[users@httpd] Guidance on Fail Over for Load Balancers

[Michael Duffy <md...@gmail.com>]

I did search the archives of this list before posting.

We are going to use the Apache HTTP Server as a load balancer for a cluster
of Tomcat servers.

We do not want the load balancer to be a single point of failure.

Is it possible to cluster or set up fail over for the load balancers?  Any
guidance would be appreciated.

There are of course expensive corporate solutions available.  We would like
to stay with Apache if we can.

Kind regards,

Mike

Re: [users@httpd] Guidance on Fail Over for Load Balancers

[Yehuda Katz <ye...@ymkatz.net>]

You could set up two HTTPD servers with a shared IP address (using VRRP
with keepalived).
In that case, I highly recommend using a configuration management tool (for
example, puppet or chef) to make sure both systems have the same settings.

If the two load balancers need to share session information (for example,
for visitor pinning) then it gets a bit more complicated.

- Y

On Mon, Jan 23, 2017 at 6:32 PM, Michael Duffy <md...@gmail.com> wrote:

> I did search the archives of this list before posting.
>
> We are going to use the Apache HTTP Server as a load balancer for a
> cluster of Tomcat servers.
>
> We do not want the load balancer to be a single point of failure.
>
> Is it possible to cluster or set up fail over for the load balancers?  Any
> guidance would be appreciated.
>
> There are of course expensive corporate solutions available.  We would
> like to stay with Apache if we can.
>
> Kind regards,
>
> Mike
>
>