[jboss-jira] [JBoss JIRA] Commented: (JBAS-3242) EE spec requires that web container aborts any active txn at the end of a request
[ http://jira.jboss.com/jira/browse/JBAS-3242?page=comments#action_12340624 ]
Adrian Brock commented on JBAS-3242:
------------------------------------
I'll think you'll find the "performance penalty" is actually the
checking done by the
CachedConnectonManager (CCM) when debug=true.
I'd say we enable it by default (if people don't want it, they can always disable
it themselves).
But we get QA to look at the relative performance of this
1) Disabled
2) Enabled with debug=true
3) Enabled with debug=false
We could also split this into two valves.
1) For the CCM checking
2) For the transaction checking
A further optimation would be to look at the resource type.
e.g. There is no need to check for incomplete transactions if this is a static file being
served.
EE spec requires that web container aborts any active txn at the end
of a request
---------------------------------------------------------------------------------
Key: JBAS-3242
URL: http://jira.jboss.com/jira/browse/JBAS-3242
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Web (Tomcat) service
Affects Versions: JBossAS-4.0.4.GA
Reporter: Gavin King
Assigned To: Remy Maucherat
Priority: Critical
Fix For: JBossAS-4.0.6.CR1
The Java EE spec requires that if there is any active JTA transaction at the end of a
servlet request, then the container must abort the txn. JBoss currently does not do this.
Instead it leaves orphaned transactions dangling about on pooled threads, ready and
waiting to be resumed by any random request that comes along.
This is really Bad behavior and I'm kinda shocked.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira