[squid-users] Possible to user reply_header_add directive with acl random access list ?

Alex Rousskov rousskov at measurement-factory.com
Fri Jul 19 02:44:06 UTC 2019


On July 18, 2019 18:15:30 --Ahmad--  wrote:


> reply_header_add start "a" !markedProcessed half10000 markProcessed

...

I would replace 1/1 random ACL with "all", but OK.


> tcp_outgoing_address 12.13.200.10 half10000

Strange: Your outgoing address decisions appear to be random, completely 
independent from your Start values. Is that what you want?


> 12.13.200.13 --> B
> 12.13.200.14 --> a
> 12.13.200.12 ---> E
>
>
> And those are wrong ….
>
>
>
>
> above are wrong reply values , the correct should be as below based on the 
> Acls we configured .
>
>
>
>
>
>
> 12.13.200.13 --->D
> 12.13.200.12 ---->C
> 12.13.200.14  ---->E

I see nothing in your configuration that would tie outgoing address to 
Start values. Where did you configure Squid to use "D" for .13 or vice versa?

Alex.



>
>
>
>> On 18 Jul 2019, at 23:08, Alex Rousskov wrote:
>>
>> On 7/18/19 3:48 PM, --Ahmad-- wrote:
>>
>>> Any recommendation alex ?
>>
>> I recommend isolating the problem to the minimum number of transactions
>> (probably one or two in your case) and then posting your Squid
>> configuration, actual transaction headers, and an explanation why those
>> actual headers are wrong (and what headers you expected to see).
>>
>> Alex.
>>
>>
>>
>>>> On 17 Jul 2019, at 18:36, Alex Rousskov <rousskov at measurement-factory.com> 
>>>> wrote:
>>>>
>>>> On 7/17/19 10:40 AM, --Ahmad-- wrote:
>>>>
>>>>
>>>>> 2019/07/17 09:21:42| FATAL: Invalid ACL type ‘annotate_client'
>>>>
>>>>
>>>>> do i need to recompile squid to enable this kind of ACLS ?
>>>>
>>>> These ACLs are only supported in the development version of Squid
>>>> (future v5): https://github.com/squid-cache/squid/commit/63e82d8
>>>>
>>>> Alex.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squid-cache.org/pipermail/squid-users/attachments/20190718/208894e6/attachment.html>


More information about the squid-users mailing list