On Wed, 16 Nov 2011 13:49:23 -0500, Schwenker, Stephen wrote:
Hello,
We've installed squid as an accelerator for our website, and it's
working pretty well, but we have found out that the collapsed
forwarding
is not working as expected. We ran some load tests to a single url
with
a 1 minute expiry and squid is still sending approximately 12000
request
to the back end server in a 5 minute span. If I understand the
documentation correctly, squid should only send 5 requests to the
back
end server.
Object variance can affect that. 5 requests is the best-case scenario
for a single static object representation.
What are the reply headers coming back with the object?
I've configured the following.
http_port 80 accel defaultsite=<site name> vhost
cache_peer <web server> parent 80 7 no-query no-digest originserver
sourcehash monitorurl=/statuscheck/ name=<web server> http11
... acls
... header access
Your "header access" re-writing may also affect the responses. What are
these lines?
Amos
