RE: [squid-users] cache_peer and acl random problem

From: Jenny Lee <bodycare_5_at_live.com>
Date: Fri, 30 Jul 2010 10:21:46 +0000

>> Thank you for all the effort you are putting in squid.
>>
>> Please consider a lean and mean version for a skinny squid. Some people have no interest in ecap, icap, xcap, adaptations, wwpwc2sp, 22s... I am lost!
>>
>
> I've propose it and been shot down a few times now.
 
Some folks like me need squid for access control of our corporate users. Users also do not like to get weird data due to caching because of problems with their broken browsers. Therefore, we don't like to cache them. We don't need 99% features of squid. I understand that the purpose of squid is to cache... But now that you include everything in it along with the kitchen sink... So why not go for a skinny one too?
 
NOTE: Hotmail apparently does not send directly to the mailing list, but instead sends to user. We have to manually add squid-users every time. More times I notice this after I press send. This paragraph is not on original mail sent to you directly. Apologies.

> At least I can point to a full disabling list. Look for test-suite/build
> tests/layer-01-minimal.opts. Thats everything that can safely be turned
> off. Though you may find some of them are wanted.
 
That is what I am looking for! Unfortunately, there is no such list or reference to it in any file in:
3.1.5
3.HEAD
 
>> --enable-useragent-log \
>> --enable-underscores \
>> --enable-referer_log \
>
> All the special log types (useragent and referer logs) can be turned off
> and replaced with a logformat entry in the config file.
I wanted to ask about that actually. I think those options are obsolete from 1.6 days and confusing now. I already do that with logformat and I was not sure what purpose they served in configure.

> underscores can be turned off. It enables a DNS violation which makes
> Squid accept and lookup invalid domains like broken_host.example.com
People want that. They say, your cache is not working (when accessing these hosts). Better to keep it on than dealing with them.

>> How to remove these cache digests?
>>
>> I removed from configure... I have "cache deny all" and I have "no digest" in parents... yet I get: "TCP_MISS:CD_SIBLING_HIT P3"
>>
>
> Looks like a bug. The code for CD_SIBLING_HIT including display is all
> wrapped in the code removal macros.
>
> Can you please check your build include/autoconf.h for
> USE_CACHE_DIGESTS and config.log for "cache-digests" and "cache_digests"
> to see if there is anything helpful showing up?
autoconf.h/config.log:
#define USE_CACHE_DIGESTS 0
Proper --disable-cache-digests is in every line in config.log.
 
Here is the configuration leading to this:
 
cache deny all
acl 33PCT random 1/3
acl TESTUSER proxy_auth TESTUSER
 
cache_peer 2.2.2.3 parent 3128 0 proxy-only no-digest no-netdb-exchange login=PASS name=P2
cache_peer 2.2.2.4 parent 3128 0 proxy-only no-digest no-netdb-exchange login=PASS name=P3
 
cache_peer_access P2 allow TESTUSER 33PCT
cache_peer_access P2 deny all
 
cache_peer_access P3 allow TESTUSER 33PCT
cache_peer_access P3 deny all
 
prefer_direct off
 
Here is what I get in logs:
 
TCP_MISS:FIRSTUP_PARENT P2
TCP_MISS:CD_SIBLING_HIT P3
TCP_MISS:FIRSTUP_PARENT P2
TCP_MISS:HIER_DIRECT 2.2.2.2
TCP_MISS:CD_SIBLING_HIT P2
TCP_MISS:FIRSTUP_PARENT P3
 
 
Is this because of an error on my configuration [Digest and Sibling parts]?
 
Behaviour is how I want. Only this log entries bothering me. Over the years, I came to greatly trust squid logs. They don't lie. If they point out an error, it is because of me... or Microsoft.
 
On my main question would the hit go direct with this config if parent2 or parent3 down?
 
Thank you!
J
                                                
Received on Fri Jul 30 2010 - 10:21:53 MDT

This archive was generated by hypermail 2.2.0 : Fri Jul 30 2010 - 12:00:04 MDT