How to investigate web script not running?
 

Hello

I'm trying to run my very first FastCGI script on an Apache shared
host that relies on mod_fcgid:
==============
#!/usr/bin/python
from fcgi import WSGIServer
import cgitb

# enable debugging
cgitb.enable()

def myapp(environ, start_response):
start_response('200 OK', [('Content-Type', 'text/plain')])
return ['Hello World!\n']

WSGIServer(myapp).run()
==============

After following a tutorial, Apache complains with the following when I
call my script:
==============
Internal Server Error

The server encountered an internal error or misconfiguration and was
unable to complete your request.
==============

Generally speaking, what tools are available to investigate issues
when running a Python web app?

Thank you.

Re: How to investigate web script not running?
 

On Fri, 28 Sep 2012 13:37:36 +0200, Gilles <nospam@nospam.com> wrote:

>==============
>Internal Server Error
>
>The server encountered an internal error or misconfiguration and was
>unable to complete your request.
>==============

Looks like fcgi.py doesn't support WSGI:

Traceback (most recent call last):
File "hello.fcgi", line 2, in ?
from fcgi import WSGIServer
ImportError: cannot import name WSGIServer

Re: How to investigate web script not running?
 

On Fri, 28 Sep 2012 13:37:36 +0200, Gilles <nospam@nospam.com> wrote:

> Hello
>
> I'm trying to run my very first FastCGI script on an Apache shared
> host that relies on mod_fcgid:
> ==============
> #!/usr/bin/python
> from fcgi import WSGIServer
> import cgitb
>
> # enable debugging
> cgitb.enable()
>
> def myapp(environ, start_response):
> start_response('200 OK', [('Content-Type', 'text/plain')])
> return ['Hello World!\n']
>
> WSGIServer(myapp).run()
> ==============
>
> After following a tutorial, Apache complains with the following when I
> call my script:
> ==============
> Internal Server Error
>
> The server encountered an internal error or misconfiguration and was
> unable to complete your request.
> ==============


Do it the other way around:

# cgitb before anything else
import cgitb
cgitb.enable()

# so this error will be caught
from fcgi import WSGIServer



Regards,
Michael

Re: How to investigate web script not running?
 

On Fri, 28 Sep 2012 14:16:22 +0200, "Michael Ross" <gmx@ross.cx>
wrote:
>Do it the other way around:
>
># cgitb before anything else
>import cgitb
>cgitb.enable()
>
># so this error will be caught
> from fcgi import WSGIServer

Thanks much for the tip. The error isn't displayed when calling the
script from a web browser but it is when running the script on a shell
account.

It looks like that newer version of fcgi.py doesn't include support
for WSGI, and I need some extra (Flup?) software to sit between
mod_fcgid and a WSGI Python application.

Definitely not plug 'n play :-/

Re: How to investigate web script not running?
 

On Friday, 28 September 2012 18:45:41 UTC+5:30, Gilles wrote:
> On Fri, 28 Sep 2012 14:16:22 +0200, "Michael Ross" <gmx@ross.cx>
>
> wrote:
>
> >Do it the other way around:
>
> >
>
> ># cgitb before anything else
>
> >import cgitb
>
> >cgitb.enable()
>
> >
>
> ># so this error will be caught
>
> > from fcgi import WSGIServer
>
>
>
> Thanks much for the tip. The error isn't displayed when calling the
>
> script from a web browser but it is when running the script on a shell
>
> account.
>
>
>
> It looks like that newer version of fcgi.py doesn't include support
>
> for WSGI, and I need some extra (Flup?) software to sit between
>
> mod_fcgid and a WSGI Python application.
>
>
>
> Definitely not plug 'n play :-/

Well the plug and play standard is superseded by USB practically.

Re: How to investigate web script not running?
 

On Sat, 29 Sep 2012 10:05:25 -0700 (PDT), Ramchandra Apte
<maniandram01@gmail.com> wrote:
>> Definitely not plug 'n play :-/
>
>Well the plug and play standard is superseded by USB practically.

Indeed ;-)

Anyway, Support finally got back to me, and it turns out that they
have Flup alreay installed on shared hosts, so I just have to provide
a WSGI script. OTOH, mod_fcgid is confured to wait 5mn or so before
checking if the script was edited, so I'll have to use a test host for
development and only use the shared host for deployment.

Thank all.