Use stale-if-error cache directive! http://www.mnot.net/blog/2007/12/12/stale -----Original Message----- From: Morgan Storey [mailto:me@xxxxxxxxxxxxxxxx] Sent: Tuesday, March 29, 2011 10:42 AM To: squid-users@xxxxxxxxxxxxxxx Subject: Re: Squid reverse proxy with cache. Anyone got any other ideas on why this isn't delving into the cache when the backend server is down? On Tue, Mar 22, 2011 at 10:37 AM, Morgan Storey <me@xxxxxxxxxxxxxxxx> wrote: > Thanks Amos I tried your suggestions, still no good when the webserver > is offline. I get only a single line in the cache.log "2011/03/22 > 10:28:50| TCP connection to myAccel (10.0.0.18:80) failed" > I have put the squid.conf in the email to make corrections easier (I > have tried that offline mode on but it doesn't seem to work); > > dns_nameservers 10.0.0.2 10.0.0.1 > > cache_dir ufs d:/cache 8000 100 256 > cache_mem 512 MB > maximum_object_size_in_memory 10240 KB > > access_log e:/squidlogs/access.log squid > cache_log e:/squidlogs/cache.log > cache_store_log e:/squidlogs/store.log > logfile_rotate 10 > > > acl all src all > acl manager proto cache_object > acl localhost src 127.0.0.1/32 > acl to_localhost dst 127.0.0.0/8 0.0.0.0/32 > acl localnet src all # RFC1918 possible internal network > > acl SSL_ports port 443 > acl Safe_ports port 80 # http > acl Safe_ports port 443 # https > acl CONNECT method CONNECT > > ############################################################## > > http_port 80 accel defaultsite=www.site.com vhost > > # proxy icp > # hostname type port port options > # -------------------- -------- ----- ----- ----------- > cache_peer 10.0.0.18 parent 80 0 no-query originserver name=myAccel > > > acl our_sites dstdomain www.site.com > forwarded_for on > > http_access allow our_sites > cache_peer_access myAccel allow our_sites > cache_peer_access myAccel deny all > > cache_effective_group proxy > > ############################################################## > > # > #Allow ICP queries from All > icp_access allow all > > # TAG: max_stale time-units > # This option puts an upper limit on how stale content Squid > # will serve from the cache if cache validation fails. > # > #Default: > max_stale 1 week > > #Suggested default: > #refresh_pattern ^ftp: 1440 20% 10080 > #refresh_pattern ^gopher: 1440 0% 1440 > refresh_pattern -i (/cgi-bin/|\?) 0 0% 0 > refresh_pattern -i (/umbraco/|\?) 0 0% 0 > refresh_pattern . 120 20% 4320 > > acl QUERY urlpath_regex umbraco \? > acl POST method POST > no_cache deny QUERY > no_cache deny POST > > # TIMEOUTS > forward_timeout 15 seconds > > # Do not tell the world that which squid version we're running > httpd_suppress_version_string on > > # Remove the Caching Control header for upstream servers > header_access Cache-Control deny all > > cache_mgr someone@xxxxxxxxxxxxx > > #offline_mode on > > > On Mon, Mar 21, 2011 at 10:26 PM, Amos Jeffries <squid3@xxxxxxxxxxxxx> wrote: >> On 21/03/11 21:44, Morgan Storey wrote: > >> Some notes on those configs: >> >> (squid1) >> * You can deny CONNECT entirely. It should never happen to a reverse-proxy. >> They get native HTTPS connections to their https_port. >> >> * Your reverse-proxy settings, for cache_peer and the related http_access >> rule MUST be above the default forward-proxy rules. Since this appears to be >> a pure reverse-proxy you can drop the default config file http_access lines >> entirely. >> NP: do that and I believe your Squid will stop needing to wait for DNS. It >> only uses it now to check the localnet ACL :) >> >> * header_access is not a very safe thing to do for Squid-2. It affects both >> the client request headers passed to your main server AND the reply headers >> destined for the visitor. >> refresh_pattern flag "reload-into-ims" is much safer and helps avoid the >> forced reload clients sometimes send. >> >> (squid2) >> * you are missing the http_access reverse-proxy settings. squid1 was >> better. >> >> >> Amos >> -- >> Please be using >> Current Stable Squid 2.7.STABLE9 or 3.1.11 >> Beta testers wanted for 3.2.0.5 >> >
