Re: What if Squid goes down?

I already use cutover systems, where I have the main proxy (with large cache)
and a redundant (two of them, actually, with minimal caches...just a few MB).

Each machine has two addresses (standard linux aliasing). If the main box
goes out, or stops giving service, the redundant machine 'picks up' the main
machine's primary IP address (essentially giving it a third address) and
takes over it's job until it determines (by testing the main machine's
secondary address) that service has been restored. When it has, the secondary
lets go of the primary address, and lets the original box take over again.

The software I cobbled up to do this is outrageously crude, and of little
better quality than rubbing two sticks together. At some point I'll do
something more elegant than shell scripts, ping and echoping.

D

Ong Beng Hui wrote:

> > A simple, expensive and inelegant solution would be to have two
> machines, both with the same IP address, one connected to your LAN, the
> other on standby (physically disconnected from your network). If you have
> problems, unplug one and plug in the other...(Im sure there are other far
> more elegant solutions)
>
> Some guys in Hong Kong Digital did a Highly Available
> Proxy setup with shared disks, automatic failover
> facility for some big company in Hong Kong. A HA
> proxy is possible. A DNS round robin can be use to
> "load balance" the two servers.
>
> But I think a more elegant solution will be to use ICP
> as a heart beat between servers, and use ICP to fail
> over proxy services.

--
Note to evil sorcerers and mad scientists: don't ever, ever summon powerful
demons or rip holes in the fabric of space and time. It's never a good idea.
ICQ UIN: 3225440