Ryan McIntosh wrote:
> Long time listener, first time caller. I tried this patch out and
> it is good stuff :) I'm not much of a programmer, but I have a
> suggestion that it might be beneficial to set a bandwidth limit per
> connection (as well as an overall limit). Wouldn't this completely
> avoid that DOS situation without having to eat any CPU cycles?
Welcome Ryan! :)
Well, we have proposed three types of limitations:
- Per server: How much bandwidth will use the HTTP server.
- Per virtual server: Idem, at virtual server level.
- Resource: It will allow to set limits on extensions and
directories.
About the CPU cycles, I think we will have to look for more complex
solution for this issue: wait is not the right way, and try to keep
the connection inside the mail loop inhibits the server to slow down
even if there is nothing to do.
--
Greetings, alo.
http://www.alobbs.com
_______________________________________________
Cherokee mailing list
[email protected]
http://www.alobbs.com/cgi-bin/mailman/listinfo/cherokee
