Importing Your Self Managed Neo4j 3.5.x Database into Aura

With the release of Neo4j 4 on Aura, you may wish to take advantage of the new features using an existing AuraDB Instance. If you have already been using 3.5 on Aura, please see this article for the steps to convert to using Neo4j 4.

If you have been using either the enterprise or community edition of Neo4j on your local server or workstation, this is the right article to review. 

Creating a dump file of Neo4j

For the full explanation of how to export a dump of Neo4j, please see our documentation on the topic. The abbreviated version is:

1. Log into the server running your Neo4j database.
2. cd to the directory containing your Neo4j installation. If you are unsure of the location and you are running on your local machine, you can open a terminal in Neo4j Desktop to navigate to it.
3. Stop the existing database by running bin/neo4j stop.
4. Make sure the backups directory we're going to use exists by running mkdir -p backups/graph.db. 
5. Take the backup bin/neo4j-admin dump --database=graph.db --to=backups/graph.db/`date +%Y%m%d`.dump.

Create a new AuraDB Instance

On the main Aura console page, scroll to the bottom of the page and choose "Create a database".

Under "Neo4j Version" choose 4, and select the region you want your AuraDB Instance deployed in. This does not need to be the same region as your existing AuraDB Instance. Note that if you have never deployed an AuraDB Instance before you may not see the option to choose 3.5, that option now only exists for customers who have used 3.5 in the past. 

As always, copy the password that is displayed for future use. Let the AuraDB Instance creation complete before moving to the next step. 

Import to the AuraDB Instance just created

The final step will be to import your data.

If the dump file is smaller than 4GB you can do this in the Aura console otherwise please use the command line tool push-to-cloud.

Open your new AuraDB Instance and click the Import Data tab. Then click the Choose file button to navigate to the location of the graph.db.dump file saved in the first step.

You'll be prompted to confirm you want to overwrite all of the existing data in your new AuraDB Instance; click Upload to continue. 

Wait for import to complete

We recommend getting a coffee, tea, or other favourite beverage at this step. It will take anywhere from a few minutes for a small dataset, to potentially an hour or more if you have an exceptionally large amount of data and a slow internet connection, for the import to finish. 

Check that the data was imported properly

Before you switch your application over to use the new AuraDB Instance, you'll want to perform some tests to make sure that everything is working as expected. One quick way to do this is to open a browser window for both the original and new AuraDB Instance and run this Cypher query against them, comparing the output:

MATCH (n)
RETURN DISTINCT labels(n) as Labels, count(labels(n)) as NumberOfNodes

The numbers of labels and nodes with them should be the same as your original database. If they are not, please open a ticket with our support team. 

Note that you'll still want to run whatever tests your developers have in place to be safe. 

Enjoy the new features available

With the exception of running multiple databases, there are a number of improvements available in Neo4j 4. We recommend reviewing the complete list of deprecations, additions, and compatibility notes for both versions 4.0 and 4.1 to gain a complete view of these changes.