I just upgraded from CentOS 5.5 to CentOS 5.6, while running Squid v3.1.12.1 in
both environments, and somehow created a race condition in the process.
Besides updating the 200+ software packages that are the difference between 5.5
and 5.6, I configured and enabled DNSSEC on my nameserver.
On Wednesday 27 August 2008 08:48:36 am Steve Bertrand wrote:
Altrock, Jens wrote:
So there is no significant change in features, only in programming
language, is that right?
IPv6... ;)
There is no IPv6 support in Squid v3.0. It is schedule for inclusion in
v3.1
On Monday 25 August 2008 09:54:37 am Steve Bertrand wrote:
Amos Jeffries wrote:
Steve Snyder wrote:
Is there a current guesstimate as to when an IPv6-capable version
of Squid will be released?
According to the FAQ, IPv6 support is already checked into the
Squid3 trunk. It doesn't
Is there a current guesstimate as to when an IPv6-capable version of
Squid will be released?
According to the FAQ, IPv6 support is already checked into the Squid3
trunk. It doesn't give any indicate, though, of when this
functionality will be available in a non-development release.
On Thursday 06 March 2008 11:05:24 am Adrian Chadd wrote:
Well, the way I'd approach it is to first get an idea of how to throw
things into 'threads', and probably draft and craft a basic event loop
and submission queue for stuff to happen across threads.
Then Squid can run as one thread, and
On Tuesday 04 March 2008 7:36:50 am Adrian Chadd wrote:
Hi everyone,
I'm quite disappointed in the lack of feedback from the community over
this. Its hard to figure out what people want if noone speaks up, so
this is your time to speak up.
I see nothing attractive in Squid v3.0.
I don't
How do I allow/disallow access to specific sites by the day of the week?
Example: allow connections to http://www.fun_n_games.com only on Saturdays
and Sundays.
It's easy enough to see that
acl fun_n_games dstdomain .fun_n_games.com
acl funtime time SA # Sunday Saturday
But I don't
On Wednesday 30 January 2008 5:59:36 pm Go Wow wrote:
What are the pre-requisites software in order to have squid installed
on a machine.
An example. These are the libraries Squid wants to see on my Fedora 7
system at runtime:
$ ldd /usr/sbin/squid
linux-gate.so.1 = (0x0011)
Would those who regularly benchmark their Squid 2.6/2.7 installations
comment on the procedure they use?
I'm setting up a (Linux-based) Squid v2.7 system and have several
decisions to make about the software/hardware to use. I'd like to tune
the system for best responsiveness and don't want
On Thursday 20 December 2007 12:13:03 pm Alex Rousskov wrote:
On Thu, 2007-12-20 at 16:43 +, Tony Dodd wrote:
Jason wrote:
I'm about to build a cache for a production environment. Which do
I want, 2.7 or 3.0? If you have time for a detailed answer, what
are the differences? BTW,
On Wednesday 19 December 2007 9:56:03 am Adrian Chadd wrote:
Hi everyone,
The Squid-2.7 release should be tagged any day now, so we'd appreciate
it if people currently using Squid-2.6 in high-traffic environments
could give Squid-2.HEAD a whirl.
Is there a 2.6S17 -- 2.7S1 changelog?
What
On Tuesday 18 December 2007 4:12:42 pm [EMAIL PROTECTED] wrote:
Is it possible to get this plugin video player to work over squid? If
so, how?
http://movenetworks.com/support/renderer.html
Requires IE. We have squid/2.6.STABLE13
Thanks
1. Squid (2.6S17) doesn't cause any problems with
Seen with Squid 2.6S13 in cache.log. Whose problem is this, Squid's or
Yahoo's?
2007/06/03 09:54:22| ctx: enter level 0: 'http://www.yahoo.com/s/595507'
2007/06/03 09:54:22| WARNING: unparseable HTTP header field {al-estate}
2007/06/03 09:54:22| ctx: exit level 0
2007/06/03 09:54:22| WARNING:
Looking around, I find plenty if examples on how to block object types
and/or domain, but not on a full URL.
Can someone point to an example of how, on Squid 2.6S13, to block access
to a specific URL?
Thanks.
How do I force a specific URL to always be fetched from the Squid (2.6S10)
cache?
Thanks.
What do the new epoll and COSS options offer to me as the administrator of
a lightly-loaded Squid server? Anything?
I usually read of epoll in the context of being recommended for a Squid
server with very high CPU utilization. I'm not clear on the advantages
of COSS over other disk storage
On Monday 10 July 2006 12:08 pm, Rick Brooks wrote:
[snip]
All of you are quite aware that Internet access speeds are higher and
higher (15Mb is not unusual in the USA) but our squid servers are not
exceeding 3.5Mb to the client desktop behind it.
[snip]
What are you using to measure that
Regarding configuration of Squid v2.6:
--enable-kqueue / --disable-kqueue
Given that this is an OS-specific option, shouldn't the name reflect that?
I'm thinking --enable-bsd-kqueue.
--enable-linux-tproxy
Why use this instead of --enable-linux-netfilter on Linux 2.6.x systems?
The
Have you tried reiserfs with noatime,notail?
That's said to be among the fastest performers among mainstream
filesystems for Squid.
On Tuesday 27 June 2006 3:09 pm, O'Brien, Kevin wrote:
Unfortunately we already use ext2 and noatime, but thanks for the
suggestion.
-=Kevin=-
-Original
On Thursday 08 June 2006 6:30 pm, Chris Robertson wrote:
[snip]
hardware raid with sata good enough? Dual core any good? Does the
menory system benefit on dual chanel?
Get the fastest disks you can afford. The more spindles the better
(within reason). Don't RAID the cache_dir drives. With
On Squid 2.5S13, using ufs/aufs, I am using the recommended 256 L2 cache
directories. I find though that I never use more than about 140 of them.
That is, roughly 115 of every 256 directories are untouched since the
initial creation of the cache directory hierarchy.
Is there any advantage
I don't recall you saying which OS you're running, but if it is Linux I
suggest building Squid with the Intel compiler rather with GCC.
Intel's ICC compiler (Linux version is free for non-commercial use)
typically generates much better code for Intel CPUs than GCC does. I've
been building
Red Hat is enabling the --enable-truncate option in their Squid packages.
Is the use of this option advisable in my circumstances?
I'm running Squid 2.5S12 on RHEL4. My single cache is on a ReiserFS (v3.6)
filesystem, mounted with the noatime and notail options, on a very low
latency SCSI
I don't know the answer to your problem, but I can confirm that
transparent mode and squid_redirect do work together, at least on Linux.
Here's a snippet of my 2.5.STABLE12 squid.config:
httpd_accel_port 80
httpd_accel_host virtual
httpd_accel_with_proxy on
httpd_accel_uses_host_header on
Yesterday 48 occurrances of this showed up in my cache.log:
Malformed UTF-8 character (unexpected continuation byte 0xad, with no preceding
start byte) in pattern match (m//) at (eval 1) line 5, STDIN line 21982.
This with Squid v2.4S9 + all patches through April 20th, on a Linux system.
What
On Saturday 19 March 2005 3:32 am, Henrik Nordstrom wrote:
On Sat, 19 Mar 2005, Steve Snyder wrote:
Same here.
Applied all 4 patches on my Linux machine and Squid would restart
every few minutes. In cache.log is the note that a signal 6 was
received on an assert error.
Do you have
On Saturday 19 March 2005 7:25 am, Henrik Nordstrom wrote:
On Sat, 19 Mar 2005, Steve Snyder wrote:
Yes. It always fails here:
2005/03/18 23:12:51| assertion failed: store_swapout.c:277:
mem-inmem_hi == mem-swapout.queue_offset
Thats an odd one.. but due to the nature of the bug just
It seems that recent Red Hat packages of Squid are using the
--enable-truncate config option by default. Aparently someone there
thinks there is a gain to be made by truncating rather than unlinking the
cached objects.
I wonder if this is applicable to my Squid (v2.5S9) installation. I'm
Same here.
Applied all 4 patches on my Linux machine and Squid would restart every
few minutes. In cache.log is the note that a signal 6 was received on an
assert error.
I backed out the aufs path and all is well again.
On Friday 18 March 2005 11:03 pm, Awie wrote:
All,
After applying
I'm running Squid v2.5S8 on a Linux (Fedora Core 2) box. It seems that I
can't get configure to, ahem, configure the library manager.
If I do:
AR=xiar ./configure --set-some-squid-options-here
or
export AR=xiar
./configure --set-some-squid-options-here
I find that configure always
It appears that my local cache is completely bypassed in favor of a parent
proxy.
I need to have my Squid (v2.5S7 + patches) call a parent proxy for
authentication. After making what seemed to be reasonable additions to
my squid.conf file, I find that I can in fact run through the parent
On Sunday 02 January 2005 6:18 pm, Henrik Nordstrom wrote:
On Sat, 1 Jan 2005, Steve Snyder wrote:
I noticed that my Squid (2.5S7 + official patches) recently has a
multitude of zero-length files in the cache directories.
Maybe your Squid is compiled with --enable-truncate? If so then you
I noticed that my Squid (2.5S7 + official patches) recently has a
multitude of zero-length files in the cache directories. I see no error
or warning messages on my (Linux v2.4.28) system, either in cache.log or
in the system log. Squid seems to be doing it's job, but I'm concerned
about
On Sunday 19 December 2004 2:58 am, Joe Cooper wrote:
[snip]
But not by much. I tried out the Intel compiler a couple of years ago,
with barely measurable (statistically insignificant) results. And
while I'm sure the Intel compiler has improved in those two years, GCC
probably has also. And
I am running Squid (2.5S7, on FC2 Linux) on a notebook machine, a machine
which is used on several networks. One of these networks requires me to
use a parent proxy for Internet access. Is it possible to specify the
parent proxy, if any, outside the squid.conf config file?
Basically I'd like
On Tuesday 07 December 2004 8:09 am, Elsen Marc wrote:
[snip]
Basically I don't see the added value of your single-user-squid setup.
Won't browser caches have the same effect ?
A few reasons:
1. Squid does a better job of caching than my browser. (Despite my large
browser cache, the Squid
On Monday 06 December 2004 10:40 pm, Lucia Di Occhi wrote:
[snip]
Now my question is: what are the compile flags used in the fedora RPMS?
The standard flags (FC3, for x86 systems) are -march=i386 -mcpu=pentium4
-O2. That is, the i386 instruction set is used, with the instructions
arranged
I understand that there are some problems associated with configuring
Squid (2.5S6 + patches) as a transparent proxy. Are there any negative
affects from having a transparent config even if the browsers are
directly addressing the cache?
On my (Linux, RedHat v9) LAN the browsers should all be
Yes, ICC (both 7.1 and 8.0) will build Squid v2.5Sx. I'll mail you the squid.spec
file (I'm building for a Red Hat environment) if you want, but it really comes down to
only 2 steps.
1. Run the script to set the ICC environment variables:
/usr/local/intel/compiler/bin/iccvars.sh
2. I add
On Sunday 16 May 2004 10:48 pm, unixware wrote:
Thanks Steve for your reply
i think -tpp6 -xK i like -march=i686 in gcc
Yes, it is equivalent to -march=i686 in GCC. Likewise, -tpp7 -xN is a
replacement for GCC's -march=pentium4 switch.
after exporting all these u suggested squid-2.4
I'm using IE v6.0/SP1, accessing the Internet through a Squid 2.5S5 proxy.
All available patches have been applied to both IE and Squid.
By default IE disables HTTP v1.1 access when a proxy is in use - see the
Advanced tab in Internet Options. (The rationale is to avoid problem
with HTTP
I'm suddenly getting a lot of these errors. Here's an example:
2004/01/14 10:29:35| ctx: exit level 0
2004/01/14 10:29:35| ctx: enter level 0:
'http://www.squid-cache.org/Versions/v2/2.5/bugs/'
2004/01/14 10:29:35| httpProcessReplyHeader: Impossible keep-alive header
from
Hello.
In the book Linux Power Tools the author says that when using a ReiserFS
file system for a Squid cache, the rupasov hash yeilds optimal
performance.
Is this true? This is the first time I've read/heard this assertion being
made, and I'm pretty attentive to Squid performance tuning
On Saturday 26 July 2003 10:10 am, Henrik Nordstrom wrote:
On Saturday 26 July 2003 05.37, Steve Snyder wrote:
Well, I don't see any reason for the delay. Here's all the
http_access references I've got. Note that the logged TCP_DENIED
I posted previously are the result of the 2nd
I have a couple of questions regarding Squid (2.5S3 + patches) and
TCP_DENIED on my Linux (RedHat v7.3) system.
This is a prohibition I have in my config file:
acl imrworldwide dstdomain .imrworldwide.com
http_access deny imrworldwide
Below are excerpts from my log file, showing the
[ This should really go to the Squid developers list, but I'm not
subscribed to that one. Sorry.]
Before deploying new software into my production environment, I like to
test it for a couple of weeks in a less critical context. Now that Squid
2.5STABLE3 has accumulated 25 patches, some
For purposes of defining an ACL in Squid v2.5S1, is there a distinction
made between .domain.tld and domain.tld (note the leading dot in the
first case)?
For example, if I say:
acl slashdot dstdomain slashdot.org
no_cache deny slashdot
(note absense of leading dot) will that
I take it then that there is no config option or similar for building
Squid statically?
-- Forwarded Message --
Subject: How to build with static linking?
Date: Sat, 22 Feb 2003 12:35:40 -0500
From: Steve Snyder [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Is there a clean way
When building Squid v2.5S1 (+ all patches) do I need to use the
--enable-ssl configuration option if I am not doing authentication of
clients?
If so, what feature(s) of Squid require the SSL services?
Thanks.
Is there a clean way to tell the Squid v2.5S1 build process (on a Linux
system) that I want to link the binaries staticly? If so, how is it
done?
Doing a ./configure --help didn't tell me anything, and it seems that
setting a LDFLAGS=-static environment variable doesn't work either. I
know
50 matches
Mail list logo