[squid-users] TIMEOUT_ROUNDROBIN_PARENT

It seems that my back-end servers periodically stop responding to
requests from squid. Below is the output of the cache.log with
debugging turned on and the corresponding entry from access.log.

When I query the back-end process independent of squid, they respond.

Can anyone shed any more light on the subject? I know that's a very
general question, but I'm at a bit of a loss. Is my analysis of the
problem correct? Any insight into what might cause this problem, or
what I can do to further diagnose the problem?

Thanks,
Mark

==> cache.log <==
2006/08/21 16:34:55| The request GET http://xxx.xxx.com:80/index.html is
ALLOWED, because it matched 'all'
2006/08/21 16:34:56| carpSelectParent: CARP Calculating hash for
http://backendpool/VirtualHostBase/http/xxx.xxx.:80/site02/VirtualHostRoot/index.html
2006/08/21 16:34:56| 'Keep-Alive' header denied by anonymize_headers
configuration
2006/08/21 16:34:56| The reply for GET
http://backendpool/VirtualHostBase/http/xxx.xxx.com:80/site02/VirtualHostRoot/index.html

is ALLOWED, because it matched 'all'
2006/08/21 16:34:57| clientReadRequest: FD 95: no data to process ((11)
Resource temporarily unavailable)

==> access.log <==
1156203297.497 2038 yyy.yyy.10.124 TCP_MISS/200 5244 GET
http://xxx.xxx.com/index.html - TIMEOUT_ROUNDROBIN_PARENT/10.0.20.8
text/plain
Received on Mon Aug 21 2006 - 18:03:00 MDT