Multiple F5 refreshes create multiple threads

[Mateusz Mieszkowski <mm...@objectivity.co.uk>]

Hi,

I'm having a problem with quite old version of Wicket (1.3.6) and I'm wondering if this still exists in the latest version as we're getting ready for upgrading.
The thing is that if the page opened in browser uses SortableDataProvider, iterator method is executed. So far so good...
But each time user do F5, page refresh creates new http thread and every one of them is trying to process iterator method.
Since this is the same user/session, I think it is safe to say that only one thread(that is doing the refresh) is necessary and more refreshes should kill http threads that are working to get the same info.

The problem in my system was, that one user's F5 blocked and it caused hundreds of refresh and started hundreds of http thread. I imagine that in worse case scenario it could use up whole thread pool and system would be useless for some time.

I've checked changelogs and jira issues but haven't found anything similar. Could you please let me know if there is a way to avoid such situation?

Kind regards,
Mateusz Mieszkowski

Re: Multiple F5 refreshes create multiple threads

[Martin Makundi <ma...@koodaripalvelut.com>]

Hi!

Not knowing what 1.3 does, but 1.4 locks pagemap in org.apache.wicket.Session.

**
Martin

2013/7/18 Mateusz Mieszkowski <mm...@objectivity.co.uk>:
> Hi,
>
> I'm having a problem with quite old version of Wicket (1.3.6) and I'm wondering if this still exists in the latest version as we're getting ready for upgrading.
> The thing is that if the page opened in browser uses SortableDataProvider, iterator method is executed. So far so good...
> But each time user do F5, page refresh creates new http thread and every one of them is trying to process iterator method.
> Since this is the same user/session, I think it is safe to say that only one thread(that is doing the refresh) is necessary and more refreshes should kill http threads that are working to get the same info.
>
> The problem in my system was, that one user's F5 blocked and it caused hundreds of refresh and started hundreds of http thread. I imagine that in worse case scenario it could use up whole thread pool and system would be useless for some time.
>
> I've checked changelogs and jira issues but haven't found anything similar. Could you please let me know if there is a way to avoid such situation?
>
> Kind regards,
> Mateusz Mieszkowski

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@wicket.apache.org
For additional commands, e-mail: users-help@wicket.apache.org