rsr,

Sounds like you need 2 apaches, on separate physical servers - 

The front end is called thin, and does not have mod perl at all, since nothing will 
run there. It lives mainly to proxy the perl requests to the second apache (often 
called heavy) will live on an inside IP address only (eg. 192.168.x.x or as needed for 
your network). This one has your standard mod perl  stuff and just executes the 
requests it receives. It doesn't need to know anything about the fact that its 
requests are coming and going via proxy at all.

There is excellent documentation on apache.org about how to do this - I just learned 
how to do it myself a couple of weeks ago, although in my case both apaches live on 
the same machine. Basically, the secure thin server needs to support mod_rewrite and 
mod_proxy plus whatever else you need it to do.

If this meets your needs (and it seems like it will) you do not need any socket level 
stuff at all which should make you very happy. Just build your perl code to do what 
you want it to do, and make it run on the inside only server.

The this server can serve fixed html files, images, and so on, or you could proxy them 
from the other apache, or yet another apache on an inside machine if the secure 
machine is not even allowed to do that.

HTH,

GV

At 11:34 AM 1/8/2003 +0530, Chandrasekhar R S wrote:
>Dear All,
>        Let me explain my scenario.
>
>        We are having a secure OS and we plan to integrate mod_perl with the Apache
>webserver already available on the OS.  Being a secure server, it mandates
>that none of the exec's be done on the server itself, as these potentially
>can be malicious.  Hence we delegate the execution of perl scripts
>elsewhere.
>
>        Hence, should I be able to integrate mod_perl to my secure web server, I
>still would need to receive perl requests through mod_perl only, but rather
>than executing the scripts there itself, I should instruct mod_perl to
>delegate the interpretation/execution of perl scripts to elsewhere.
>
>        This being the requirement, I planned to proceed this way -
>                1. I would modify mod_perl code to forward a perl request to another
>standalone server.  This I planned to do with sockets (rather than employing
>Queues, files etc.) since, I would need persistent connections to serve
>dynamic perl scripts.
>
>                2. I should look for a server that could communicate with mod_perl.  
>This
>server would execute the perl scripts and return the response back to
>mod_perl.  This I call the standalone server.
>
>        I will probe all the suggestions you all had sent me.
>
>        I shall look for at the POE mail lists.
>
>with thanks
>rsr.
>
>-----Original Message-----
>From: Perrin Harkins [mailto:[EMAIL PROTECTED]]
>Sent: Tuesday, January 07, 2003 10:08 PM
>To: [EMAIL PROTECTED]
>Cc: [EMAIL PROTECTED]
>Subject: Re: Query
>
>
>Chandrasekhar R S wrote:
>> I am having a requirement as follows :
>>
>>       I need to execute/interpret the perl requests away from mod_perl.
>
>Can you explain why you want to do this?  Your stated requirement is
>already met by CGI, FastCGI, SpeedyCGI, and a bunch of other things, but
>we can't really recommend anything specific without more information.
>
>- Perrin

Reply via email to