On 11.09.2012 12:39, grant lowe wrote:
Hi all.
Still having problems with squid. I'm past the step where I got the
UDP_DENIED messages, but now I'm getting UDP_MISS and TCP_MISS
errors.
It is working then.
a) those are NOT errors. They are MISS responses.
b) having a peer does not suddenly download the entire Internet into
your cache.
This very confusing. From the cache.log, I see the the startup
messages:
.
.
.
2012/09/10 17:15:55 kid1| Logfile Daemon: opening log
/home/squid-seo2/apps/squid-3.2.0.17/var/logs/access.log
2012/09/10 17:15:55 kid1| Squid plugin modules loaded: 0
2012/09/10 17:15:55 kid1| Store logging disabled
2012/09/10 17:15:55 kid1| DNS Socket created at [::], FD 15
2012/09/10 17:15:55 kid1| DNS Socket created at 0.0.0.0, FD 16
2012/09/10 17:15:55 kid1| Adding domain corp.bbi.com from
/etc/resolv.conf
2012/09/10 17:15:55 kid1| Adding domain prd.slacker.com from
/etc/resolv.conf
2012/09/10 17:15:55 kid1| Adding nameserver 10.2.20.10 from
/etc/resolv.conf
2012/09/10 17:15:55 kid1| Adding nameserver 10.2.11.14 from
/etc/resolv.conf
2012/09/10 17:15:55 kid1| Configuring Parent 10.2.12.13/8080/0
2012/09/10 17:15:55 kid1| Configuring Parent 10.2.12.14/8080/0
2012/09/10 17:15:55 kid1| Loaded Icons.
2012/09/10 17:15:55 kid1| Accepting reverse-proxy HTTP Socket
connections at local=10.2.12.73:80 remote=[::] FD 17 flags=9
HTTP protocol.
2012/09/10 17:15:55 kid1| Accepting ICP messages on [::]:3130
2012/09/10 17:15:55 kid1| Sending ICP messages from [::]:3130
ICP protocol.
2012/09/10 17:15:55 kid1| Accepting HTCP messages on [::]:4827
2012/09/10 17:15:55 kid1| Sending HTCP messages from [::]:4827
HTCP protocol.
2012/09/10 17:15:55 kid1| Configuring Sibling
prdseoproxy02.corp.bbi.com/3128/3130
I then run a telnet to port 80 on the box and do a simple GET. When I
do that I get this:
^F2012/09/10 17:26:55 kid1| temporary disabling (Service Unavailable)
digest from prdseoproxy02.corp.bbi.com
Cache Digest protocol seeking a copy of the sibling peers cache index.
The sibling is out of contact / not responding.
I tried repeating the telnet to the same client a couple more times,
thinking that squid would have in its cache. But it doesn't. In face
it repeats this message:
2012/09/10 17:36:11 kid1| Detected DEAD Sibling:
prdseoproxy02.corp.bbi.com
So I'm doing something wrong, but all I can see is that it's the
peer-to-peer communications somehow. Here is the pertinent data from
the squid.conf file:
http_port 10.2.12.73:80 accel vhost
# acl prdseoproxy01 src 10.2.12.73/32
# acl prdseoproxy02 src 10.2.12.74/32
acl squid_peer src 10.2.12.185/32
acl squid_peer src 10.2.12.186/32
acl lb_source src 10.2.12.0/28
http_access allow lb_source
icp_port 3130
htcp_port 4827
icp_access allow squid_peer
miss_access deny squid_peer
htcp_access allow squid_peer
cache_peer prdseoproxy02.corp.bbi.com sibling 3128 3130
Where is this sibling?
how is it configured?
does TCP from your squid to its port 3128 working okay? (the error
suggests not)
# http_access allow squid_peer
cache_peer 10.2.12.13 parent 8080 0 originserver default no-query
no-digest connect-timeout=15
cache_peer 10.2.12.14 parent 8080 0 originserver default no-query
no-digest connect-timeout=15
Yes, this is a reverse proxy, and I'm really stumped. Please help.
Thanks.
Amos
