You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@geronimo.apache.org by "David Jencks (JIRA)" <ji...@apache.org> on 2010/02/19 21:06:28 UTC
[jira] Created: (GERONIMO-5152) Transaction manager doesn't retry
stuff that might work later
Transaction manager doesn't retry stuff that might work later
-------------------------------------------------------------
Key: GERONIMO-5152
URL: https://issues.apache.org/jira/browse/GERONIMO-5152
Project: Geronimo
Issue Type: Bug
Security Level: public (Regular issues)
Components: connector
Affects Versions: 2.2, 3.0
Reporter: David Jencks
Assignee: David Jencks
Fix For: 2.2.1, 3.0
There are a number of situations where the transaction manager should respond to failure by waiting a while and trying again.
These include getting a retry error and the backend not being available when the tm starts so a recovery connection is not available.
To solve this we'll need to start by reorienting the relationship between connectors and the tm by giving the tm a NamedXAResourceFactory so it can ask for some way to retry stuff for a given name.
Then we'll need a list of retry tasks and a timer to execute them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (GERONIMO-5152) Transaction manager doesn't retry
stuff that might work later
Posted by "David Jencks (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/GERONIMO-5152?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12836089#action_12836089 ]
David Jencks commented on GERONIMO-5152:
----------------------------------------
In trunk (component version currently 2.2)
rev 911974 does some genericization and sets up the connector to tm relationship so recovery retry is possible.
rev 912058 adds retry tasks for recovery and retry errors on commit.
I think there are some more situations to take care of during commit failures where the original XAResource becomes unusable and we need to use the NamedXAResourceFactory to get a replacement but I haven't figured out which XA error codes would indicate this situation.
> Transaction manager doesn't retry stuff that might work later
> -------------------------------------------------------------
>
> Key: GERONIMO-5152
> URL: https://issues.apache.org/jira/browse/GERONIMO-5152
> Project: Geronimo
> Issue Type: Bug
> Security Level: public(Regular issues)
> Components: connector
> Affects Versions: 2.2, 3.0
> Reporter: David Jencks
> Assignee: David Jencks
> Fix For: 2.2.1, 3.0
>
>
> There are a number of situations where the transaction manager should respond to failure by waiting a while and trying again.
> These include getting a retry error and the backend not being available when the tm starts so a recovery connection is not available.
> To solve this we'll need to start by reorienting the relationship between connectors and the tm by giving the tm a NamedXAResourceFactory so it can ask for some way to retry stuff for a given name.
> Then we'll need a list of retry tasks and a timer to execute them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (GERONIMO-5152) Transaction manager doesn't retry
stuff that might work later
Posted by "David Jencks (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/GERONIMO-5152?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12897169#action_12897169 ]
David Jencks commented on GERONIMO-5152:
----------------------------------------
rev 984318 ports rev 939929 to branches 2.1 txmanager.
> Transaction manager doesn't retry stuff that might work later
> -------------------------------------------------------------
>
> Key: GERONIMO-5152
> URL: https://issues.apache.org/jira/browse/GERONIMO-5152
> Project: Geronimo
> Issue Type: Bug
> Security Level: public(Regular issues)
> Components: connector, transaction manager
> Affects Versions: 2.2, 3.0
> Reporter: David Jencks
> Assignee: David Jencks
> Fix For: 2.2.1, 3.0
>
>
> There are a number of situations where the transaction manager should respond to failure by waiting a while and trying again.
> These include getting a retry error and the backend not being available when the tm starts so a recovery connection is not available.
> To solve this we'll need to start by reorienting the relationship between connectors and the tm by giving the tm a NamedXAResourceFactory so it can ask for some way to retry stuff for a given name.
> Then we'll need a list of retry tasks and a timer to execute them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Closed: (GERONIMO-5152) Transaction manager doesn't retry
stuff that might work later
Posted by "David Jencks (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/GERONIMO-5152?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
David Jencks closed GERONIMO-5152.
----------------------------------
Resolution: Fixed
This seems to work!
> Transaction manager doesn't retry stuff that might work later
> -------------------------------------------------------------
>
> Key: GERONIMO-5152
> URL: https://issues.apache.org/jira/browse/GERONIMO-5152
> Project: Geronimo
> Issue Type: Bug
> Security Level: public(Regular issues)
> Components: connector, transaction manager
> Affects Versions: 2.2, 3.0
> Reporter: David Jencks
> Assignee: David Jencks
> Fix For: 2.2.1, 3.0
>
>
> There are a number of situations where the transaction manager should respond to failure by waiting a while and trying again.
> These include getting a retry error and the backend not being available when the tm starts so a recovery connection is not available.
> To solve this we'll need to start by reorienting the relationship between connectors and the tm by giving the tm a NamedXAResourceFactory so it can ask for some way to retry stuff for a given name.
> Then we'll need a list of retry tasks and a timer to execute them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (GERONIMO-5152) Transaction manager doesn't retry
stuff that might work later
Posted by "David Jencks (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/GERONIMO-5152?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
David Jencks updated GERONIMO-5152:
-----------------------------------
Component/s: transaction manager
> Transaction manager doesn't retry stuff that might work later
> -------------------------------------------------------------
>
> Key: GERONIMO-5152
> URL: https://issues.apache.org/jira/browse/GERONIMO-5152
> Project: Geronimo
> Issue Type: Bug
> Security Level: public(Regular issues)
> Components: connector, transaction manager
> Affects Versions: 2.2, 3.0
> Reporter: David Jencks
> Assignee: David Jencks
> Fix For: 2.2.1, 3.0
>
>
> There are a number of situations where the transaction manager should respond to failure by waiting a while and trying again.
> These include getting a retry error and the backend not being available when the tm starts so a recovery connection is not available.
> To solve this we'll need to start by reorienting the relationship between connectors and the tm by giving the tm a NamedXAResourceFactory so it can ask for some way to retry stuff for a given name.
> Then we'll need a list of retry tasks and a timer to execute them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (GERONIMO-5152) Transaction manager doesn't retry
stuff that might work later
Posted by "David Jencks (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/GERONIMO-5152?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12850448#action_12850448 ]
David Jencks commented on GERONIMO-5152:
----------------------------------------
rev 928119 ports the server trunk changes to the 2.2 branch. Changes will be needed to port this to 2.1 since much of the work was using gbean annotations instead of gbeanInfo.
> Transaction manager doesn't retry stuff that might work later
> -------------------------------------------------------------
>
> Key: GERONIMO-5152
> URL: https://issues.apache.org/jira/browse/GERONIMO-5152
> Project: Geronimo
> Issue Type: Bug
> Security Level: public(Regular issues)
> Components: connector, transaction manager
> Affects Versions: 2.2, 3.0
> Reporter: David Jencks
> Assignee: David Jencks
> Fix For: 2.2.1, 3.0
>
>
> There are a number of situations where the transaction manager should respond to failure by waiting a while and trying again.
> These include getting a retry error and the backend not being available when the tm starts so a recovery connection is not available.
> To solve this we'll need to start by reorienting the relationship between connectors and the tm by giving the tm a NamedXAResourceFactory so it can ask for some way to retry stuff for a given name.
> Then we'll need a list of retry tasks and a timer to execute them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (GERONIMO-5152) Transaction manager doesn't retry
stuff that might work later
Posted by "David Jencks (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/GERONIMO-5152?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12850242#action_12850242 ]
David Jencks commented on GERONIMO-5152:
----------------------------------------
rev 927980 ports the txmanager trunk changes above to the 2.1 txmanager branch. No changes beyond a merge required.
> Transaction manager doesn't retry stuff that might work later
> -------------------------------------------------------------
>
> Key: GERONIMO-5152
> URL: https://issues.apache.org/jira/browse/GERONIMO-5152
> Project: Geronimo
> Issue Type: Bug
> Security Level: public(Regular issues)
> Components: connector
> Affects Versions: 2.2, 3.0
> Reporter: David Jencks
> Assignee: David Jencks
> Fix For: 2.2.1, 3.0
>
>
> There are a number of situations where the transaction manager should respond to failure by waiting a while and trying again.
> These include getting a retry error and the backend not being available when the tm starts so a recovery connection is not available.
> To solve this we'll need to start by reorienting the relationship between connectors and the tm by giving the tm a NamedXAResourceFactory so it can ask for some way to retry stuff for a given name.
> Then we'll need a list of retry tasks and a timer to execute them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.