[jira] Updated: (GERONIMO-5649) txmanager could try to replace dead XAResources in commit and rollback tasks

["David Jencks (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/GERONIMO-5649?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

David Jencks updated GERONIMO-5649:
-----------------------------------

    Attachment: GERONIMO-5649-1.patch

This attempts to solve getting a new NamedXAResource when the existing one breaks, retry on recovery (GERONIMO-5648), and setting the tx timeout on XAResources (GERONIMO-452).

Assumes TMFAIL means we need a new XAResource, TMRETRY means retry the existing one.  Sets tx timeout to be time left on existing tx.

This is for the 2.2 txmanager branch.

> txmanager could try to replace dead XAResources in commit and rollback tasks
> ----------------------------------------------------------------------------
>
>                 Key: GERONIMO-5649
>                 URL: https://issues.apache.org/jira/browse/GERONIMO-5649
>             Project: Geronimo
>          Issue Type: Bug
>      Security Level: public(Regular issues) 
>          Components: transaction manager
>    Affects Versions: 2.2, 3.0
>            Reporter: David Jencks
>            Assignee: David Jencks
>             Fix For: 2.2.1, 3.0
>
>         Attachments: GERONIMO-5649-1.patch
>
>
> Generally if a RM crashes, outbound connections to it need to be reestablished.  So if we have a in-doubt tx using an outbound connection, we should try to get a new XAResource to finish up the branch in the CommitTask and proposed RollbackTask.
> AFAIK inbound connections will automatically reestablish connections and use the RETRY error code so we don't need to do this for inbound.  I still need to find out which error codes indicate that we should get a new connection and try again.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.