We have a Neo4J 2.1.3 database, and we have a uniqueness constraint that was created as follows:
CREATE ABBREVIATION ON (segment: SEGMENT) ASSERT segment.segmentId IS UNIQUE
When we check this from the browser, it works great when detecting violations, for example:
CREATE (n: SEGMENT {name: "duplicate", segmentId: 110484}) RETURN n
leads to
Node 589 already exists with the SEGMENT label and the property "segmentId" = [110484]
Neo.ClientError.Schema.ConstraintViolation
what well.
We have the installation of a load tester with 3 machines and several threads in a box using Cypher over REST, talking to Neo4J, and using transactional endpoints to create similarly above (but, of course, many more properties related to our application) and always recording Neo4J master to configure HA.
In this installation, we can reliably reproduce multiple uniqueness violation violations that are NOT caught by Neo4J, they are executed without errors, and in the resulting db we can see several nodes with the SEGMENT label and the same value for the segmentId property (we intentionally generate repeated segmentId values for our test )
Does anyone else face the same issue? Is this a Neo4J bug?
Thank.