Hi! Got the following "problem" with squid 3.2.7: We got several ICAP services with request and response-modification, that rewrites the request URL and modifies the response (especially the URLS in the html)on the return path. That means that a request: http://a.service.name/www.google.com/... gets translated by the request_modification to: http://www.google.com/... This then passes the proxy/cache and on the return path we translate the URLS inside the response (like /images/srpr/logo4w.png) back to http://a.service.name/www.google.com/images/srpr/logo4w.png. This works with a simple config like this: adaptation_access modify_request_a allow HTTP GETPOST adaptation_access modify_response_a allow HTTP GETPOST So far so good. Now we got several such services running - each one running a separate squid instance. We want to consolidate those into a single squid instance. So now in preparation for this second service "b.service.name" with a separate icap handler (with a similar config), I have tried to limit the icap request/response via acls to a specific URL: everything that comes in to the destinationdomain "a.service.name": acl hosts_allowed_a dstdomain a.service.name adaptation_access modify_request_a allow HTTP GETPOST hosts_allowed_a adaptation_access modify_response_a allow HTTP GETPOST hosts_allowed_a Unfortunately this does not work for the response modification. Only icap request modification takes place, but no response-modification! It seems as if I need to use a different ACL for this to work, as dstdomain is now "www.google.com"and not "a.service.name" but I have not found anything that would be of help. (I am also using adapted_http_access, so this might have an impact as well) So, how can I achieve this goal? Is there another acltype besides "dstdomain" to match on the unmodified dstdomain instead of the icap-request modified dstdomain? As a workaround: do I need to set an additional header (via header_access/header_replace or similar) and trigger on this acl (acl ... req_header <headername>) instead of modify_response_a? Other ideas? Thanks, Martin This message and the information contained herein is proprietary and confidential and subject to the Amdocs policy statement, you may review at http://www.amdocs.com/email_disclaimer.asp
