Why and how hibernate_sequence is using for id generatior

Question

I'm faced out an issue with ID generation for my entities generated by JHipster:

1. Dev environment on H2 database
2. Prod environment on Postgres
3. I have one entity "station" with two fields "id" and "name"
4. Creating a liquibase script which imports dictionary in "station" table like INSERT INTO station (name) VALUES ('Adygeya') without ID definition
5. Trying to add station on dev environment - OK
6. Trying to add station on prod - Hibernate tries to add new station with duplicated ID

WHY?

My research was shown that only for postgres and oracle in initial scheme jhipster created a new sequince "hibernate_sequence" which is used for new entities creation.

So I'm fixed out this wrong behaivour by adding specific sequence name for my entity ID generation rule

@Id
@GeneratedValue(strategy = GenerationType.AUTO, generator = "station_gen")
@SequenceGenerator(name = "station_gen", sequenceName = "station_id_seq")
private Long id;

Now I have just 3 questions:

1. Why JHipster using one sequince for all tables for postgres and oracle?
2. Where is it configured?
3. What am I doing wrong?

Answer 1

Will post here my answer wich was found with help of @GaëlMarziou and @Julien Dubois.

Thats because of history and default behaivour of Hibernate hilo algorithm with GenerationType.AUTO. MySQL doesn't support sequencies so JHipster needs to use this stupid algorithm

JHipster team has found right solution wich will fix my problem. The solution is to use GenerationType.SEQUENCE for all DB but use GenerationType.IDENTITY for MySql. All the details in this commit https://github.com/jhipster/generator-jhipster/commit/4516b4ff4d49a96a75fd963b0c7667f198bd9b79

This way I will configure my entities now too.