wccp2
Everyone, I've decided to tidy up the wccp2 code and get it import-ready for squid-2. My aims: * configurable standard and dynamic services * handle multiple routers * stay sane during a cache configuration reload I'm basing it on Steven Wiltons code in our monotone repository (http://www.cacheboy.net/). So far I have: * configurable standard services * framework for dynamic services * multiple router support; this was already completed by Steven and I believe he has this bit in production So the bits left: * finish the dynamic service configuration options * spend some time thinking about what should happen during a cache configuration reload * dump the config out properly through cachemgr Anyone is welcome to contribute; the code can be found in my branch (net.cacheboy.squid.adrian). I'll hold back from merging it into net.cacheboy.squid until its at least stable in my current home setup (A couple of machines behind a Cisco 2651 speaking WCCP2.) Adrian
Re: problems with the squid-2.5 connection pinning
- Original Message - From: Henrik Nordstrom [EMAIL PROTECTED] To: Steven Wilton [EMAIL PROTECTED] Cc: squid-dev@squid-cache.org Sent: Saturday, April 15, 2006 11:15 PM Subject: Re: problems with the squid-2.5 connection pinning lör 2006-04-15 klockan 09:10 +0800 skrev Steven Wilton: Having seen your patch, I've added the Proxy-Support: headers, and also added a pinning flag to the request-flags struct to allow identification of a pinned connection. Looking at your patch I think you got the logics slightly wrong when adding the flag. Pinning is a property of the connections, not the individual requests. From the point where the server connection has indicated use of Microsoft authentication scheme the server-side connection should be exclusively reserved for the specific client connection, and requests from the same client connection should be handled both as pinned looking for a matching reserved server connection and as authenticated even if there is no Authorize header (Microsoft authentication only sends Authorize headers on the first request on the connection, subsequent requests automatically inherit the same credentials) Hmm, you're right. I'll follow the example in your patch to mark the client connection as pinned, and use this information to modify the pconn key. Regards Steven
AC_LTDL_DLLIB
anyone remember what this call in configure.in was for? It seems to be not present on current autoconf + libtool. --- configure.in21 Mar 2006 23:15:19 - 1.402 +++ configure.in17 Apr 2006 11:47:41 - @@ -32,7 +32,6 @@ AC_PROG_CXX AC_CANONICAL_HOST AC_DISABLE_SHARED AC_PROG_LIBTOOL -AC_LTDL_DLLIB AC_PROG_RANLIB -- GPG key available at: http://www.robertcollins.net/keys.txt. signature.asc Description: This is a digitally signed message part
Re: wccp2
On Mon, Apr 17, 2006, Adrian Chadd wrote: So the bits left: * finish the dynamic service configuration options * spend some time thinking about what should happen during a cache configuration reload * dump the config out properly through cachemgr * dynamic service option is now implemented * It hasn't been tested yet but the cache does associate to the router * the rebuild logic still hasn't been implemented; thats a next week job. Adrian
Re: problems with the squid-2.5 connection pinning
- Original Message - From: Henrik Nordstrom [EMAIL PROTECTED] To: Steven Wilton [EMAIL PROTECTED] Cc: squid-dev@squid-cache.org Sent: Saturday, April 15, 2006 11:15 PM Subject: Re: problems with the squid-2.5 connection pinning lör 2006-04-15 klockan 09:10 +0800 skrev Steven Wilton: Having seen your patch, I've added the Proxy-Support: headers, and also added a pinning flag to the request-flags struct to allow identification of a pinned connection. Looking at your patch I think you got the logics slightly wrong when adding the flag. Pinning is a property of the connections, not the individual requests. From the point where the server connection has indicated use of Microsoft authentication scheme the server-side connection should be exclusively reserved for the specific client connection, and requests from the same client connection should be handled both as pinned looking for a matching reserved server connection and as authenticated even if there is no Authorize header (Microsoft authentication only sends Authorize headers on the first request on the connection, subsequent requests automatically inherit the same credentials) Thanks for pointing this out. I've updated the pinning patch to fix this problem, and tested on my home connection. I can confirm that it works for a simple http GET command, and I'll do further testing and update this list with the results using frontpage (which uses a variety of other http methods to transfer data). Due to other changes in the squid source, I needed to set the must_keepalive flag on the request to stop squid from closing the client-side connection, and I also had to remove the Connection: Proxy-support header from being sent back to the client (this caused IE to get really confused). regards Steven pinning.patch Description: Binary data
Re: Hi there
I approved your subscription. Welcome to the fold! adrian On Tue, Apr 18, 2006, Doug Dixon wrote: Hi all I've been a user of Squid for a few years now, and am interested in devoting some time to Squid-3.0 development, bugfixing and testing. I'd be most interested in helping to get the 3.0 branch up to the next PRE and STABLE, but obviously there will be some more feature development. In terms of features, I'm interested in helping with the reverse proxy stuff, e.g. collapsed forwarding etc. Hopefully you'll let me join the squid-dev mailing list which would be a good starting point :) Cheers Doug
Re: AW: [squid-users] Howto Clear Cache Periodicaly
HI, My select loop called average time is already about 50ms, with only 20 users. Where/How should I fine tune? FYI, I'm having 2 X 25GB cache_dir now with only 3% currently filled up. I'm going to put about 400 users more later. here is the cache mgr output. Kindly let me know which to be monitored and any suggestion on the baseline. Thank you so much.. Connection information for squid: Number of clients accessing cache: 20 Number of HTTP requests received: 31772 Number of ICP messages received:0 Number of ICP messages sent:0 Number of queued ICP replies: 0 Request failure ratio: 0.00 Average HTTP requests per minute since start: 22.0 Average ICP messages per minute since start:0.0 Select loop called: 1480208 times, 58.636 ms avg Cache information for squid: Request Hit Ratios: 5min: 0.0%, 60min: 16.7% Byte Hit Ratios:5min: 4.8%, 60min: 6.9% Request Memory Hit Ratios: 5min: 0.0%, 60min: 0.0% Request Disk Hit Ratios:5min: 0.0%, 60min: 2.4% Storage Swap size: 1924632 KB Storage Mem size: 7020 KB Mean Object Size: 31.68 KB Requests given to unlinkd: 414 Median Service Times (seconds) 5 min60 min: HTTP Requests (All): 0.61549 0.61549 Cache Misses: 0.61549 0.72387 Cache Hits:0.0 0.00307 Near Hits: 0.0 0.58309 Not-Modified Replies: 0.0 0.00194 DNS Lookups: 0.0 0.0 ICP Queries: 0.0 0.0 Resource usage for squid: UP Time:86794.142 seconds CPU Time: 47.467 seconds CPU Usage: 0.05% CPU Usage, 5 minute avg:0.01% CPU Usage, 60 minute avg: 0.01% Process Data Segment Size via sbrk(): 16552 KB Maximum Resident Size: 0 KB Page faults with physical i/o: 0 Memory usage for squid via mallinfo(): Total space in arena: 16552 KB Ordinary blocks:16415 KB 85 blks Small blocks: 0 KB 0 blks Holding blocks: 1228 KB 2 blks Free Small blocks: 0 KB Free Ordinary blocks: 136 KB Total in use: 17643 KB 99% Total free: 136 KB 1% Total size: 17780 KB Memory accounted for: Total accounted:12856 KB memPoolAlloc calls: 5633363 memPoolFree calls: 5415605 File descriptor usage for squid: Maximum number of file descriptors: 1024 Largest file desc currently in use: 21 Number of file desc currently in use: 16 Files queued for open: 0 Available number of file descriptors: 1008 Reserved number of file descriptors: 100 Store Disk files open: 0 Internal Data Structures: 60840 StoreEntries 1465 StoreEntries with MemObjects 1463 Hot Object Cache Items 60759 on-disk objects --- Adrian Chadd [EMAIL PROTECTED] wrote: On Thu, Apr 13, 2006, nonama wrote: Hi, Thank you for your help. Can you let me know on other setting baseline that I should look into using the cache manager? i.e Select loop called average time - should be under 20 milliseconds , in order for me to monitor my squid performance? Thats just one indication that squid is taking too long doing stuff to adequately service network requests. It'll increase quite quickly under load if your running short of CPU cycles. There's other variables you want to look at. The request rate is a good place to start. What is it at? adrian __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
Re: AC_LTDL_DLLIB
On Monday 17 April 2006 19:50, Robert Collins wrote: anyone remember what this call in configure.in was for? It seems to be not present on current autoconf + libtool. --- configure.in21 Mar 2006 23:15:19 - 1.402 +++ configure.in17 Apr 2006 11:47:41 - @@ -32,7 +32,6 @@ AC_PROG_CXX AC_CANONICAL_HOST AC_DISABLE_SHARED AC_PROG_LIBTOOL -AC_LTDL_DLLIB AC_PROG_RANLIB Its for dynamicly loadable libraries/modules/plugins. Its part of libtool and is present in libtool-1.5.10 and libtool-1.5.22 It exists in an aclocal file - ltdl.m4 -- regs MR E_T ___ \ \ \ OOHH I hate TYPOS \ \ \
wccp2 and squid reload semantics
Hiya, I've completed the WCCP2 configuration changes. Here's an example setup from the cachemgr config output: wccp2_router 192.168.2.2:2048 wccp2_version 4 wccp2_forwarding_method 1 wccp2_return_method 1 wccp2_service dynamic 80 wccp2_service standard 0 wccp2_service_info 80 priority=240 flags=src_ip_hash,ports_defined ports=80,81,82,83 protocol=tcp wccp2_incoming_address 0.0.0.0 wccp2_outgoing_address 255.255.255.255 (Its a bit redundant, but you get the idea.) What should the cache reload semantics be when Squid is reloaded? Should I just start sending the updated here_i_am messages and let the routers sort everything out? Or is there a better way to handle this? Adrian
Re: wccp2 and squid reload semantics
tis 2006-04-18 klockan 10:49 +0800 skrev Adrian Chadd: What should the cache reload semantics be when Squid is reloaded? Should I just start sending the updated here_i_am messages and let the routers sort everything out? Or is there a better way to handle this? It's probably possible to do it smart, but it's sorted out automatically pretty quickly so there is not much to gain from trying to be smart, at least not compared to the complexity to get it correct.. I guess it would be possible without too much effort to keep an MD5 checksum or similar of the wccp2 config section, and only restart the WCCP if there has actually been any relevant configuration changes.. Regards Henrik signature.asc Description: Detta är en digitalt signerad meddelandedel
Re: wccp2 and squid reload semantics
Hm, reading the IETF draft again does show a shutdown command which we should probably send if we've removed a service/router. adrian On Tue, Apr 18, 2006, Henrik Nordstrom wrote: tis 2006-04-18 klockan 10:49 +0800 skrev Adrian Chadd: What should the cache reload semantics be when Squid is reloaded? Should I just start sending the updated here_i_am messages and let the routers sort everything out? Or is there a better way to handle this? It's probably possible to do it smart, but it's sorted out automatically pretty quickly so there is not much to gain from trying to be smart, at least not compared to the complexity to get it correct.. I guess it would be possible without too much effort to keep an MD5 checksum or similar of the wccp2 config section, and only restart the WCCP if there has actually been any relevant configuration changes.. Regards Henrik
Re: AC_LTDL_DLLIB
On Tue, 2006-04-18 at 09:13 +0800, Mr E_T wrote: On Monday 17 April 2006 19:50, Robert Collins wrote: anyone remember what this call in configure.in was for? It seems to be not present on current autoconf + libtool. --- configure.in21 Mar 2006 23:15:19 - 1.402 +++ configure.in17 Apr 2006 11:47:41 - @@ -32,7 +32,6 @@ AC_PROG_CXX AC_CANONICAL_HOST AC_DISABLE_SHARED AC_PROG_LIBTOOL -AC_LTDL_DLLIB AC_PROG_RANLIB Its for dynamicly loadable libraries/modules/plugins. Its part of libtool and is present in libtool-1.5.10 and libtool-1.5.22 It exists in an aclocal file - ltdl.m4 ah, I was missing a new split out package, libltdl3-dev on debian derived systems. Cheers, Rob -- GPG key available at: http://www.robertcollins.net/keys.txt. signature.asc Description: This is a digitally signed message part
Re: wccp2 and squid reload semantics
On Tue, Apr 18, 2006, Henrik Nordstrom wrote: Not critical on reconfigure, as we still most likely still accept redirected traffic and the router will automatically take us offline after a short while. What is more important is to send the shutdown event when Squid is actually shut down.. this to promptly make the router not direct any traffic to us. Thats not a bad thing to do actually. I'll go and add that in. Adrian