On 17/12/2011 5:39 a.m., J Smith wrote:
Hello everyone,
I'm new to squid. I think I have it configured to do what I want, but I wanted
to run my squid.conf options (pertinent parts only) by more experienced eyes to
double-check.
I'm running squid (2.7.STABLE7) as a forward proxy on my personal computer,
with the sole purpose of caching frequently-used static content from a few
specific sites. Nothing else should get cached. Furthermore, once the content
is cached, it does not need to be revalidated.
Here goes:
# List sites to be cached
acl cache_us dstdomain a.example.com
acl cache_us dstdomain b.example.com
# Allow access from my computer only
http_access allow localhost
http_access deny all
# Only cache cache_us
cache deny !cache_us
# Once cached, do not revalidate
offline_mode on
A couple of questions:
* From my scan of the log files, this appears to work. Any reason why it won't
work, hidden pitfalls, etc.?
The usual pitfalls of force-caching things which are borderline for
cacheability. Broken page contents when any piece of the original page
changes.
The best way to achieve caching performance is to alter the origin
server(s) for these sites to produce cache-controls allowing their
content to be cached for the correct timespan.
* With offline_mode on, I'm assuming that other cache-tuning options (e.g.,
max_stale, refresh_pattern, minimum_expiry_time, etc.) are irrelevant;
squid will cache matching content once, always fetch it from the cache in the
future, and never revalidate it. Right?
They are still relevant. Squid just does not erase the cached content
and serves up stale objects without explicit permission from the origin
when failures occur.
* How/where are validation requests logged?
It shows up in the access.log as the result codes (TCP_*) with the
REFRESH or RELOAD sub-tag when each transaction finishes.
http://wiki.squid-cache.org/SquidFaq/SquidLogs#Squid_result_codes has
the details on what these tags are and mean.
Amos
