SQL Camel wrote:
Hello,
When running with reverse proxy mode, when should we consider to use cache
cluster rather than all separate cache boxes?
From what I know, no matter with parents or with siblings every cache has all the objects of others, so is cache cluster really meaningful?
When the origin server is too slow or too far away or needs to do too
much work to generate and/or check objects.
Also when you don't want each proxy to maintain its own duplicate cache
use a peer clustering protocol that prevents duplicates. They are listed
in the cache_peer documentation.
Amos
--
Please be using
Current Stable Squid 2.7.STABLE9 or 3.1.1
