Supported for the workflow capability (24.0.x): Db2 (11.5 / 12.1, including Db2 on Cloud), PostgreSQL (community 13-16 and EDB Postgres), Oracle (19c / 21c), and Microsoft SQL Server on some releases and components - the detailed system requirements per version list the exact combinations, and not every capability supports every database (e.g. Case / Content needs Db2, PostgreSQL or Oracle; BAI uses OpenSearch). The databases run outside the cluster (existing servers, cloud managed services) or inside it through a database operator - CP4BA itself does not create them, except the built-in PostgreSQL of the starter / trial profile and the foundational services' own metastore.
Reusing existing databases: yes for a migration (the documented traditional-to-containers migration points the CR at the upgraded BPMDB / PDW / CMNDB of the traditional installation), no for a fresh installation of the same version next to a running traditional server (two engines must not share a BPMDB). Preparation is the same as on traditional: create the databases with the scripts the operator generates (cp4a-prerequisites.sh produces the DDL for your database type), users, and the JDBC secrets.
# secrets the workflow CR references for the engine databases
oc create secret generic bawins1-db-secret --from-literal=dbUser=bawuser --from-literal=password='...' -n cp4ba
spec:
baw_configuration:
- name: bawins1
database:
type: db2 # db2 | postgresql | oracle | sqlserver (per release)
server_name: db2.example.com
port: 50000
database_name: BAWDB
secret_name: bawins1-db-secret
cm_max_pool_size: 100
enable_ssl: true
db_cert_secret_name: db2-ssl-cert # server certificate when TLS is enforcedSizing and tuning are unchanged (the same tables, indexes and cleanup jobs); PostgreSQL is the common choice for new deployments and the only option on Workflow Process Service; Db2 stays the safest for large migrated estates.
References