[jbossts-issues] [JBoss JIRA] (JBTM-1866) HeuristicMixedException reported to client, although backends don't support heuristic completion

[Mark Little (JIRA)]

    [ https://issues.jboss.org/browse/JBTM-1866?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12794209#comment-12794209 ] 

Mark Little commented on JBTM-1866:
-----------------------------------

Christian, I also suggest you go and read the OTS specification from the OMG. Here's a snippet:

"A heuristic decision is a unilateral decision made by one or more participants in a transaction to commit or rollback updates without first obtaining the consensus outcome determined by the Transaction Service. A participant can only make a heuristic decision once the two-phase-commit protocol has begun, in particular it cannot make such a decision if it receives a rollback without a previous prepare. Heuristic decisions are normally made only in unusual circumstances, such as communication failures, that prevent normal processing. When a heuristic decision is taken, there is a risk that the decision will differ from the consensus outcome, resulting in a loss of data integrity."

And if you read all of it you will see that heuristic decisions are not necessarily just thrown (caused) by a single RM. I've already given this example, but I'll repeat: 2 RMs in the commit list, both prepare successfully and the coordinator logs the intention to commit; during the 2nd phase of 2PC if the 1st RM commits and the 2nd rolls back then we have a heuristic (HeuristicMixed). A heuristic decision that is reported by the coordinator to the client indicates that we have a non-atomic outcome.
                
> HeuristicMixedException reported to client, although backends don't support heuristic completion
> ------------------------------------------------------------------------------------------------
>
>                 Key: JBTM-1866
>                 URL: https://issues.jboss.org/browse/JBTM-1866
>             Project: JBoss Transaction Manager
>          Issue Type: Bug
>      Security Level: Public(Everyone can see) 
>          Components: Transaction Core
>    Affects Versions: 4.17.4
>            Reporter: Christian von Kutzleben
>            Assignee: Tom Jenkinson
>
> this bug is related to:
> https://issues.jboss.org/browse/JBTM-1865
> While testing recovery and transaction failures we came across
> two related issues, the test scenario is almost the same, the other
> issue is filed as JBTM-1865
> Scenario: 
> 2 enlisted XAResources (one for each database used by the test),
> The prepare phase is successful.
> We crash the second database server during the XAResource.commit invocation,
> _after_ it has committed it's branch, right before the control flow would return to the XAResource (which as a result throws a XAException.XA_RBCOMMFAIL).
> The TM can not know at this point in time, whether the database successfully committed the transaction or not. In this particular case, all changes have been committed.
> By definition of the 2PC protocol, the transaction should be considered as "to be committed", because the prepare phase has ended successfully.
> The first branch was successfully committed (as it should be spec.)
> But now things go awry:
> 1. The client receives a HeuristicMixedException, although the backend
> does not support heuristic completion at all, i.e. there should be no Heuristic* exception be thrown
> The correct behavior is, to report the transaction as success to the client, commit all other branches. All XAResources which failed during XAResource.commit() are supposed to be recovered in future.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira