With the help of the other members of the BPM A-Team (Sushil Shukla, John Featherly, Siming Mu) I have put together a concise list of points that should be at least visited prior to moving into production with BPM 11g/12c. Note that it is BPM 11g/12c specific: although it touches on areas common to other common parts of SOA Suite (such as clustering etc...) it does not cover in any way other products within SOA Suite (such as BPEL, Mediator etc...).
The list is by no means exhaustive, just a collection of points that experience has shown need to be understood and validated. Some may not be relevant in all customer projects.
The items on the list are left deliberately without explanations - if the item means little or nothing to you, go away and research it, there are a collection of links to useful documents at the end of this blog.
□ Validate Clustering & failover
□ Verify DR
□ Verify patching
□ Validate load balancer, specifically algorithm and LB health check to soa-infra
□ Validate IDM, access paths, security, authentication and authorization
□ Move Policy Store to DB / OID
□ Research and test major component type and version (JDK - HotSpot, JRockit, 6/7, Browser - IE, Chrome, Firefox)
□ All paths in all processes traversed
□ Empty XML elements – Selection Failures
□ Failure areas negatively tested
□ Boundary events
□ Event sub-processes
□ Fault policies – retries, manual intervention etc..
□ Business exceptions
□ Tests for non-BPM code; pojo’s, stored procedures etc.
□ OBR functional test (if not under separate lifecycle management)
□ Validate dashboards, reports etc.
□ Load testing – average/peak
□ Stress testing – how far can the system be pushed
□ Negative testing under load
□ Inbound & outbound failures to simulate increasing load
□ Managed server failure / cluster failure
□ Unavailable infrastructure components – network routers, switches, DB etc.
□ Service unavailability – LDAP etc.
□ Engine & workspace performance when DB both sparsely & heavily populated
□ DB partitioning implemented
□ Purging strategy understood and tested
□ Other standard database operations for backup, recovery test etc.
□ Process instances queried, recovered & continued after failure both individually and in bulk
□ Re-Deployment / Versioning of new processes
□ Instance patching
□ Instance migration
□ Query individual process instances based on business data – EM / Workspace / scripts etc...
□ Alter Flow & alter data
□ Intrusion Detection (PEN)
□ Release Management, UAT and Promote to Production
□ Version control system
□ Environment Spec and Configuration definition