Re: heartbeat monitoring/failover list

From: Martin Hamilton <>
Date: Thu, 20 Aug 1998 21:26:10 +0100


Carlos Horowicz writes:

| Miquel van Smoorenburg wrote:
| > Well there's some info in the Linux High Availability HOWTO.. most
| > of it applies to a lot more systems than just Linux.

It's very good from a conceptual point of view, but we seem to be
missing the free software implementations. I know that several people
have been hacking at journalling/log-structured filesystems for Linux,
but haven't discovered anything which supports IP address takeover.

| Even at the server level, I think that having more than one server for each
| squid functionality (at the intranet or at the gateway level) ensures higher
| availability. Conclusion: am I missing the point ?

I think so - though my message was probably a bit too vague in the
first place!

If you can force all your WWW cache users to download a proxy
autoconfig script, there may not be a problem to solve in caching

The real question is - given the failure of a mission critical machine
or service, how to automatically get its IP address handed over to
another machine with a working installation of the service. Perhaps
it's easier to appreciate with something like a DNS server - although
resolvers are supposed to try multiple servers, you often get a nasty
pause whilst waiting for the first server in the search list to time

You can buy (for example) a nice Layer 4 switch from Alteon:
<URL:>, which will do
transparent failover, but this is going to cost you a lot of money. In
theory there's nothing preventing us from doing this in software on
Unix boxes using ARP. The Linux High Availability HOWTO explains how,
but it looks like nobody's gotten around to implementing it yet.
(unless you know better!)

Does that help to make things clearer ?



Version: 2.6.3i
Charset: noconv

Received on Thu Aug 20 1998 - 13:27:00 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:41:40 MST