Not easy replacement of Apache with Ngnix due to error log

[Jonathan Matthews]

On 3 April 2012 06:47, niraj <nginx-forum at nginx.us> wrote:
> Hi  Anton
>
>               As u suggested,

It's "you", not "u". We're not teenagers swapping text messages here.

>               words "upstream closed prematurely FastCGI stdout" show,
> that problem is
> probably in FastCGI server.
>
>                  This is what I am telling u that Nginx error log
> misguide us.

No, it told you exactly where to aim your troubleshooting lens. At the
FCGI process.

>              Actually Problem was one perl module was not installed in
> that so I was getting the above error.
>
>     Earlier I used apache so if any module related error then it was
> reflect in apache's error log ..............

That's because of an architectural decision for Apache to host the
perl process, and for nginx to outsource it. Each has their advantages
and disadvantages and both are valid approaches. Learn to use the
tools in your toolkit.

> that actually not happening with Nginx , this create a problem for
> system admin.

If your system administrator is unable to configure a debug log from
your perl process or its FCGI wrapper, after being given the earlier
information that the problem is *inside* the FCGI process, then your
problem is not with nginx: it is with your incompetent system
administrator.

> If any idea to get details error logs for nginx  then please sugeest.
>
> how to set  fcgiwraper error log ??????

I don't know if this thread is useful: http://www.ruby-forum.com/topic/180064
You may have to search for other people's solutions to getting the
verbose logs into nginx logs. It appears people have achieved this. I
don't have any pertinent information other than what I found searching
for the problem.

Jonathan
-- 
Jonathan Matthews
London, Oxford, UK
http://www.jpluscplusm.com/contact.html