LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: Interception / redirection of traffic to transparent web cache (not

To: "LinuxVirtualServer.org users mailing list." <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: Interception / redirection of traffic to transparent web cache (not squid)
From: Mack.Joseph@xxxxxxxxxxxxxxx
Date: Thu, 10 Mar 2005 20:09:46 -0500
Joseph Mack PhD, High Performance Computing & Scientific Visualisation
LMIT, Supporting the EPA Research Triangle Park, NC 919-541-0007
Federal Contact - John B. Smith 919-541-1087 - smith.john@xxxxxxx

lvs-users-bounces@xxxxxxxxxxxxxxxxxxxxxx wrote on 03/10/2005 12:59:38
PM:



>
> |>Where I'm at:
> |>I understand that the 2.2 kernel series made this sort of thing
> |>relatively easy, but this is not the case with the 2.4 and
> |>2.6 kernels.
> |
> |
> | yes.
> |
>
> But there are workarounds are there not?  This is where I
> got lost in the available documentation.

you can get the functionality you want other (more complicated)
ways. There's no workaround to get the original functionality
(without rewritting large amounts of kernel, which you'll have
to test).


> I am aware that LVS' primary use is as a load balancer, but I also
> thought that it could behave much the way commercial L4
> switches do.  My
> hope was that I could direct traffic to the cache, but if the cache
> fails, the traffic will go straight to the net instead.

For LVS: use all realservers A,B,C...Z: if A fails then use all B,C...Z

I think you want

use only A: if A fails then use only B, if A recovers, use only A.

this is what hearbeat does (it's used in LVS to failover directors).

Joe




<Prev in Thread] Current Thread [Next in Thread>