On 26/02/2015 5:59 a.m., Tory M Blue wrote: > Greetings > > I'm getting further along with my testing and am trying to go the route of > lcp or htcp since I have a 3 squid cache cluster. > > So the questions are.. > > 1) What is the best debug level to set in order to see either icp or htcp > information? (I've been using tcpdump at the server level, just to see if I > can see traffic (and I'm not) :) > Its UDP traffic ;-P. "debug_options 12,9" will show everything. PS. you can use one or the other, not both at the same time to a given peer. HTCP is far more accurate than ICP with HTTP/1.1 (thus recommmended) but more bandwidth heavy. > 2) I've got the reported pinger exiting errors and wondering if this is why > icp or htcp doesn't seem to be working? No pinger is doing ICMP. It will mean the ICMP metrics for peer latency, and up/down status will be a bit more inaccurate than they could otherwise be. But generally nothing to worry over. It can be fixed by setting pinger binary owner to "root" with permissions 755. > > 3) how do I tell squid to ignore ipv6 and concentrate on ipv4? "dns_v4_first on" Squid will not ignore IPv6 entirely (see RFC 6540 / BCP 177 for why), but will prefer IPv4 servers. This is generally not a good idea. For ~80% of networks IPv6 operates faster than IPv4. The ~20% where it only operates just-as-fast as IPv4 are suspected to be operating over tunnels. Amos _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
