[jbossts-issues] [JBoss JIRA] (JBTM-2575) When checking for orphaned subordinate transactions in the middle of a tree branches that are eligible for orphan detection will be rolled back

RH Bugzilla Integration (JIRA) issues at jboss.org
Mon Apr 25 06:13:01 EDT 2016


    [ https://issues.jboss.org/browse/JBTM-2575?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13196379#comment-13196379 ] 

RH Bugzilla Integration commented on JBTM-2575:
-----------------------------------------------

Vladimir Dosoudil <dosoudil at redhat.com> changed the Status of [bug 1289386|https://bugzilla.redhat.com/show_bug.cgi?id=1289386] from MODIFIED to ON_QA

> When checking for orphaned subordinate transactions in the middle of a tree branches that are eligible for orphan detection will be rolled back
> -----------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: JBTM-2575
>                 URL: https://issues.jboss.org/browse/JBTM-2575
>             Project: JBoss Transaction Manager
>          Issue Type: Bug
>          Components: Recovery
>            Reporter: Tom Jenkinson
>            Assignee: Tom Jenkinson
>            Priority: Blocker
>             Fix For: 4.17.31, 5.2.9.Final
>
>
> There is a check in the subordinate orphan detection that not only checks for matching gtrid but also for matching subordinate name. This will not match correctly for an intermediary node. E.g.
>  a->b b->c
> When b scans c the xid it gets back will have subordinate name of c, b will look in its object store and match the subordinate on gtrid but the subordinate node ID in b subordinateatomicaction will be "b".
> This check is actually superfluous anyway. We already know that the Xid returned from c was for b because of transport level checks.



--
This message was sent by Atlassian JIRA
(v6.4.11#64026)


More information about the jbossts-issues mailing list