Introduction
By default Neo4j is set to use the following RAM:
| CAST Imaging ≥ 2.5.0 | 8GB of RAM memory (initial and maximum) |
|---|---|
| CAST Imaging ≤ 2.4.x | 3GB of RAM memory (initial and maximum) |
If you are facing performance issues, it is possible to modify the allocated memory. See also https://neo4j.com/docs/operations-manual/current/performance/memory-configuration/#_considerations for more information.
How do performance issues manifest themselves?
In general, a sure sign that you need to modify the RAM allocated to Neo4j is when objects fail to load in CAST Imaging - particularly if there are a large number of objects to display - this is because the query used to fetch the objects for display is not completing before a timeout occurs. This timeout will manifest itself as a 504 (gateway timeout) error in the browser. Allocating additional RAM to Neo4j can help reduce the time required to complete the query.
In ≥ 2.19, during installation, if the Neo4j fails to start (due to lack of memory) it will not block the entire installation process. Instead, the installation will continue but the corresponding services will not start. The issue can be resolved by modifying the allocated memory.
Update process
Edit the configuration file
Open and edit the following file:
Microsoft Windows traditional installer:
%APPDATA%\CAST\ImagingSystem\neo4j\neo4j.conf
Docker Installer extension (located in the folder created when unzipping the extension):
neo4j\configuration\neo4j.conf
Microsoft Windows via traditional installer
This file is located in the protected %APPDATA% location, therefore you must open the file with elevated permission (this is usually achieved by right clicking your text editor in the Windows start menu and selecting Run as administrator):
Linux
You may need to use elevated permissions to edit this file (for example use sudo).
Alter the following lines and increase the RAM allocation as necessary:
dbms.memory.heap.initial_size=8G dbms.memory.heap.max_size=8G
Apply the changes
Traditional Microsoft Windows installation
If you have installed CAST Imaging using the traditional Microsoft Windows installer, you must run the following command to ensure that the new settings are applied correctly to the Neo4j Windows Service:
neo4j update-service
You should run the command from a CMD window in the following location:
%PROGRAMFILES%\CAST\ImagingSystem\neo4j\bin
Finally, restart the Neo4j Microsoft Windows service:
Docker deployment
Restart the Neo4j Docker container to ensure the changes are taken into account.
Memrec optimization test
You can also run an optimization test (memrec) that will recommend memory allocation and database settings for you. To do so, run the following commands:
Microsoft Windows via traditional installer
Input the amount of available RAM on the server (for a server with 16GB of RAM, enter "16G" as in the example below):
cd %PROGRAMFILES%\CAST\ImagingSystem\neo4j\bin neo4j-admin memrec --memory=16G
The tool will recommend settings as follows:
# Based on the above, the following memory settings are recommended: dbms.memory.heap.initial_size=5g dbms.memory.heap.max_size=5g dbms.memory.pagecache.size=7g
Docker
Login to the neo4j docker container to run the memrec command. Make sure your neo4j container is up and running.
$ docker exec -it neo4j /bin/bash root@2d9e51d4be43:/var/lib/neo4j# cd bin root@2d9e51d4be43:/var/lib/neo4j/bin# ./neo4j-admin memrec --memory=16G
The tool will output the recommended settings. Exit the container:
root@2d9e51d4be43:/var/lib/neo4j/bin# exit
Database size test
The database size test should only be performed once you have imported Application data into CAST Imaging. It displays the size in MB of the Neo4j database, which can then be used to modify the following settings:
- dbms.memory.pagecache.size
Run the following command:
cd %PROGRAMFILES%\CAST\ImagingSystem\neo4j\bin neo4j-admin memrec --database=graph.db
The tool will recommend settings as follows:
# The numbers below have been derived based on your current data volume in database and index configuration of database 'graph.db'. # They can be used as an input into more detailed memory analysis. # Lucene indexes: 0.0 # Data volume and native indexes: 114647.0
As opposed to the memory recommendations from the Neo4j tool, you can also set the dbms.memory.pagecache.size value comparable to your data value.
