No, I'm not caching the user account page. I only cache front page.
And usually the account page works fine for everyone. Only under heavy
stress does it run into problem, randomly.
We wrote a program which always fakes the same session id to test the
server. But no matter how many threads do we run, it wouldn't happen
in intranet. Only when I run this test program here in china, and
request the server in USA, would it happen.
I can get the configuration file later. Thanks for your help.

Sent from my iPhone

在 2010-1-21，17:00，nginx-request@nginx.org 写到：
>
> Message: 2
> Date: Wed, 20 Jan 2010 16:53:05 +0100
> From: "Piotr Sikora" <piotr.sikora@frickle.com>
> To: <nginx@nginx.org>
> Subject: Re: Messed Response Stream
> Message-ID: <F32D30C73D7B4186855F19A9FCAB3237@nightmare>
> Content-Type: text/plain; format=flowed; charset="iso-8859-1";
> reply-type=original
>
>> 3. Access a page that is different for different users, the issues
>> would
>> happen randomly.
>> (...)
>> I'm sure it's not because of proxy cache
>
> Are you saying that you're using proxy cache to cache pages that are
> different for every user? Can you show your configuration? I'm
> pretty sure
> that bad configuration is source of your problems.
>
> Best regards,
> Piotr Sikora < piotr.sikora@frickle.com >

_______________________________________________
nginx mailing list
nginx@nginx.org
http://nginx.org/mailman/listinfo/nginx