Re: Transperant HTTP Redundancy.

From: Lyndon Levesley <lol@dont-contact.us>
Date: Wed, 25 Mar 1998 14:21:44 +0000

>>>>> On Wed, 25 Mar 1998 at around 10:37:16,
>>>>> "GV" == Goncalo Valverde penned:

 GV> Greetings!

 GV> On 25-Mar-98 Javier Puche. CSIC RedIRIS wrote:

 +> - Have both machines with IP addresses different than x.x.x.x but with
 +> the ability to listen to x.x.x.x with a virtual interface when needed.
 +> - When machine 1 attending x.x.x.x falls down, have the other machine
 +> (2) detect that situation and set up a virtual interface for attending
 +> x.x.x.x , do it also the other way around.
 +> - Make sure that if either machine goes up again doesn't listen to IP
 +> x.x.x.x before checking that the other machine is not doing that
 +> already.

 GV> Theres an url explaining that somewhere... but the problem is that this kind
 GV> of solution is not bulletproof.. imagine that for some reason the kernel panics
 GV> with an harddisk failure.. sometimes when this happens the machine still
 GV> replies to pings, squid is up and running, but obviously doesnt replies to
 GV> requests.. so you could grab the ip address because the other machine is still
 GV> using it... or is there any way around this problem?

 This is just me rambling, but couldn't you do :

                          |CISCO|
                             |
                             | <- 10.10.10.1/29
                            /'\
                           / \
                          / \
10.10.10.3/29 -> |SQUID-A| |SQUID-B| <- 10.10.10.2/29
(10.22.22.1/32) (10.22.22.1/32)

 Where the squid boxes have virtual interfaces configured. Now run a
routing protocol between the Cisco and the squid boxes (OSPF) and
do a "set ip next-hop 10.22.22.1" on the cisco. Make sure the squid
boxes don't peer with each other.

 Your next-hop doesn't need to be directly connected, providing you
have an IGP route to reach it with.

 If you use OSPF and set equal costs to reach the next-hop/32 it will
even load-balance for you, but you'll need to /not/ do per packet
balancing (i.e. run a route cache).

 You may need to mess with gated a tad to get the correct look to the
LSA it passes to the cisco.

 I haven't really thought about this - just written it down, but it
looks OK to me. In fact, it probably works better with the boxes on
different nets.

Cheers,

Lyndon

-- 
Penis Envy is a total Phallusy.
Received on Wed Mar 25 1998 - 06:33:58 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:39:27 MST