Our company has two service providers, and at various times, they both have
routing problems, but usually not both at the same time, and sometimes only
with certain sites or regions. Most of the time, we want to use our main
connection, where we have two full T1's, dual homed with BGP. Sometimes
though it'd be advantageous to use our more local ISP, which is only half a
T1.
The machine running Squid has two IP addresses, one in the class C that gets
routed through the main connection, and one in the class C that gets routed
to the other provider. As a bonus, the other provider is also running Squid
and we are peered with them.
What I would like to do would be to have Squid try the main connection
first, then fail over to the other connection. Is this possible? Maybe by
running two copies of Squid, the main one going to the main connection and
caching, and the other a peer in passthrough mode? I just don't know how to
make Squid try to get it itself before trying a peer; normally it works the
other way around.
Maybe this isn't even the best way to configure; if anyone has different
ideas, I'd be happy to hear them.
Does anyone have any suggestions, pointers to howto's, etc? If this is an
old topic, maybe someone can point me to the proper thread in the archives
and I'll go read and shut up.
Received on Wed Mar 08 2000 - 08:45:21 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:51:58 MST