[rabbitmq-discuss] Upgrading erlang, rabbitmq in a cluster.

[Jean Paul Galea]

On 11/12/2012 07:17 PM, Matt Pietrek wrote:
> For what it's worth, we have a system here that allows us to upgrade
> clusters without downtime. It works for us, but may or may not be
> workable for other teams.
>
> Short synopsis is that our cluster is behind a VIP. When we want to
> upgrade a cluster, we spin up an alternate cluster and switch the VIP
> over to the alternate cluster. Meanwhile, we have tooling that moves
> messages between clusters. When the 'master' cluster's update is done,
> we reverse the process.
>
> This does require that your clients be tolerant to dropped connections.
> There's a lot of details not covered in the above description but it
> gives the basic gist of it.

We thought about this, but figured it's too much work for our small team 
to payoff. We can live with a few minutes downtime rather than having to 
manage a second cluster + shoveling messages.

Although interesting to know that there are people out there using this 
method to upgrade without downtime.