ching or
not, to my understanding it shouldn't affect things, but just wanting
some clarification.
Thanks in advance,
Adrian Miller
On 7 July 2012 19:02, Amos Jeffries wrote:
> On 7/07/2012 6:10 p.m., Adrian Miller wrote:
>>
>> Squid via Network Wireless Router & Wireless Clients
>>
>> Hopefully this is a simple question, with an equally quick answer.
>>
>> I have set up traditiona
Squid via Network Wireless Router & Wireless Clients
Hopefully this is a simple question, with an equally quick answer.
I have set up traditional squid setups before, with the 2 NIC setup.
This time though i have only a wireless router connected via ethernet
to the squid box (1 NIC only).
All c
texts are considered equivalent by varnish and
it is NOT invalidating the other variants.
Perhaps something worth doing for squid too? I mean, it's obviously not
a squid bug but it will be consistent to the way developer tools
(firebug, chrome dev tools etc) are treating this header.
Adrian
was repeated for the anonymous users
(", Accept-encoding" added at the end of the list). Developer tools in
firefox and chrome were silently correcting this and even squid
access.log with headers (see below) was only showing an extra coma that
was difficult to spot.
On 11/09/2010 08:56
? Alternatively could you point me
to the place in squid sources where these decisions are taken? Maybe I
can better understand what I should do by looking at the code.
Many thanks,
Adrian
has one of the headers in vary with a
different value.
On 11/08/2010 06:33 PM, Adrian Dascalu wrote:
Done some new tests and I found out that caching an URL that has a
different X-Username header will invalidate the other version of that
object.
Is this the intended behaviour? I mean, Vary
method for cacheing multiple versions of the same URL ?
Adrian
On 11/08/2010 02:33 PM, Adrian Dascalu wrote:
Hi,
I'm out of ideeas trying to debug cache misses that I cannot explain. As a last
resort I'm sending this problem to the list with the hope that you could come
up with some e
passed ,from the same browser, I would get a MISS.
The request headers are IDENTICAL and there is no PURGE. What else might
invalidate the cached object?
Thank you,
Adrian
> On 08/10/10 05:42, Adrian Dascalu wrote:
> > Hi to all squid users!
> >
> > I'm new to this list so please hold the big guns.
>
> The problem you outline is discussed last in this reply. I've taken the
> opportunity to comment on the config improvements
On 10/7/2010 7:39 PM, Nick Cairncross wrote:
behind my ADSL router. Squid would be listening on, say, 80
>>to incoming requests. ADSL router would be port forwarding 80 to it. I
>>have a machine which runs a lot of web browseable add-ons all listening
>>on different ports (nzb, image library,
- NONE/- text/html
I've searched this list and internet in general for ideas of what I'm doing
wrong and came up empty.
I'm open to any suggestion for improvement in this setup. Performance is my
main goal.
Many thanks,
Adrian
On 2/6/10, Amos Jeffries wrote:
> Adrian Buciuman wrote:
> > Hi,
> >
> > Running squid 2.6 STABLE 21 from Centos 5.
> >
> > I've seen a issue with a site. The problem is that the browser locks
> > for some time, and only afterwards the user is able
and is not yet in stable
releases.
Thanks,
Adrian Buciuman
Mine too.. using Squid Version 3.1.0.15-20091212, and running on
configuration of TPROXY 4, have a problem on upload, squid often error
when upload facebook photo, attache an email on GMAIL or Yahoo, etc..
Irvan
On 1/23/2010 5:53 AM, Amos Jeffries wrote:
Linda Walsh wrote:
jay60103 wrote:
Please create an Issue and attach the patch. I'll see about including it!
adrian
2010/1/6 Rajesh Nair :
> Thanks for the response, Matt!
>
> Unfortunately the "cooperating" HTTP service solution would not work
> as I need to set the cookie for the same domain for wh
isn't already included.
I can make more educated guesses about where the likely CPU hog
culprits are given workload and configuration file information.
Adrian
2009/12/10 Guy Bashkansky :
> Is there an oprofile version for FreeBSD? I thought it is limited to
> Linux. On FreeBSD I tr
we want to try using CentOS 5.4. and soon after we install and
configure it with TPROXY, and test it, i will post it in here right a
way.. wish me luck :p
Irvan Adrian
Everything
Amos Jeffries wrote:
On Mon, 09 Nov 2009 20:46:19 +0700, Irvan Adrian K
wrote:
Dear Mr Amos, thanks for your res
ay be CentOS, for TPROXY..
Thanks,
Irvan Adrian
Dan wrote:
To throw in my 2 cents. I have tried to using both ubuntu server 9.04
and 9.10 neither of them I could get to work. I experienced the same
problem. So to make sure it wasn't me making a mistake somewhere I
tried the same config an
Dear Mr Amos, thanks for your respond, very helpfull..
Amos Jeffries wrote:
Irvan Adrian K wrote:
So, What the solution for these threads ? because i'm in the same
trouble to make TPROXY4 work in UBUNTU 9.10 Server
Explicit "Server" release or normal? I have recently
I use DROP. If it is working
for you then that is great. :) As for why
it works that way I don't know. When I use ACCEPT the
traffic is bridged through and not redirected to squid.
Thanks,
Irvan Adrian
Marko Kotar wrote:
Ok
My ebtable rules are(without -i option):
ebtables -t br
one from each Squid. There's no atomic
"append a line" IO method in UNIX so doing it that way won't work.
You could try hacking things up to lock/unlock the file for each
logfile write but I have no idea what the impact would be.
Adrian
2009/8/20 Joel Ebrahimi :
> Hi
The pipelining used by speedtest.net and such won't really get a
benefit from the current squid pipelining support.
Adrian
2009/8/15 Daniel :
> Henrik,
>
> I added 'pipeline_prefetch on' to my squid.conf and it still isn't
> working right. I've past
users faster.
* then in the helper, cache the IP results for a short period (say, 5
to 10 seconds) so frequent page accesses wouldn't result in a flurry
of requests to the backend
* keep the number of helpers low - you're doing it wrong if you need
more than 5 or 6 helpers doing this..
Have you asked the videocache group why it functions the way it functions?
adrian
2009/8/6 pavel kolodin :
> On Thu, 06 Aug 2009 05:34:09 -, Amos Jeffries
> wrote:
>
>
>>> Why?
>>
>> Possible reasons:
>>
>> 1) 302 being the status you reall
.. :)
If you have a set of "hot" traffic that moves over time, upping
cache_mem may not help.
2c,
Adrian
How much disk IO is going on when the CPU shows 70% IOWAIT? Far too
much. The CPU time spent in CPU IOWAIT shouldn't be that high. I think
you really should consider trying an alternative disk controller.
adrian
2009/8/4 smaugadi :
>
> Dear Adrian and Heinz,
> Sorry for the dela
Is this still involving the videocache stuff?
If it is, why aren't you asking them?
Adrian
2009/8/4 ░▒▓ ɹɐzǝupɐɥʞ ɐzɹıɯ ▓▒░ :
> (repost)
>>>> and how about caching online game patcher ? e.g ragnarok online, rohan
>>>> online, etc ?
>>>> is that
Investigate "tproxy"
Adrian
2009/8/4 Ja-Ryeong Koo :
> Hello,
>
> I am writing this email to ask something regarding ways to hide Caching
> Server IP address.
>
> I have one apache server, one caching server (squid2.6.stable22).
> (Client -- Caching Server
umented, and some coder makes a change which they
think won't have as much of an effect as it does. It has certainly
happened before in squid. :)
So, "more information please."
Adrian
Linux in the distant past, a lot of the IDE
drivers would have to manually transfer the data using PIO rather than
having a bus-master DMA transfer occur like many SCSI cards did. This
was counted to "IO wait."
Investigate what your storage driver is doing. :)
HTH,
Adrian
2009/8/
ng on your storage media and what the controller is doing
with it. You still haven't eliminated the possibility that the
controller/driver is somehow not helping.
You should also graph at least read/write IO count and byte counts;
investigate what is going on.
2c,
Adrian
2009/8/2 smaugadi :
&
Are you seeing high IO wait CPU use, or high IO wait times on IO?
Adrian
2009/8/2 smaugadi :
>
> Dear Adrian,
> Well my conclusion that this is an IO problem came from the fact that I see
> huge IO waits as the volume of traffic increase (with tools such as mpstat),
> when usin
2009/8/2 smaugadi :
>
> Dear Adrian,
> During the implementation we encountered issues with all kind of variables
> such as:
> Limit of file descriptors (now the squid is using 204800).
> TCP port range was low (increased to 1024 65535) TCP timers (changed them)
> The ip_co
but that is it.
The only support is some magic support for sharing the same incoming
HTTP socket between multiple, separate squid processes.
If you care about performance, Squid-2.7 is probably the best for you
at the moment from the Squid codebases..
Adrian
t finished deploying COSS. But then, I'm hitting other
issues with TCP stack tuning and CPU use.
Adrian
2009/8/2 smaugadi :
>
> Dear ALL,
> We have a squid server with high volume of traffic, 200 – 300 MB.
> The server is in transparent mode and using 18GB of ramdisk. With this
>
The donations were always few and far between. I'm not sure if there's
been any real active donations in the last twelve months; I think only
Duane knows.
Adrian
2009/8/2 Juan C. Crespo R. :
> Guys
>
> Checking the site I found there is no donation from December 2008, or it
Change "ufs" to "aufs" - assuming you compiled in aufs.
Consider upgrading to Squid-2.7.STABLEx - I did a whole lot of little
performance tweaks between 2.6 and 2.7.
Learn about oprofile and submit some performance information to help
developers. :)
Adrian
2009/7/28
This doesn't surprise me. They may be trying to maximise outbound
bits, or try to retain control over content, or not understanding
caching, or all/combination of the above.
I'd suggest contacting them and asking.
adrian
2009/7/26 Jason Spegal :
> A little bit messy but
m and ask why they've made their content uncachable?
Having cachable video content on websites will make them much, much
less likely to begin being blocked by bandwidth-strapped end-sites. :)
Adrian
gives me the green light.)
2c,
Adrian
Just break on SIGABRT and SIGSEGV. The actual place in the code where
things failed will be slightly further up the callstack than the break
point but it -will- be triggered.
Just remember to ignore SIGPIPE's or you'll have a strangely failing Squid. :)
adrian
2009/7/21 Marcus Koo
I'm using squid 2.6 on Centos 5.3
I've seen that the squid process keeps a saved user id of root (real
and effective uid are that of user squid). Is this normal? How can I
change it, and make squid give up root privileges completely?
Thanks,
Adrian Buciuman
2009/7/21 Soporte Técnico @lemNet :
> rep_mime_type can´t be used for parent selection because this is evaluated
> before content has been reached ?
Correct.
Adrian
I was going to say; I'm tweaking the performance of a cache with 21
million objects in it now. Thats a bti bigger than 2^24.
2009/7/16 Henrik Nordstrom :
> tor 2009-07-16 klockan 14:29 +1200 skrev Amos Jeffries:
>
>> For you with MB->GB files in Squid-2 that changes to faster Squid due to
>> limit
ithout investing some time into the Squid-2 fixes to do it.
I've toyed with it before and its reasonably easy to fix without
hurting performance.
2c,
Adrian
Are you using a url rewriter program?
Also, why haven't you just emailed redhat support?
Adrian
2009/7/15 Jarosch, Ralph :
> I found the section which rewrite the request in my cache.log.
>
> Can someone explain what happens there.
>
> 2009/07/15 06:51:56| cbdataValid: 0x1
ntact Redhat for support? That is why you're paying them for.
adrian
." I have no idea whether it is doing this without explicit
"don't further redirect" rules (eg by deny entries in the redirect
list, or "wccp exclude in", etc) because that may absolutely be
platform, IOS and WCCPv2 negotiation type dependant.
So please, poke the admin in question to get as much information about
the configuration and setup of everything.
Adrian
ch mentions bypass with class 90. So I'm very curious what
exactly it is that people are seeing, with what exact
configuration(s).
Adrian
to do if someone has some motivation. :)
Adrian
2009/7/9 tintin_vefg54e654g :
>
> Hi everyone,
>
> my configuration is as follow :
>
> I have a Mandriva 2009.1 OS, with squid ( + sarg, and mrtg) proxy.
> so, in the purpose to keep few free brandwicth for working using ^^ I wo
in the Wiki complete with explanations about how
all of the various parts of the puzzle hold together.
2c,
adrian
2009/7/2 Alexandre DeAraujo :
> I am giving this one more try, but have been unsuccessful. Any help is always
> greatly appreciated.
>
> Here is the setup:
> Router:
Upgrade to a later Squid version!
adrian
2009/6/30 goody goody :
>
> Hi there,
>
> I am running squid 2.5 on freebsd 7, and my squid box respond very slow
> during peak hours. my squid machine have twin dual core processors, 4 ram and
> following hdds.
>
> Filesystem
troller, but no
RAID." I wasn't given time to benchmark RAID vs non-RAID but in this
particular workload, RAID has never ever been faster in my testing in
cases other than the RAID card itself being buggy. Others have a
differing opinion.
Adrian
iter
interface and maintain rulesets for caching youtube, maps, windows
updates, etc. It just doesn't seem like it'll happen.
Adrian
ules. Oh, and
compile Squid "right".
2c,
Adrian
ve (which is at least as fast as Squid-2.HEAD) as a forward
proxy on some current generation hardware. It's peaking at 700
requests/sec and ~120mbit a sec with a ~ 30% byte hit rate.
A reverse proxy with a high hit rate should do quite a bit better than that.
Adrian
2009/6/26 Phibee Network Operation Center :
> ok the bug are not resolved no ?
The bugs get resolved when someone contributes a fix.. :)
Adrian
Squid-2.HEAD has some internal rewriting support.
I'm breaking it out into a separate module in Lusca (rather than being
an optional part of the external rewriter) to make using it in
conjunction with the external URL rewriter possible.
Adrian
2009/6/26 Jeff Pang :
> Does squid
Its a per-cache_dir option in Squid-2.7 and above; I'm not sure about 3.
Adrian
2009/5/20 Jason Spegal :
> Just tested and verified this. At least in Squid 3.0 minimum_object_size
> affects both memory and disk caches. Anyone know if this is true in 3.1 as
> well? Any thought
Squid doesn't currently implement any smarts for the WCCPv2 return path.
Adrian
2009/5/6 kgardenia42 :
> On Fri, May 1, 2009 at 5:28 AM, Amos Jeffries wrote:
>> kgardenia42 wrote:
>>>
>>> On 4/30/09, Ritter, Nicholas wrote:
>>>>
>>>> *
pshots.
I'll commit whatever other fixes are needed in this environment here :)
Thanks,
Adrian
#x27;m going to have invent my own..
2c,
Adrian
2009/5/2 Roy M. :
> In http://highscalability.com/youtube-architecture , under "Serving
> Thumbnails", it said:
>
> .
> - Used squid (reverse proxy) in front of Apache. This worked for a
> while, but as load increased per
me develop and test my
performance and feature related Squid work.
Adrian
d 64k
Squid-3 "sort of" fixed this. It wasn't ever fully fixed, much like
how the problem could be fixed in Squid-2 if someone wanted to do the
slight trickery required.
Adrian
people will write concurrent software,
and so will run multiple threads to do the socket and network stuff
(copyin/copyout/tcp/ip stuff, with a kernel thread handling part of
the NIC stuff and potentially some of the TX/RX.)
Adrian
Then it may be a bug. :)
Adrian
2009/1/20 Tawan Won :
> As you see the out of object dump in my previous mail, there is no client
> fetching the object.
> If an object has clients fetching it, object dump should print out the
> client list information too, if any.
> In additi
If it hasn't been swapped out to disk, the object has to stay in RAM
until the client(s) currently fetching from it have fetched enough for
part of the object (ie, the stuff at the beginning which has been sent
to clients) to be freed.
Adrian
2009/1/20 Taehwan Weon :
> Hi,
>
&
my URL rewriter now instead of a custom bit of perl code.
Hopefully having this simple rewriter out there will tease a few of
you to start using it and sharing configuration file snippets, which
is a whole lot easier than trying to share rewriter code. :)
Have fun,
Adrian
(http://code.google.com/p/squidtools/)
File a bug. :)
Adrian
2009/1/17 Chudy Fernandez :
> squid configuration shows weird on cachemgr
>
> @squid.conf
> delay_pools 1
> delay_class 1 2
> delay_parameters 1 18432/18432 -1/-1
> delay_initial_bucket_level 100
> delay_body_max_size 1048576 1 allow all
>
eresting squid tools they'd like to include in
the squidtools code project then please let me know. I'd like to
eventually have the whole collection available as a single "set" which
can be packaged up and installed together to enhance existing and new
Squid (and cacheboy :) installations.
Thanks,
Adrian
of hacking went into it to properly support async disk IO and thus
perform with any semblence of working well. its possible there's a bug
which I just haven't seen in production.
2c,
Adrian
>
>>> adrian
>>>
>>>
>>> 2008/9/12 Mark Powell :
>&g
y for me to fix
bugs. :)
Having statistics of your running server is the best thing to do for
debugging and provisioning, so please consider installing the package
and setting it up.
Enjoy!
Adrian
No, I don't think it can.
I'm just wrapping up some changes to FreeBSD-current and my Squid fork
to support tproxy-like functionality under FreeBSD + ipfw.
Adrian
2009/1/7 Mehmet ÇELİK :
>
>> As per usual, the easiest fix is to re-write the web app properly.
>> The RE
and present them in a web interface.
Adrian
I'm still not sure whether the correct behaviour is to send ICP for
the rewritten URL, or to rewrite the URLs being received before
they're looked up.
Hm!
Adrian
2008/12/24 Imri Zvik :
> On Wednesday 24 December 2008 17:01:39 Adrian Chadd wrote:
>> Thanks. Be sure to comm
Thanks. Be sure to comment on the bugzilla ticket too.
Oh and tell me which bug it is so I can make sure I'm watching it. :)
Adrian
2008/12/23 Imri Zvik :
> On Sunday 21 December 2008 10:52:42 Imri Zvik wrote:
>> Hi,
>>
>> On Thursday 18 December 2008 21:57:22 Adrian
fine.
adrian
2008/12/21 Oleg Motienko :
> On Tue, Jun 17, 2008 at 1:24 AM, Henrik Nordstrom
> wrote:
>> On mån, 2008-06-16 at 08:16 -0700, pokeman wrote:
>>> thanks henrik for you reply
>>> any other way to save bandwidth windows updates almost use 30% of my entir
Nope, I don't think the storeurl-rewriter stuff was ever integrated into ICP.
I think someone posted a patch to the squid bugzilla to implement this.
I'm happy to commit whatever people sensibly code up and deploy. :)
Adrian
2008/12/18 Imri Zvik :
> Hi,
>
> I'm us
limb along with the memory usage.
Grab oprofile and do some digging?
Adrian
>
> Mark.
>
>
>
> -Original Message-
> From: Kinkie [mailto:gkin...@gmail.com]
> Sent: Wednesday, December 17, 2008 4:50 PM
> To: Mark Kent
> Cc: squid-users@squid-cache.org
>
ASYNC_CREATE 1
#define ASYNC_WRITE 0
#define ASYNC_READ 1
Thats by default on Squid-2.HEAD. I've just changed them all to be
async under cacheboy-1.6 and this performs great under freebsd-7 +
AUFS with my testing.
Adrian
2008/12/11 Bin Liu :
> Thanks for your reply, Adrian. I'm ver
Its a hack which is done to defer a storage manager transaction from
beginning whilst another one is in progress for that same connection.
I'd suggest using your OS profiling to figure out where the CPU is
being spent. This may be a symptom, not the cause.
adrian
2008/12/7 Bin Liu &l
There isn't. Sorry.
Adrian
2008/12/7 Kaustav Dey Biswas <[EMAIL PROTECTED]>:
> Hi Adrian,
>
> Thanks a lot for your prompt reply.
>
> Actually, I need to implement the quota system as a part of my final year
> Engineering project. I am planning to make it a
2008/12/5 Nyamul Hassan <[EMAIL PROTECTED]>:
> Thx for the response Adrian. Earlier I was using only AUFS on each drive,
> and the system choked on IOWait above 200 req/sec. But, after I added COSS
> in the mix, it improved VASTLY.
Well, thats why its there, right? :)
>
on implementing the
functionality in Squid-2.
Thanks,
Adrian
2008/12/5 Kaustav Dey Biswas <[EMAIL PROTECTED]>:
> Hi,
>
> I am a squid newbie. I am trying to set up daily download quotas for NCSA
> authorized users. I have a daemon running which checks the log files, and
> whn
hard-core tuning - or at least,
they've done it, but haven't published the results anywhere. :)
Adrian
2008/12/3 Nyamul Hassan <[EMAIL PROTECTED]>:
> Why aren't there any (or marginal / insignificant) improvements over 3
> spindles? Is it because squid is a single threade
Good detective work! I'm not sure whether this is a requirement or
not. Henrik would know better.
Henrik, is this worthy of a bugzilla report?
adrian
2008/11/30 Itzcak Pechtalt <[EMAIL PROTECTED]>:
> Hi
>
> I found some inefficiency in Squid TCP connection handling toward
Does Squid-2.7.STABLE5 exhibit this issue?
Adrian
2008/11/28 Marcel Grandemange <[EMAIL PROTECTED]>:
> Looks like squid broke itself again.
> If anybody could advise me as to what's happening here it would be great.
>
> Im thinking the move to v3 has been disasterous so
Heh. The best way under unix is a hybrid of threads and epoll/kqueue
w/ non-blocking socket IO.
Adrian
2008/11/28 Ken DBA <[EMAIL PROTECTED]>:
>
>
>
> --- On Sat, 11/29/08, Adrian Chadd <[EMAIL PROTECTED]> wrote:
>
>> From: Adrian Chadd <[EMAIL PROTECT
proxies
that achieve much more than 150mbit/sec, even considering the
shortcomings of the codebases, I can't help but think there's
something else going on that isn't specifically Squids' fault. :)
Adrian
2008/11/28 Ken DBA <[EMAIL PROTECTED]>:
>
>
>
> ---
efficiently across all Squid versions.
If its kept low then it performs just fine.
Adrian
nder Linux but operations such as write() and close()
block under FreeBSD (think 'writing out metadata', for example) and
this mostly gives rise to the notion of Linux "being better" by most
people who haven't studied the problem in depth. :)
hope that helps,
Adrian
2008/11/
Is that per-flow, or in total?
Adrian
2008/11/24 Ken DBA <[EMAIL PROTECTED]>:
> Hello,
>
> I was just finding the flow capacity for Squid is too limited.
> It's even hard to reach an upper limit of 150 MBits.
>
> How can I improve the flow capacity for Squid in the
Do it yourself, benchmark, post results?
2008/10/5 Rafael Gomes <[EMAIL PROTECTED]>:
> I have two Scsi discs. I can set a unique cache_dir and make a Raid 0,
> so i will improve the write or I can set two cache_dir one per disc.
>
> What is better?
>
> Are There any documents about information?
G'day everyone,
I'll be in San Francisco (ish area) from October 1 to October 7. Drop
me a line if you're interested in catching up for an impromptu Squid
related evening event sometime then.
Adrian
Well, what are the complete request/reply headers for each of the
requests you're testing with?
Adrian
2008/9/25 BUI18 <[EMAIL PROTECTED]>:
> My Squid Version is 2.6/STABLE14
>
> Here's my refresh_pattern from squid.conf
>
> #Suggested default:
> refresh_
uhm, "running without cache" would mean "don't use any disk storage"
I'd suggest trying to run squid with no aufs cache_dir lines, just the
NULL line (cache_dir null /). This rules out the disk storage as a
potential candidate for failure.
Adrian
2008/9/25
Firstly, you should use the internal DNS code instead of the external
DNS helpers.
Secondly, I'd do a little debugging to see if its network related -
make sure you've disabled PMTU for example, as WCCP doesn't redirect
the ICMP needed. Other things like Window scaling negotiation and such
may con
2008/9/14 chudy fernandez <[EMAIL PROTECTED]>:
> I've posted as ask by Adrian.
> http://wiki.squid-cache.org/WikiSandBox/Discussion/YoutubeCaching
>
> I wanna know if somebody out there has a better idea of how to fix
> it(temporarily) it inside the squid.
Keep an eye o
Well, I fixed the thing up under FreeBSD so it certainly was working
for me at some point.
I'm one server away from getting my polygraph test cluster going and
I'll hopefully be installing that tomorrow; I'll make sure COSS gets a
decent thrashing when thats all up and running.
1 - 100 of 1379 matches
Mail list logo
