Re: [squid-users] redirector with Shutdown flag
Henrik, I might be wrong but here is what I found. In helper.c:GetFirstAvailable There is no check at helper.c:GetFirstAvailable for flags.shutdown and if the helper is still alive and already flagged with shutdown flag, squid will keep sending requests to it. Thanks, Val On Tue, 2 Nov 2004, Henrik Nordstrom wrote: On Tue, 2 Nov 2004, Valentin Chopov wrote: According to the Squid FAQ the only way to shutdown the redirectors is to close their stdin (squid's stdout). My understanding is that the shutdown flag is only for squid and there is no way redirectors to know about it. The redirector knows about it by the simple fact that stdin is EOF. So if for some reason redirector send something to stdout (squid's stdin) after squid set the Shutdown flag, squid will keep this redirector in service ... am I right? Thanks, Val Squid keeps the redirector in the table until the redirector acknowledges the close (by closing it's stdout) and exits, but it won't (and can't) send any new requests there as the communication channel squid->helper has been closed. It can happen perfectly fine that a helper is shut down while processing a request. In this case the helper should send the response when done processing the current request and then exit. Regards Henrik
Re: [squid-users] redirector with Shutdown flag
Henrik thanks. According to the Squid FAQ the only way to shutdown the redirectors is to close their stdin (squid's stdout). My understanding is that the shutdown flag is only for squid and there is no way redirectors to know about it. So if for some reason redirector send something to stdout (squid's stdin) after squid set the Shutdown flag, squid will keep this redirector in service ... am I right? Thanks, Val On Mon, 1 Nov 2004, Henrik Nordstrom wrote: On Mon, 1 Nov 2004, Valentin Chopov wrote: Do you think it will be a good idea Squid continuously to remind these redirectors to shutdown (e.g. every 1h)? The shutdown indication to the helper is permanent, always there. As soon as the helper tries to read the next request it notices.. Regards Henrik
Re: [squid-users] redirector with Shutdown flag
Henrik, Do you think it will be a good idea Squid continuously to remind these redirectors to shutdown (e.g. every 1h)? Thanks, Val On Sat, 30 Oct 2004, Henrik Nordstrom wrote: On Thu, 28 Oct 2004, Valentin Chopov wrote: How squid handles the redirectors with shutdown flag on. It looks that the redirector functions normaly even there is a shutdown flag. The "shutdown" flag indicates Squid has requested this redirector to shut down, but the redirector has not yet acknowledged the shutdown by closing it's input to Squid.. (or put in simpler words, Squid has requested this redirector to exit, but it has not exited yet) Regards Henrik == Valentin S. Chopov, CC[ND]P Sys/Net Admin SEI Data Inc. E-Mail: [EMAIL PROTECTED] ==
[squid-users] redirector with Shutdown flag
Hi, How squid handles the redirectors with shutdown flag on. It looks that the redirector functions normaly even there is a shutdown flag. Here is the "URL Redirector Stats" Thanks, Val ---cut here--- Cache Manager menu Redirector Statistics: program: /usr/local/squid25/bin/redirectorun number running: 9 of 8 requests sent: 3931155 replies received: 3931155 queue length: 0 avg service time: 5.21 msec # FD PID # Requests Flags TimeOffset Request 1 6 557 3484795 A S 0.001 0 (none) 1 9 31012 25593 A 0.003 0 (none) 2 10 31013 2587A 0.002 0 (none) 3 11 31014 574 A 0.003 0 (none) 4 12 31015 291 A 0.008 0 (none) 5 13 31016 198 A 0.008 0 (none) 6 15 31017 166 A 0.008 0 (none) 7 251 31018 148 A 0.008 0 (none) 8 255 31019 130 A 0.008 0 (none) Flags key: A = ALIVE B = BUSY C = CLOSING S = SHUTDOWN Number of requests bypassed because all redirectors were busy: 32897 ---cut here--- == Valentin S. Chopov, CC[ND]P Sys/Net Admin SEI Data Inc. E-Mail: [EMAIL PROTECTED] ==
RE: [squid-users] diskd VS aufs
On Wed, 30 Jun 2004, Elsen Marc wrote: Thanks. I suppose I do not need to rebuild the cache (./squid -z). But what No, you don't. makes you conclude that "FreeBSD diskd is the better choice. Under Solaris and Linux use aufs."? I seem to remember this was a FAQ, I may be wrong, at least I can't find it for the moment. From what I remember : under FreeBSD support for shared mem is better which diskd heavily uses. Linux and Solaris are better on thread support, which aufs uses. M. We are running Squid on FreeBSD 5.2-CURRENT with aufs. It looks that KSE improved the thread support in the FreeBSD. Val
Re: [squid-users] Squid CPU Performance
On Fri, 21 Nov 2003, Joel Jaeggli wrote: > On Fri, 21 Nov 2003, Mark Pelkoski wrote: > > > List, > > I FINALLY implemented a Squid server into my production environment > > today. It is squid-2.5.STABLE4-20031029 on a Quad Proc Xeon 500MHz with > > 1M Cache and 3x9Gig Raid-5 dedicated for cache and 2x9Gig Raid-1 for OS > > Redhat 9.0. I tested this exact server config with 74 users and > > performance was pretty good for 800 Reqs/Min. Now in Production I have > > 325+ users at 2700 Reqs/Min and performance stinks. It's like being on a > > dial-up connection. 3 of the procs are sitting below 1%. The other is > > used by Squid at 99.9%. It there any way to speed up performance on a > > multi-proc system? TIA. > > not really... it will however go a lot faster if you unraid the 3 9gb > drives and treat each one as a seperate cache dir... > I agree with Joel, and also you can use "cache_dir diskd" and this maybe will give some job for other processors too, there will be will be 3 separate "diskd" processes. > > > > Mark Pelkoski > > > > > > -- > -- > Joel Jaeggli Unix Consulting [EMAIL PROTECTED] > GPG Key Fingerprint: 5C6E 0104 BAF0 40B0 5BD3 C38B F000 35AB B67F 56B2 > > >
Re: [squid-users] cisco wccp problem
what is the ouput of the "tcpdump proto gre" on the Squid box ? Val On Wed, 27 Aug 2003, Kuba Leszewski wrote: > Damian-Grint Philip wrote: > > >What messages do you get if you type the following in at your Cisco router > >while browser traffic is crossing it? > >Term mon > >Debug ip icmp > > > > > Well, actually there's nothing. > Some ICMP packets, but none connected with the http traffic I generate. > > Kuba > > > > > > > == Valentin S. Chopov, CC[ND]P Sys/Net Admin SEI Data Inc. E-Mail: [EMAIL PROTECTED] ==
[squid-users] msblaster worm, squid & wccp ?
Hi, According to the W32.Blaster.Worm technical details, it will start DoS attack to Windows Update Site after Aug 16, 2003 . If this DoS attack will be on port 80 what will happen with the squid running as a transparent proxy with wccp? BTW, I couldn't find an info if the attack will be on port 80 or port 443. Thanks, Val == Valentin S. Chopov, CC[ND]P Sys/Net Admin SEI Data Inc. E-Mail: [EMAIL PROTECTED] ==
Re: [squid-users] Squid, WCCP, and Loading?
Did you try "no ip route-cache" on the 7200 FastEthernet Interface (the interface wich sends/receives wccp GRE packets to/from your Squid box. Val == Valentin S. Chopov, CC[ND]P Sys/Net Admin SEI Data Inc. E-Mail: [EMAIL PROTECTED] ==
[squid-users] cache.log varyEvaluateMatch, clientProcessHit
Hello, I'm getting the folloing in the cache.log file and it seems this happens with kazaa sites only. Any idea what that means. Thanks, Val 2003/02/28 07:56:34| varyEvaluateMatch: Oops. Not a Vary match on second attempt , 'http://desktop.kazaa.com/us/images/home/60_60_braidedhair_h.gif' 'accept-enco ding="gzip,%20deflate", user-agent="Mozilla%2f4.0%20(compatible%3b%20MSIE%206.0% 3b%20Windows%20NT%205.1)"' 2003/02/28 07:56:34| clientProcessHit: Vary object loop! 2003/02/28 08:09:41| varyEvaluateMatch: Oops. Not a Vary match on second attempt , 'http://search.kazaa.com/us/images/spacer.gif' 'accept-encoding="gzip,%20defla te", user-agent="Mozilla%2f4.0%20(compatible%3b%20MSIE%206.0%3b%20Windows%20NT%2 05.1%3b%20YComp%205.0.0.0)"' 2003/02/28 08:09:41| clientProcessHit: Vary object loop! == Valentin S. Chopov, CCNP Sys/Net Admin SEI Data Inc. E-Mail: [EMAIL PROTECTED] ==
Re: [squid-users] FrontPage sites cacheing
Henrik, thanks for the suggestion. Unfortunately not using WCCP is not an option for us;) I don't know if you mean "ie_refresh" option but it's tutned on in our case. Is there something else what we can do? Thanks again, Val On Thu, 20 Feb 2003, Henrik Nordstrom wrote: > Don't use interception (including WCCP). > > Using interception hides the fact that you are using a proxy cache > from the browsers, so the browsers do not know they need to include > cache instructions to instruct caches to fetch a fresh copy when the > user pushes the reload button.. > > Configure the browser to use a proxy and things should work much > better. > > > There is also some (one) options in squid.conf which may help.. > > Regards > Henrik > > > On Thursday 20 February 2003 19.48, Valentin Chopov wrote: > > Hi, > > > > It seems I have a problem with FrontPage sites cacheing. > > I got complaints that after making changes to their web > > pages (which are hosted on the servers outside our network) the > > updates can not be seen behind our squid proxy servers, even after > > hitting refresh on the browser. > > I think the most of complains are about FrontPage web sites but I > > didn't check all of them. > > We are running Squid 2.5 + WCCP. Currently I'm fixing this with > > ACLs on the Cisco routers to exclude the IPs from the WCCP. > > Is there a way to fix this globally? > > > > Thanks, > > > > Val > > > > == > > Valentin S. Chopov, CCNP > > Sys/Net Admin > > SEI Data Inc. > > E-Mail: [EMAIL PROTECTED] > > == > > >
[squid-users] FrontPage sites cacheing
Hi, It seems I have a problem with FrontPage sites cacheing. I got complaints that after making changes to their web pages (which are hosted on the servers outside our network) the updates can not be seen behind our squid proxy servers, even after hitting refresh on the browser. I think the most of complains are about FrontPage web sites but I didn't check all of them. We are running Squid 2.5 + WCCP. Currently I'm fixing this with ACLs on the Cisco routers to exclude the IPs from the WCCP. Is there a way to fix this globally? Thanks, Val == Valentin S. Chopov, CCNP Sys/Net Admin SEI Data Inc. E-Mail: [EMAIL PROTECTED] ==