You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We currently have a PostgreSQL database for identity management by the Kratos service and a MySQL database for the other Serlo data used by database-layer and notification-mail-service.
Having two (or more) databases is a principle of microservices architecture, but in our small team it leads to some unnecessary complexities:
relations between databases are not reliable, since foreign key contraits is not available
migrating the content to another database is complex (v.g. we would have to create a quite clever migration in order to deal with current fk constraints regarding users in the mysql database)
Steps:
rename database of kratos in postgres to serlo
adapt database connection string of Kratos
migrate data from MySQL into the new database in PostgreSQL
adapt code of database-layer
Note:
Also given the opportunity we would long term like to get rid of MySQL because it is owned by Oracle who are closing the once open project. PostgreSQL is real FOSS and the industry preference for the last years + some good features of PostgreSQL.
RICE Score
Reach and impact: 5 -- For standard tasks in database-layer, we may not notice a difference but makes things like dumping, migrations easier.
Confidence: 8 -- Such migrations have been done before but we might have to research best tools to migrate.
Effort: 6 -- Requires investigation first, configuration with Google Cloud, have to think it through and test very well because errors have high impact.
The text was updated successfully, but these errors were encountered:
We currently have a PostgreSQL database for identity management by the Kratos service and a MySQL database for the other Serlo data used by database-layer and notification-mail-service.
Having two (or more) databases is a principle of microservices architecture, but in our small team it leads to some unnecessary complexities:
Steps:
serlo
Note:
Also given the opportunity we would long term like to get rid of MySQL because it is owned by Oracle who are closing the once open project. PostgreSQL is real FOSS and the industry preference for the last years + some good features of PostgreSQL.
RICE Score
The text was updated successfully, but these errors were encountered: