Hi,
I try to set up Squid servers as web accelerator. Each Squid server
should check if their siblings have the object before they get the
object from the real web server. All of the object has unique URL, so it
is safe to cache them forever in Squid servers. The purpose is to reduce
the load on the real web server. Here is my configuration:
I N T E R N E T
| |
+----+----+ +----+----+
| Squid 1 | | Squid 2 |
+----+----+ +----+----+
192.168.1.41| |192.168.1.40
+--------+--------+
|192.168.1.14
+-----+------+
| Web server |
+------------+
My squid.conf looks like this:
##-----------------------
## squid.conf on Squid 1
##-----------------------
acl all src 0.0.0.0/0.0.0.0
acl local_dst dst 192.168.1.0/255.255.255.0
acl local_src src 192.168.1.0/255.255.255.0
http_access allow local_dst
http_access allow local_src
http_access deny all
visible_hostname squid1.testing.com ## my host name
cache_dir diskd /var/spool/squid 50000 256 256
cache_store_log /usr/local/squid/var/logs/store.log
http_port 80
httpd_accel_host 192.168.1.14
httpd_accel_port 8080
cache_access_log /usr/local/squid/var/logs/access.log
cache_mem 1 GB
cache_effective_user squid
cache_effective_group squid
mime_table /usr/local/squid/etc/mime.conf
client_persistent_connections off
cache_log /usr/local/squid/var/logs/cache.log
icp_access allow local
icp_access deny all
cache_peer 192.168.1.40 sibling 80 3130
refresh_pattern . 10080 100% 10080 ignore-reload override-lastmod
##----------
I checked the log files, but it doesn't seem that the Squid servers talk
to their sibling. Is this the right way to do it? I tried to check the
web, but I didn't find any configuration for web accelerator with
siblings setting. Should I use 2 layer of Squid servers? Please advise.
Thank you very much.
--bk
Received on Thu Nov 24 2005 - 21:05:22 MST
This archive was generated by hypermail pre-2.1.9 : Thu Dec 01 2005 - 12:00:10 MST