org.janusgraph.graphdb.internal.AbstractElement.equals (Object) (5,3%)
java.security.AccessController.doPrivileged[native] (java.security.PrivilegedExceptionAction) (4,9%)


вторник, 23 мая 2017 г., 10:41:15 UTC+3 пользователь CGen написал:

Hello!

I need to load CSV files into JanusGraph with a size of 1.7 gigabytes.

How can I increase the load speed? 

I added to the config:
storage.batch-loading=true
schema.default=none

I have almost no access to the disk, but the CPU is always busy.

In the profiler, the hot spot is: org.janusgraph.graphdb.internal.AbstractElement equals (8%).


CGen <new...@...>
 

Hello!

I need to load CSV files into JanusGraph with a size of 1.7 gigabytes.

How can I increase the load speed? 

I added to the config:
storage.batch-loading=true
schema.default=none

I have almost no access to the disk, but the CPU is always busy.

In the profiler, the hot spot is: org.janusgraph.graphdb.internal.AbstractElement equals (8%).