[squid-users] Squid 3.5 Reverse proxy. Pinger... icp/htcp

[Amos Jeffries]

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