Re: Interesting behavior observed in the missed bytes ratio in relation with missed requests ratio

[Lucas Brasilino]

Hi:


I noticed when I analysed Squid's access.log using
calamaris that a miss ratio of 54,41% of the requests
made generated 81,92% of miss ratio of the total
requested bytes. This means that the proxy server
isn't 
reducing the bandwidth usage effectively... What is

the most probable reason for it?
 


It's off-topic, but are you using LRU ?
You should try using LFU-DA or GDSF. Maybe
GDSF is a better choice for you since it
relies over objects size, and your byte
hit rate isn't good.

--
[]'s

Lucas Brasilino
[EMAIL PROTECTED]
Procuradoria da República no Estado de Roraima
+55-95-6239642

Question about suid's default behaviour.

[Ghislain Garcon]

Hello,

I've tested squid 2.5-12 without http_acces acl and all pages are
blocked. Then I created two deny http_access acls which don't match with
my  http requests and all pages are unfiltered.

For me it is strange. I've looked in source code and found :

debug(28, 3) (aclCheck: NO match found, returning %d\n, allow !=
ACCESS_DENIED ? ACCESS_DENIED : ACCESS_ALLOWED);
aclCheckCallback(checklist, allow != ACCESS_DENIED ? ACCESS_DENIED :
ACCESS_ALLOWED);

at the end of aclCheck(aclCheck_t * checklist) in file acl.c. Is this
swap really needed for something else?

In fact I try to add external acl support in icap patch. The acl
structures are not exactly the same in Icap and in squid, then I have to
call aclNBCheck with only one acl. Then I had aclCheck: NO match found,
returning 1 as a result of aclCheck.

Ghislain Garçon

Re: Question about suid's default behaviour.

[Henrik Nordstrom]

On Wed, 16 Nov 2005, Ghislain Garcon wrote:


at the end of aclCheck(aclCheck_t * checklist) in file acl.c.


Yes..


Is this swap really needed for something else?


It's just how it works.


From the documentation on http_access:


If none of the access lines cause a match, the default is the
opposite of the last line in the list.  If the last line was
deny, the default is allow.  Conversely, if the last line
is allow, the default will be deny.  For these reasons, it is a
good idea to have an deny all or allow all entry at the end
of your access lists to avoid potential confusion.

Regards
Henrik

Re: exceptions.AttributeError: str object has no attribute group

[Eric]

Also had to apply the following patch to icap_handler.py to fix this
bug in handling encapsulated requests...

139,141c139,147
   try:
   request.encap_reqs= encaps.group('req')
   except:
---
   if encaps:
   try:
   request.encap_reqs= 
 encaps.group('req')
   except:
   request.encap_reqs = None
   request.encap_res= encaps.group('res')
   request.encap_body= encaps.group('body')
   request.res_body =  
   else:
143,145c149,150
   request.encap_res= encaps.group('res')
   request.encap_body= encaps.group('body')
   request.res_body =  
---
   request.encap_res = None
   request.encap_body= None
[EMAIL PROTECTED] /usr/local/icap_server diff -u icap_handler.py.orig 
icap_handler.py
--- icap_handler.py.origSun Jul  7 20:21:46 2002
+++ icap_handler.py Wed Nov 16 12:36:58 2005
@@ -136,13 +136,18 @@
elif request.command == 'respmod':
#print DBG: respmod request is : 
,request.header
encaps = get_header_match 
(ENCAPSULATED_respmod,request.header)
-   try:
-   request.encap_reqs= encaps.group('req')
-   except:
+   if encaps:
+   try:
+   request.encap_reqs= 
encaps.group('req')
+   except:
+   request.encap_reqs = None
+   request.encap_res= encaps.group('res')
+   request.encap_body= encaps.group('body')
+   request.res_body =  
+   else:
request.encap_reqs = None
-   request.encap_res= encaps.group('res')
-   request.encap_body= encaps.group('body')
-   request.res_body =  
+   request.encap_res = None
+   request.encap_body= None
else:
print Currently ignoring command 
,request.command
# Collect all the following data

On Sun, Oct 23, 2005 at 07:41:11PM -0500, Eric wrote:
 Great, I set up the python icap server and wrote a basically empty
 proxylet since it automatically updates Content-Length if you modify
 the body at all...guess it would work with any ICAP server.  Works great.
 
 Thanks,
 eric.

I'm interested in adding gzip support in squd-2.5,does it make sense?

[ro vencentro]

I'm interested  in adding gzip support in squd-2.5,does it make sense?