[
https://issues.apache.org/jira/browse/WICKET-2946?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12891033#action_12891033
]
Hudson commented on WICKET-2946:
--------------------------------
Integrated in Apache Wicket 1.4.x #53 (See
[http://hudson.zones.apache.org/hudson/job/Apache%20Wicket%201.4.x/53/])
WICKET-2946
Catch just Exception, not Throwable
> Exceptions raised in overridden sessionDestroyed() event (of WebApplication
> class) are completely swallowed by the session-store classes without even
> logging anything
> ----------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: WICKET-2946
> URL: https://issues.apache.org/jira/browse/WICKET-2946
> Project: Wicket
> Issue Type: Bug
> Components: wicket
> Affects Versions: 1.3.6, 1.4.9
> Environment: Redhat Linux 4. JDK 1.5
> Reporter: Farhan
> Assignee: Juergen Donnerstag
> Fix For: 1.4.10
>
>
> A week ago our production server was being pounded continuously for days by
> the msnbot. Since all our wicket-pages, including LoginPage, are currently
> stateful in nature, each hit would result in session-store directory getting
> created (as per the default DiskPageStore session implementation). This also
> at one time came to a point, where the number of session-folders exceeded 32k
> sub-directory limit, of linux-kernel, often resulting in some (valid)
> existing users session getting expired suddenly.
> Digging into the problem revealed that there was a semantic error in our
> MyCustomApplication.sessionDestroyed() event, where a NullPointer Exception
> was getting raised, but since framework session-store (DiskPageStore) class
> was completely swallowing the exceptions raised, it took us quite a while to
> identify, what could have been a very obvious NPE, had it been at least
> logged by the catch() block. So if we could modify the catch block to at
> least log e.getMessage(), whenever an exception is raised.? that would save a
> lot of time of developers in such exceptional scenarios.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
