RE: [business-transaction] Issue - maint-13 - CONFIRM_ONE_PHASE report-hazard default

["Furniss, Peter" <Peter.Furniss@choreology.com>]

Title: Message

When this came up at today's meeting, checking the text led to me to view that CONFIRM_ONE_PHASE shouldn't have a report-hazard field at all.

 

CONFIRM_TRANSACTION and CANCEL_TRANSACTION are control protocol messages - essentially the distributed api to the coordinator, and so it is reasonable for them to decide whether they want to receive hazard reports from the tree.

 

But CONFIRM_ONE_PHASE is an outcome message, sent only if there turns out to be a single remaining enrolled inferior of a superior. If there had been two inferiors, the superior would have sent PREPARE then CONFIRM to each (with appropriate variations). The determination of whether hazards are reported to a superior is made independently by the inferior (from configuration, probably) - an inferior can choose to suppress hazard/heuristic events if it so chooses. Since CONFIRM_ONE_PHASE is only sent by chance (so to speak), why should it be able to control the reporting.

 

Peter

 

Issue - maint-13 - CONFIRM_ONE_PHASE report-hazard default

Status: open
Date added: 11 May 2004
Submitter: Alex Ceponkus Ceponkus
Date submitted: 30 March 2004
Description:

CONFIRM_ONE_PHASE report-hazard has a default, and is not consistent with report-hazard found on CONFIRM_TRANSACTION and CANCEL_TRANSACTION.
Submitter's solution:

• Remove "Default is false." from report-hazard element description found in section 7.7.11 CONFIRM_ONE_PHASE.
• Remove bold style from "false" in confirm-one-phase abstract xml. line 3741 Remove: minOccurs="0" default="false" from report-hazard element in confirm-one-phase (line 346 of schema.)

Changes: 11 May 2004 - new issue

---