eclipse-vertx
Add support for IBM Power System i
General
We're encountering issues using Reactive Hibernate with Vert.x reactive db2 client and DB2 hosted on IBM Power i (AS400) v7.4. It appears that the Vert.x Reactive DB2 Client explicitly supports only DB2 on Linux, Unix, and Windows and and possibly DB2 for z/OS as mentioned here.
Problem Description:
When attempting to connect, the server's release version string on IBM Power i begins with "V" (e.g., V07R04M00). The Vert.x reactive db2 client expects the release version to start with either SQL (for Linux/Unix/Windows) or DSN (for DB2 z/OS). This mismatch triggers the following exception:
java.lang.IllegalArgumentException: Received unknown server product release level: V07R04M00
As a result, the DB connection initialization immediately fails.
Environment:
Reactive Hibernate
Vert.x
IBM Power i v7.4
Feature
Is the lack of support for DB2 on IBM Power i due to known compatibility issues or simply because it is a proprietary system and hasn't been tested/supported yet?
Clarification on this point would help us understand if there are inherent limitations or if future support could be expected.
the current implementation for DB2 seems good, however there is a lack of contributor with DB2 specific knowledge.
we would welcome any contributor helping there
things are documented but the protocol seems to be quite complicated to get (unlike PostgreSQL for instance)
Okay, thank you for your answer. Using DB2 on Power i together with vert.x is probably quite a niche, yes. Do you think it's worth a try for us to get around the thrown exception? Or is the protocol that different from z/OS and DB2 on Windows/Linux?
Protocol should be the same regardless of the OS.
I don't exactly know what means limited for zOS in this context, we should ask the original contributor what he meant about this I think. @aguibert any idea
Do you think it's worth a try for us to get around the thrown exception? Or is the protocol that different from z/OS and DB2 on Windows/Linux?
If I remember correctly the OS check that's being hit here is because initial handshake between Db2/z and Db2 LUW are different. I don't remember how much it varies after that point, but my guess is even if you tried to go down the LUW or z path for Db2 for i you would hit issues.
