Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 30 Jun 1998 10:32:34 +0800 (CST)
From:      Bo Fussing <bmf@gateway.net.hk>
To:        Tim Tsai <tim@futuresouth.com>
Cc:        Luis Munoz <lem@cantv.net>, Evren Yurtesen <yurtesen@ispro.net.tr>, freebsd-isp@FreeBSD.ORG
Subject:   Re: cisco
Message-ID:  <Pine.LNX.3.96.980630102245.20371F-100000@gate.gateway.net.hk>
In-Reply-To: <19980629203659.50524@futuresouth.com>

next in thread | previous in thread | raw e-mail | index | archive | help
>   I suppose with some hardware help you can cut (open circuit) the ethernet
> connection or simply power down the failed primary server.

If you are running a HA setup you will need someform of backup server that
can step in if the primary dies. In this case you can use a second machine
running a backup proxy and monitoring the primary proxy. Should it fail it
would assume the same IP address using ARP. For this to work on a
multi-purpose server you can allocate an IP address specifically for the
proxy service and run multiple IPs on the ethernet interface.

I would assume that should the primary proxy fail you would try to revive
it on the primary server, if this fails you should delete the ARP entry
for the IP address and publish a new one for the backup server NIC using
the same IP address.

Would like to hear your thoughts on such a scheme.

Bo


> 
> > Another is to have some remote monitoring of the proxy and notification to
> > the system admin (by pager) of the failure.
> 
>   We do that but that's hardly what I'd call redundancy.  :-)


On Mon, 29 Jun 1998, Tim Tsai wrote:

> On Tue, Jun 30, 1998 at 09:22:18AM +0800, Bo Fussing wrote:
> > One way to counter this on the server side is to have some form of
> > backup server standing by and stepping in when the normal proxy dies.
> 
>   This is assuming that the primary server completely dies (does not
> answer to the IP or MAC address) - which is not necessarily the case
> (i.e. perhaps only the proxy application died and not the server).
> 
>   With policy route you can set the next-hop-ip address but the question
> becomes how do you get the backup server to answer to that IP address
> correctly, even if the primary server still answers to that IP in someway.
> All the setup's I know of involves another additional level (either a
> hardware switch that supports virtual IP's or something like Cisco's
> localdirector) and the actual proxy servers sit behind this localdirector
> type device.
> 
> 
>   Tim
> 
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-isp" in the body of the message
> 


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-isp" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.LNX.3.96.980630102245.20371F-100000>