I was confused by the fact that we can allow to use JTA transactions with a non-XA-datasource. Link to the documentation. So what is the difference between XA/non-XA datasources? Why should we use XA-datasources at all?
An XA transaction, in the most general terms, is a "global transaction" that may span multiple resources. A non-XA transaction always involves just one resource.
An XA transaction involves a coordinating transaction manager, with one or more databases (or other resources, like JMS) all involved in a single global transaction. Non-XA transactions have no transaction coordinator, and a single resource is doing all its transaction work itself (this is sometimes called local transactions).
Note: The explanation above was taken from: theserverside (Mike Spille)
jta="true", Transaction commit automatically.
I was wondering about this myself ("use JTA" option in a non-XA Datasource) so I tested several configurations. I have a distributed transaction connecting to two MySQL servers.
Here are my results. If I have:
- Two non-XA datasources, both have JTA="true"
Result: Error "Could not enlist in transaction on entering meta-aware object."
- Two non-XA datasources, with one JTA="true"
Result: They won't participate in the distributed transaction. Each will commit separately.
- One XA and one non-XA with JTA="false",
Result: same as #2
- One XA and one non-XA with JTA="true".
Result: Works!
From these, it looks like "use JTA" option indicates if it will participate in a distributed transaction if there's an XA datasource.
© 2022 - 2024 — McMap. All rights reserved.