Strange prepending of 0 to response

[Maxim Dounin]

Hello!

On Tue, Feb 01, 2011 at 10:14:19AM +0000, Maxwell Lamb wrote:

> Hi All,
> 
> Sorry that my first posting here is a support request/bug 
> report, but I've gotta start somewhere!
> 
> So - basically, we're seeing responses occasionally (not always 
> - in fact, quite infrequently) coming back with a 0 prepended to 
> the entire response - before the headers.
> 
> This means that the response received by the browser ends up 
> looking like
> 
> ************
> 0
> 
> HTTP/1.1 200 OK
> Server: nginx/0.8.53
> Date: Tue, 01 Feb 2011 09:07:27 GMT
> Content-Type: text/html; charset=utf-8
> Transfer-Encoding: chunked
> Connection: keep-alive
> Set-Cookie: blu-session-id=2baf936adc11f79aad72827d46fd31aa; path=/; HttpOnly
> Expires: Mon, 1 Jan 2001 00:00:00 GMT
> Last-Modified: Tue, 01 Feb 2011 09:07:27 GMT
> Cache-Control: no-store, no-cache, must-revalidate, max-age=0, post-check=0, pre-check=0
> Pragma: no-cache
> p3p: CP="IDC DSP COR ADM DEVi TAIi PSA PSD IVAi IVDi CONi HIS OUR IND CNT"
> Content-Encoding: gzip
> 
> and then a load of gzip'd data.
> *************
> 
> This of course breaks all sorts of behaviour. We initially 
> thought it was coming from HttpProxyModule, however after 
> reconfiguring to serve certain assets directly from disk, rather 
> than via the proxy, we saw the same issue repeating itself, 
> meaning that it's probably independent from the proxy module.
> 
> Has anyone seen this behaviour before? Any ideas on where it's 
> coming from?

This looks like duplicate last chunk ("0" CRLF CRLF) in previous 
request.  There is at least one known problem with proxy_cache 
(with proxy_cache_use_stale) which results in such behaviour.

If you use proxy_cache you may test if switching it off (or at 
least disabling proxy_cache_use_stale completely) resolves the 
problem.  Alternatively, apply patch from here:

http://nginx.org/pipermail/nginx-devel/2010-January/000102.html

Maxim Dounin