How To Enable URL Filtering With Just Squid & C-ICAP

Question

Before I posted this question, I read several set up guides online, the squid docs, c-icap docus, and also the ICAP RFC (3507). Unfortunately, the documentation for SQUID and C-ICAP does not provide much in the way of examples when it comes to URL filtering, only when it comes to the simple set up of using ClamAv.

So, I'm trying to figure out how to set up simple URL filtering using Squid and the C-ICAP Server. Specifically, I'm starting out trying to use SquidGuard Databases (but not the Squidguard program) as a source of black lists. I have Squid set up and running (proxying) just fine, and I have C-ICAP set up and running (responding to options request) just fine. However, I have no clue why the URL filtering is not running, as I've followed the examples as closely as possible.

My squid.conf is set up as follows for ICAP:

#Icap Options
icap_enable on
icap_service svcBlocker reqmod_precache icap://127.0.0.1:1344/srv_url_check bypass=off

At the bottom of the c-icap.conf file, I have included the icap service I want to use:

# End module: ldap_module
#URL Check Services
Include srv_url_check.conf

I've set up the srv_url_check.conf file, following the c-icap documentation:

# Default:
#None set
#Example: url_check.LoadSquidGuardDB audio-video /home/dranfu/Downloads/blacklists/audio-video/

When I send an options request to C-ICAP, it works fine:

ICAP server:localhost, ip:127.0.0.1, port:1344

OPTIONS:
    Allow 204: Yes
    Preview: 1024
    Keep alive: Yes

ICAP HEADERS:
    ICAP/1.0 200 OK:
    Methods:RESPMOD, REQMOD
    Service:C-ICAP/0.2.4 server - Echo demo service
    ISTag:CI0001-XXXXXXXXX
    Transfer-Preview:*
    Options-TTL:3600
    Date:Tue, 26 Feb 2013 10:57:13 GMT
    Preview:1024
    Allow:204
    X-Include:X-Authenticated-User, X-Authenticated-Groups
    Encapsulated:null-body=0

But when I send a standard web request to a site that is on the blacklist, I get a fairly empty response:

/usr/local/c-icap/bin/c-icap-client -p 8080 -req http://mp3.com.au 

ICAP server:localhost, ip:127.0.0.1, port:8080

And furthermore, I am still able to proxy to the website with no problem. So, if anyone has any experience with setting up squid with pure C-ICAP using URL filtering, it would be awesome to have any assistance or basic guide on how to get this working. The documentation is just too sparse, for me at least, to figure out what is going wrong. No doubt it's something I'm missing, though.

Solution

I'm trying to do the same. I got success configuring c-icap to block a request.

Your srv_url_check.conf seems to be incomplete. Mine is configured as follows:

Service urlcheck /usr/lib/x86_64-linux-gnu/c_icap/srv_url_check.so
url_check.LookupTableDB denyhosts url hash:/etc/c-icap/denyhosts.txt "Denied Host"
url_check.Profile denyProfile block denyhosts
url_check.ProfileAccess denyProfile all

The file denyhosts.txt, is a simple text file. Each line should contain a host to be block, such as:

mp3.com.au
xvideos.com
sex.com

And finally, you should uncomment line acl all src 0.0.0.0/0.0.0.0 into c-icap.conf.

Start your c-icap server like /usr/bin/c-icap -D -N -d 1 (adjust the log level (-d) as you wish) and test it using /usr/bin/c-icap-client -s url_check -req http://sex.com -v -d 1.

As a response, you will receive:

ICAP HEADERS:
    ICAP/1.0 200 OK
    Server: C-ICAP/0.4.2
    Connection: keep-alive
    ISTag: CI0001-XXXXXXXXX
    X-ICAP-Profile: denyProfile
    X-Attribute: denyhosts
    X-Attribute-Prefix: 7
    X-Response-Info: BLOCKED
    X-Response-Desc: URL category denyhosts is BLOCKED
    Encapsulated: res-hdr=0, res-body=108

RESPMOD HEADERS:
    HTTP/1.0 403 Forbidden
    Server: C-ICAP
    Content-Type: text/html
    Connection: close
    Content-Language: en

This is what I did so far...

Comments

[Admin]

I expected this may be hard to answer, but I'm researching, and once I figure it out...or someone informs me, I'll post back here so someone can make use of it.

[Scott Pack]

Welcome to Server Fault! Whilst this may theoretically answer the question, it would be preferable to include the essential parts of the answer here, and provide the link for reference.