G'day, A few notes. * Diskd isn't stable, and won't be until I commit my next set of patches to 2.7 and 3.0; use aufs for now. * Caching windows updates will be possible in Squid-2.7. It'll require some rules and a custom rewrite helper. * 3.0 isn't yet as fast as 2.6 or 2.7. Adrian On Tue, Feb 12, 2008, pokeman wrote: > > Well I experience with squid cache not good works on heavy load I 4 core > processor machine with 7 scsi drives 4 gb ram average work load in peak > hours 3000 users 30 mb bandwidth on that machine using RHEL ES 4. I search > many articles on high cache performance specially windows update these days > very headache to save PSF extension i heard In squid release 3.0 for better > performance but why squid developers could???nt find solution for cache > windows update in 2.6 please suggest me if I am doing something wrong in my > squid.conf > > > http_port 3128 transparent > range_offset_limit 0 KB > cache_mem 512 MB > pipeline_prefetch on > shutdown_lifetime 2 seconds > coredump_dir /var/log/squid > ignore_unknown_nameservers on > acl all src 0.0.0.0/0.0.0.0 > acl ourusers src 192.168.100.0/24 > hierarchy_stoplist cgi-bin ? > maximum_object_size 16 MB > minimum_object_size 0 KB > maximum_object_size_in_memory 64 KB > cache_replacement_policy heap LFUDA > memory_replacement_policy heap GDSF > cache_dir diskd /cache1 7000 16 256 > cache_dir diskd /cache2 7000 16 256 > cache_dir diskd /cache3 7000 16 256 > cache_dir diskd /cache4 7000 16 256 > cache_dir diskd /cache5 7000 16 256 > cache_dir diskd /cache6 7000 16 256 > cache_dir diskd /cache7 7000 16 256 > cache_access_log none > cache_log /var/log/squid/cache.log > cache_store_log none > dns_nameservers 127.0.0.1 > refresh_pattern windowsupdate.com/.*\.(cab|exe|dll) 43200 100% 43200 > refresh_pattern download.microsoft.com/.*\.(cab|exe|dll) 43200 100% 43200 > refresh_pattern au.download.windowsupdate.com/.*\.(cab|exe|psf) 43200 100% > 43200 > refresh_pattern ^ftp: 1440 20% 10080 > refresh_pattern ^gopher: 1440 0% 1440 > refresh_pattern cgi-bin 0 0% 0 > refresh_pattern \? 0 0% 4320 > refresh_pattern . 0 20% 4320 > negative_ttl 1 minutes > positive_dns_ttl 24 hours > negative_dns_ttl 1 minutes > acl manager proto cache_object > acl localhost src 127.0.0.1/255.255.255.255 > acl to_localhost dst 127.0.0.0/8 > acl SSL_ports port 443 563 > acl Safe_ports port 1195 1107 1174 1212 1000 > acl Safe_ports port 80 # http > acl Safe_ports port 82 # http > acl Safe_ports port 81 # http > acl Safe_ports port 21 # ftp > acl Safe_ports port 443 563 # https, snews > acl Safe_ports port 70 # gopher > acl Safe_ports port 210 # wais > acl Safe_ports port 1025-65535 # unregistered ports > acl Safe_ports port 280 # http-mgmt > acl Safe_ports port 488 # gss-http > acl Safe_ports port 591 # filemaker > acl Safe_ports port 777 # multiling http > acl CONNECT method CONNECT > http_access allow manager localhost > http_access deny manager > http_access deny !Safe_ports > http_access deny CONNECT !SSL_ports > http_access allow ourusers > http_access deny all > http_reply_access allow all > cache allow all > icp_access allow ourusers > icp_access deny all > cache_mgr info@xxxxxxxxxx > visible_hostname CE-Fariya > dns_testnames localhost > reload_into_ims on > quick_abort_min 0 KB > quick_abort_max 0 KB > log_fqdn off > half_closed_clients off > client_db off > ipcache_size 16384 > ipcache_low 90 > ipcache_high 95 > fqdncache_size 8129 > log_icp_queries off > strip_query_terms off > store_dir_select_algorithm round-robin > client_persistent_connections off > server_persistent_connections on > persistent_request_timeout 1 minute > client_lifetime 60 minutes > pconn_timeout 10 seconds > > > > Adrian Chadd wrote: > > > > On Thu, Jan 31, 2008, Chris Woodfield wrote: > >> Interesting. What sort of size threshold do you see where performance > >> begins to drop off? Is it just a matter of larger objects reducing > >> hitrate (due to few objects being cacheable in memory) or a bottleneck > >> in squid itself that causes issues? > > > > Its a bottleneck in the Squid code which makes accessing the n'th 4k > > chunk in memory take O(N) time. > > > > Its one of the things I'd like to fix after Squid-2.7 is released. > > > > > > > > Adrian > > > > > > > > -- > View this message in context: http://www.nabble.com/Mem-Cache-flush-tp14951540p15449954.html > Sent from the Squid - Users mailing list archive at Nabble.com. -- - Xenion - http://www.xenion.com.au/ - VPS Hosting - Commercial Squid Support - - $25/pm entry-level VPSes w/ capped bandwidth charges available in WA -
