TCP_REFRESH_MODIFIED

classic Classic list List threaded Threaded
32 messages Options
12
Reply | Threaded
Open this post in threaded view
|

TCP_REFRESH_MODIFIED

joseph
squid v4.02  
amos is there a way  to stop this mess please or i can say its bug that need to be more attention to it
TCP_REFRESH_MODIFIED/200 131702    same file same size  no mather wat i do sometime hit  some time TCP_REFRESH_MODIFIED and don't tell me pls its expired or what ever my refresh pattern ar correct and everything fine just every time i get same file or hit  or TCP_REFRESH_MODIFIED  only me on system fresh install fresh hard-rive fresh debian install i think  you guys making proxy cache as standard browser so what is the benefit of that proxy cache suppos to violate rules and cache stuff and suppos to do what we tell it to do cache object and stay until the time expired in my setting please lots emportant bug need to be fixed befor adding more futur to squid i dont see any one focus on important bug just on newer stuff and i dont think those there is somthing delete thim or change there status unless the code that shuld keep it in cache not looking at the conf bit of the file
why its modifyed and the file same size same expiration date shuld be modifyed


as i understand  squid look at
file size
date expiery
+++ if the income header of the file never change why would squid see it as TCP_REFRESH_MODIFIED
mmmmmm


soryy guys but the more good stuff you add and i admit  squid become faster and light
but   other code get worth    we are not browser here that why its called proxy cache
other cache then squid become advanced stealing idea from squid and they do better
we suppose to be ahead of them  
again pls dont take it  personally i do love you guys you help us free thank you allot
but instead of doing better those bug killing us pushing us backward
its up to us to decide witch control to allow not the regulation of the rfc? or googles?
again we are as squid cache fans and user tester   not as browser
sorry again i may b a bit upset but  
************************** ***** Crash to the future **** **************************
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

Amos Jeffries
Administrator
Hi jo,
  Pause. Breath.

If I'm reading that long sentence right you are getting REFRESH_MODIFIED
responses from servers with "new" content to be cached, which has not
actually changed.

What URL or URLs ?

>
> as i understand  squid look at
> file size
> date expiery
> +++ if the income header of the file never change why would squid see it as
> TCP_REFRESH_MODIFIED
> mmmmmm

There are many things that might make a cached object need revalidation.
Expires timestamp is just one of them. Having authentication credentials
embeded in it is another. Private cookies attached, client happened to
press force-reload button in their browser, etc. etc.

Unfortunately in this case the server is responding with an entirely new
object each time. Squid can't exactly prevent that happening. It is the
server choosing what to send.


If it is happening with a specific URL, or a few URLs the tool at
redbot.org can say whether there is anything broken with the server
responses that are causing problems.

Otherwise a bit more detail about what your particular setup is will be
needed to figure it out. What the proxy is intended to do / what its
used for, how its connected to from clients, and the squid.conf.


If you had any questions you want answers to in the rest of that text,
please re-phrase them a bit more clearly and I will try to answer as
best I can.

Amos

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

joseph
hi amos  
this is the header
Connection: keep-alive

HTTP/1.1 200 OK
Last-Modified: Mon, 02 Nov 2015 08:13:05 GMT
Content-Type: video/mp4
Date: Tue, 03 Nov 2015 18:25:03 GMT
Expires: Tue, 03 Nov 2015 18:25:03 GMT
Cache-Control: private, max-age=21297
Accept-Ranges: bytes
Content-Length: 131072
access-control-allow-origin: #######################
Access-Control-Allow-Credentials: true
Timing-Allow-Origin: #############################
Access-Control-Expose-Headers: Client-Protocol, Content-Length, Content-Type
x-content-type-options: nosniff
Server: gvs 1.0
Age: 5297  
sorry for not providing the url  no need i gess but
as you see
Content-Length: 131072  same  and all other info in header same never change
i the only person on that cache no other clients  its my test bansh
i dont hit re load at all  just passing the link click go  after 30 or more minut
somtime less right after  just click refresh botton no ctl F5  
age ar ok shuld stay in cache for that amount lets say
Cache-Control: private, max-age=21297    i had in refresh pattern
367840  99% 535600 override-expire override-lastmod ignore-reload ignore-no-store ignore-private reload-into-ims store-stale
and it cache it np with that  but why it look like care without brake lol  TCP_REFRESH_MODIFIED
or swap fail
but mostly MODIFIED  and its not
unless override-expire override-lastmod ignore-reload ignore-no-store ignore-private  bit ar unset some how
************************** ***** Crash to the future **** **************************
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

Yuri Voinov

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
 
Heh, this porn :) URL has Cache-Control: private.

And without special things will not be cached.

04.11.15 1:54, joe пишет:

> hi amos  
> this is the header
> Connection: keep-alive
>
> HTTP/1.1 200 OK
> Last-Modified: Mon, 02 Nov 2015 08:13:05 GMT
> Content-Type: video/mp4
> Date: Tue, 03 Nov 2015 18:25:03 GMT
> Expires: Tue, 03 Nov 2015 18:25:03 GMT
> Cache-Control: private, max-age=21297
> Accept-Ranges: bytes
> Content-Length: 131072
> access-control-allow-origin: #######################
> Access-Control-Allow-Credentials: true
> Timing-Allow-Origin: #############################
> Access-Control-Expose-Headers: Client-Protocol, Content-Length,
Content-Type
> x-content-type-options: nosniff
> Server: gvs 1.0
> Age: 5297  
> sorry for not providing the url  no need i gess but
> as you see
> Content-Length: 131072  same  and all other info in header same never
change
> i the only person on that cache no other clients  its my test bansh
> i dont hit re load at all  just passing the link click go  after 30 or
more

> minut
> somtime less right after  just click refresh botton no ctl F5
> age ar ok shuld stay in cache for that amount lets say
> Cache-Control: private, max-age=21297    i had in refresh pattern
> 367840  99% 535600 override-expire override-lastmod ignore-reload
> ignore-no-store ignore-private reload-into-ims store-stale
> and it cache it np with that  but why it look like care without brake lol
> TCP_REFRESH_MODIFIED
> or swap fail
> but mostly MODIFIED  and its not
> unless override-expire override-lastmod ignore-reload ignore-no-store
> ignore-private  bit ar unset some how
>
>
>
> --
> View this message in context:
http://squid-web-proxy-cache.1019090.n4.nabble.com/TCP-REFRESH-MODIFIED-tp4674325p4674327.html
> Sent from the Squid - Users mailing list archive at Nabble.com.
> _______________________________________________
> squid-users mailing list
> [hidden email]
> http://lists.squid-cache.org/listinfo/squid-users

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
 
iQEcBAEBCAAGBQJWORTgAAoJENNXIZxhPexGIS4H/2LaTSsunfhHJJ2nyUEBYEOL
Vu/T65tDeFTII1hxoyv/ZD0Alffv11ZEcqFOnK5BcqeBZ9k+CbPft+U/lFqGE8zG
OGEpLwIPlLhAbpO2HIXtKeWNoK8MrVOAByYs3ElUdVAEIKMw2AeGlBI3QeKwQYKI
G+J8w6G5d+vMsW5p9Qi9nAT5o43eYQAiAO7lmmz6On2UCMiMt97AODaZ3i4ULuIK
Zskx46S3xrZphkGLfROvcdP4xclzhq2Qmn1/16E6eurXWOLI07pZLwApcj5UhstE
W+OJbOkDZamp0LcwR6JNBARFJ8mLo1R/J4djIl5m2uFFp47Qtolc68rQltLk8wA=
=5OJd
-----END PGP SIGNATURE-----

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

joseph
Yuri Voinov  lol  no its not porn  its some url ppl ar daying to get it cached dont worry about it it dose cache as i sayd befor please read ^    im not worry about it im worry about why it dose not stay in cache as i explaned befor ^
************************** ***** Crash to the future **** **************************
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

Yuri Voinov

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
 
Because of.

Take a look onto Cache-Control header once more.

It you haven't refresh_pattern with "ignore-private" for this URL and
similar, it always not be cached.

If this URL is dynamic, it also always not be cached. Without special
solutions.


04.11.15 2:01, joe пишет:
> Yuri Voinov  lol  no its not porn  its some url ppl ar daying to get it
> cached dont worry about it it dose cache as i sayd befor please read
^    im
> not worry about it im worry about why it dose not stay in cache as i
> explaned befor ^
>
>
>
> --
> View this message in context:
http://squid-web-proxy-cache.1019090.n4.nabble.com/TCP-REFRESH-MODIFIED-tp4674325p4674329.html
> Sent from the Squid - Users mailing list archive at Nabble.com.
> _______________________________________________
> squid-users mailing list
> [hidden email]
> http://lists.squid-cache.org/listinfo/squid-users

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
 
iQEcBAEBCAAGBQJWORZlAAoJENNXIZxhPexG0e8H/3YmfE42dVz1uBB2zwZDMbqh
okVLzAWjNET+uvghcx9uhq2SWSzzRh/8270QnP1ilHjaWhkzvtygGC/oPNQeF40b
L4RyrPPU9ifIw/N43EkyUmEWsVD73qz2lE/F4AQ6j59vl0nFX7fhNz+bqIpvCZAQ
i++Ly6yZfHOyS6S2PxrtvnUjtUfxRhPNXcTQHAZqsZSLF+fFWr1xWa8ntFhDsi9I
nLkvjJBAThVgyr2vKZMuF++e9B/iVSv0CghdkARlEVXHW1MVY3SQzgBhHr3cUBS+
VpoSI0dRevV7PnUsNKxgvHXiR0yOc81v9s3aXSsFobUsztgY2bcZXZYNLGw8+OE=
=day4
-----END PGP SIGNATURE-----

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

joseph
>Take a look onto Cache-Control header once more.

>It you haven't refresh_pattern with "ignore-private" for this URL and
>similar, it always not be cached.

if this ignore-private useless then one more bug squid has is this wat you telling me

>If this URL is dynamic, it also always not be cached. Without special
>solutions.
dynamic Without special solutions right that why i do my store-url for this purpose lol
as i says before  im not complaining about caching the object if its dinamic or static please ree read ^
i do cache them lol im not beginner  many language i understand and i can tell if there is a problem in source or not with wat ever i provide on top if there is any mistake o dont see please let me know but don't push me with something funny tks    
************************** ***** Crash to the future **** **************************
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

Yuri Voinov

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
 


04.11.15 2:14, joe пишет:
>> Take a look onto Cache-Control header once more.
>
>> It you haven't refresh_pattern with "ignore-private" for this URL and
>> similar, it always not be cached.
>
> if this ignore-private useless then one more bug squid has is this wat you
> telling me
You give too few info to solve your problem.
>
>
>> If this URL is dynamic, it also always not be cached. Without special
>> solutions.
> dynamic Without special solutions right that why i do my store-url for
this
> purpose lol
> as i says before  im not complaining about caching the object if its
dinamic
> or static please ree read ^
> i do cache them lol im not beginner  many language i understand and i can
English is not in your list, of course.
>
> tell if there is a problem in source or not with wat ever i provide on top
> if there is any mistake o dont see please let me know but don't push
me with
> something funny tks    
You don't give meaningful info and asking about solution. And tell us
about bug.

We do not see with whom we are dealing. And their skills based on a pair
of illiterate letters is difficult to determine. That's why we ask you
to funny things.

U r welcome, dude.

>
>
>
>
> --
> View this message in context:
http://squid-web-proxy-cache.1019090.n4.nabble.com/TCP-REFRESH-MODIFIED-tp4674325p4674331.html
> Sent from the Squid - Users mailing list archive at Nabble.com.
> _______________________________________________
> squid-users mailing list
> [hidden email]
> http://lists.squid-cache.org/listinfo/squid-users

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
 
iQEcBAEBCAAGBQJWORwTAAoJENNXIZxhPexGV+cH/iA6Hui/QV89qif5p3fyoJ5c
JM0+42EAJRE4i+dKFyP/pvLurfqxUTpVGGi4ajY+4vFd/z2m2JkSozMXYk+rdaHP
kg4uggnGs9urX/bEqOCXW1YyUgntqOicOOcVaiw1QLBlG1yRq4vOix9/fjPs4Erf
r65C4jkN70D5roaHlqxe5IsNikLJXJneFzGGj2YUPzPg2/AePMYZjPPie/ZUBnUU
a468wuW3IrDEAlTottgXdJvfP65OeIe3XEOhOTy7XttQLqmUehQaWNgMBLKu889u
8EFCQRZaWhrXK6Lme+n2azlohuo0tTWLf4XPMwXGww7CgaY9XslQ45HKq4aGDiI=
=P9mT
-----END PGP SIGNATURE-----

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

Yuri Voinov
In reply to this post by joseph

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
 
To talk was objective: show the config (squid.conf) and URL, which
causes problems with caching. Telepaths on holiday in Bali. Personally,
I have no idea either looks like configuration or a URL you're trying to
cache. In such a situation it is impossible to say anything definite.

04.11.15 2:14, joe пишет:

>> Take a look onto Cache-Control header once more.
>
>> It you haven't refresh_pattern with "ignore-private" for this URL and
>> similar, it always not be cached.
>
> if this ignore-private useless then one more bug squid has is this wat you
> telling me
>
>> If this URL is dynamic, it also always not be cached. Without special
>> solutions.
> dynamic Without special solutions right that why i do my store-url for
this
> purpose lol
> as i says before  im not complaining about caching the object if its
dinamic
> or static please ree read ^
> i do cache them lol im not beginner  many language i understand and i can
> tell if there is a problem in source or not with wat ever i provide on top
> if there is any mistake o dont see please let me know but don't push
me with
> something funny tks    
>
>
>
> --
> View this message in context:
http://squid-web-proxy-cache.1019090.n4.nabble.com/TCP-REFRESH-MODIFIED-tp4674325p4674331.html
> Sent from the Squid - Users mailing list archive at Nabble.com.
> _______________________________________________
> squid-users mailing list
> [hidden email]
> http://lists.squid-cache.org/listinfo/squid-users

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
 
iQEcBAEBCAAGBQJWOR7TAAoJENNXIZxhPexGnCoH/1bAL6dG6FAxlx2KgFlW9W8a
vf9Ka29e/gAjufpo1el4MLRzILW20FBFQeTM8Yi/54b7HdYVkfakMg2fZIT+Diww
HOGkz8Tu9OvIYRjHmGkTEg5Sw5SkV1a5/3mnyG7zm5gnLik8FtyTmFatIrGVOpfI
PZ0vicAQpLPLYbdTPVHqbB6sszo0OvSt3lgzeoRzTX2WJ/2cB6qWBgsVRzr8KxJj
oD8A7PQq+32jzBvQvYm72qloFJ2F9LXcFOVXFcKCJ92EdKbtXs/G13cPB18LQdRE
GSmBT2EDKgb86r50LVKsRh/JLsnfnNG9/IICbh0rwMd5Nz7yKJGWSGsNvjcs3rg=
=mzrN
-----END PGP SIGNATURE-----

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

joseph
In reply to this post by Yuri Voinov

>You give too few info to solve your problem.
well   i give a good info bro this problem its all over the internet lots of ppl ar complaning about it

>English is not in your list, of course.
right dude im lebanise but i use to live in canada and in US my english is mixed hehe canadian US lol
sorry for that but its ok i gess

>You don't give meaningful info and asking about solution. And tell us
>about bug.
as u read on top any programmer can understand the issue i provide ar enough and im not the only one having this problem but it happen all my test i did so i start belive those ppl that is bug or!!!! :)
any way im going to study the source a bit and see wat i can get with  

>We do not see with whom we are dealing. And their skills based on a pair
>of illiterate letters is difficult to determine. That's why we ask you
>to funny things.
tks  but sorry to say you might be right for that but if i wana have fun i can find me a funny url and have fun there lol
>U r welcome, dude.
thank again bro :)
************************** ***** Crash to the future **** **************************
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

Amos Jeffries
Administrator
In reply to this post by Yuri Voinov
On 4/11/2015 9:53 a.m., Yuri Voinov wrote:
>
> To talk was objective: show the config (squid.conf) and URL, which
> causes problems with caching. Telepaths on holiday in Bali. Personally,
> I have no idea either looks like configuration or a URL you're trying to
> cache. In such a situation it is impossible to say anything definite.
>

It is an object apparently containing per-user information
(Cache-Control:private).

Note that Expires and max-age=N headers are meaningless on private
objects. A private object needs revalidating on every single re-use.


It is being served up from GoogleVideoServer (aka "gvs") so is either
YouTube content (they go to some extreme lengths to prevent caches
working), or possibly a video on someones personal account, or maybe
with DRM embeded. The no-change on size could be an artifact of jo being
the only one fetching it. Or the object may be padded by the server so
its always a fixed size of crypted blocks to prevent DRM decryption
attempts.


jo is using "ignore-private" configuration option to make Squid cache
the object. But that can only make it be stored. It cannot magically
make the stored content useful. The revalidation which does make it
useful still has to happen to ensure the right per-user headers are
attached when it is delivered to the client.
 The server then barfs out a whole new copy in full when asked for
revalidation details.


Facing servers like this occasionally is part of the price we
unfortunately pay for caching Cache-Control:private things without
causing all sorts of far worse nasty side effects.


jo; you could remove that "ignore-private" configuration option. That
would not affect these particular objects bandwidth usage, but would
free up the cache space they are currently using for caching other
things with possibly better savings.


Amos
_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

joseph
at least you pay attention on "gvs" :) +1

lets forget about youtube    :)  im just asking why TCP_REFRESH_MODIFIED if i don't or did not force reload
ignore-privet its working but ignore-reload its not .. suppose to prevent TCP_REFRESH_MODIFIED from hapening right ?


and talking about private control its used for public control now not just privet content
************************** ***** Crash to the future **** **************************
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

HackXBack
what joe is going to tell us is that his HIT ratio decrease and he is seeing TCP_REFRESH_MODIFIED  instead of tcp_hit when he used V4
this problem is right also with tcp swalfail miss
with V3.4 these strange problems is not exists ..
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

Amos Jeffries
Administrator
In reply to this post by joseph
On 4/11/2015 11:38 a.m., joe wrote:
> at least you pay attention on "gvs" :) +1
>
> lets forget about youtube    :)  im just asking why TCP_REFRESH_MODIFIED if
> i don't or did not force reload
> ignore-privet its working but ignore-reload its not .. suppose to prevent
> TCP_REFRESH_MODIFIED from hapening right ?

No, ignore-reload means ignore the Ctrl+F5 / reload button headers
received from clients (if any). It has nothing to do with the server
response details.

>
> and talking about private control its used for public control now not just
> privet content
>

The design model for Squid caching is "shared cache". Which means
strictly following RFCs Cache-Control:private content is not allowed to
be stored. Doing so in a shared cache design causes big problems.

We get around that danger in Squid by doing the revalidations. For most
traffic it should work find and reduce bandwidth. For some it does not.
But the worst-case there is no different from the proxy having done what
was mandatory to do in the first place.

Amos

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

Amos Jeffries
Administrator
In reply to this post by HackXBack
On 4/11/2015 11:53 a.m., HackXBack wrote:
> what joe is going to tell us is that his HIT ratio decrease and he is seeing
> TCP_REFRESH_MODIFIED  instead of tcp_hit when he used V4
> this problem is right also with tcp swalfail miss

I don't think the two are the same at all.

REFRESH is (in jo's case) an indicator that the private content is being
checked before use. If the server behaves itself the answer would be
UNMODIFIED/304 not MODIFIED/200 status, and the transfer size under 1KB.

SWAPFAIL is errors loading the on-disk file where the object ws stored
in the cache. Unless you want to serve random bytes out to the client
that failure will always have a MISS/200 or DENIED/500 result.
 In your case the bug is that you are having the disk I/O failure at
all. jo is not.


> with V3.4 these strange problems is not exists ..
>

3.4 does not cache several of the types of objects that 3.5+ do.

We are still in the process of converting Squid from HTTP/1.0 behaviour
to HTTP/1.1 behaviour. It is going slowly, one step at a time.

Amos

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

HackXBack
In reply to this post by Amos Jeffries
and how we can cache Control:private content ?
must be a choice ?
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

joseph
In reply to this post by Amos Jeffries
>I don't think the two are the same at all.
right they ar 2 diferent problem and they ar very bad for production to be on

>REFRESH is (in jo's case) an indicator that the private content is being
>checked before use. If the server behaves itself the answer would be
>UNMODIFIED/304 not MODIFIED/200 status, and the transfer size under 1KB.
i test almost all the rev...  from 4.02 down  here is wat i found  squid become as browser that all it dose not save object as i show the header on top of the topic  so min max ar use less and overide reload ar just a extra word to type  as i sayd   squid supost to be cache server not browser folowing the rools of google or the rfc is bad it depend to us admin to control witsh object need to be stored and how long
so that for the REFRESH

>SWAPFAIL is errors loading the on-disk file where the object ws stored
>in the cache. Unless you want to serve random bytes out to the client
>that failure will always have a MISS/200 or DENIED/500 result.
> In your case the bug is that you are having the disk I/O failure at
>all. jo is not.
 the swap fail   another very very bad bug  it happen that yesterday i got almost evry object scrooling one after the other with swap fail

so im thinking to go back to   v2.7 that provide beter saving for me tks
************************** ***** Crash to the future **** **************************
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

Amos Jeffries
Administrator
In reply to this post by HackXBack
On 4/11/2015 11:35 p.m., HackXBack wrote:
> and how we can cache Control:private content ?
> must be a choice ?

Yes. By adding the ignore-private refresh_pattern control.

Though be aware it still does very bad things to data in most Squid 3.x
versions for some configs. It is only fully safe in v4+.

Amos

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

Yuri Voinov

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
 


04.11.15 17:05, Amos Jeffries пишет:
> On 4/11/2015 11:35 p.m., HackXBack wrote:
>> and how we can cache Control:private content ?
>> must be a choice ?
>
> Yes. By adding the ignore-private refresh_pattern control.
>
> Though be aware it still does very bad things to data in most Squid 3.x
> versions for some configs. It is only fully safe in v4+.
4+ too cautious and dull does not cache that makes him even the
slightest doubt. And, moreover, does not allow the administrator to give
yourself direct instructions.

Which leads to an unprecedented reduction coefficient caching.
>
>
> Amos
>
> _______________________________________________
> squid-users mailing list
> [hidden email]
> http://lists.squid-cache.org/listinfo/squid-users

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
 
iQEcBAEBCAAGBQJWOelCAAoJENNXIZxhPexGTzQH/1LU6oEskIzTuek3YIWJ9HNj
i8Eyh/p72mqvPGsV1woKoeEUlihN3B/rKqIrB+0g+N0sJNh1yIYgQ5HfWLxAZVoS
tHXJHfN4+xMrxWMcVi5D70lvssi72n77OuSIdcfHPLfroM6yHQpBOoQbOtg+GV6F
S/M6TI5Pa/nm7DPVL89rpGOe0EKkoCD/rSSjxyCoVncXWH+EN3q2UFXL5TWCdOzz
+3feS4SJfC/bo3YBdu61AGRM/BPNiiidx3wKGcedVMKnPHQXtd9BLBWZQ77Ed63W
inU47fsYNdtgXbEMb+D9kY/QXdGsue0qksrXWJuplk7de+7u49fkEpY4UHFLl7E=
=zidE
-----END PGP SIGNATURE-----

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: TCP_REFRESH_MODIFIED

HackXBack
You are right Yuri,
its like a proxy bypassed system ..
12