Re: Question regarding url_param hashing

From: Willy Tarreau <w#1wt.eu>
Date: Tue, 20 Apr 2010 23:03:57 +0200


On Tue, Apr 20, 2010 at 04:56:01PM +0200, Dirk Taggesell wrote:
> Many thanks, Willy for answering my questions.
>
> > But in my opinion, when you start playing
> > with 10 to 100 times more URLs than you have servers, I really
> > don't see how you could be "lucky" enough to always reach the same
> > servers.
>
> We will test the consistent hashing again once we have production data
> (it is not in production yet). We tested only with test data (which
> are still some hundred thousand urls).

well, with that many URLs, I can't even imagine how you can observe that behaviour ! It's more than enough. Feel free to forward me your config, in case I spot anything suspect.

> > At least try to increase all weights (eg: 100). There should not be
> > any need for that but it can help.
>
> Tried that, but to no avail. I added a weight of 100 and later 256 to
> each back-end, but the distribution didn't change much.
>
> > Please, try to get an extract of your prod
> > servers' logs and try to replay them through your test machine.
>
> Will do that once we have them.
> Thanks for now Willy, I will keep you informed if we make further experiences.

OK, waiting for your feedback then.

Regards,
Willy Received on 2010/04/20 23:03

This archive was generated by hypermail 2.2.0 : 2010/04/20 23:15 CEST