[rabbitmq-discuss] Fwd: Re: Cluster Disk Node failure and cluster membership

Fri May 25 12:33:50 BST 2012

Forwarding to the list, didn't notice it was coming from google groups.

-------- Original Message --------
Subject: Re: [rabbitmq-discuss] Cluster Disk Node failure and cluster 
membership
Date: Fri, 25 May 2012 12:32:42 +0100
From: Francesco Mazzoli <francesco at rabbitmq.com>
To: Kevin Nuckolls <kevin.nuckolls at gmail.com>
CC: rabbitmq-discuss at googlegroups.com

Hi Kevin,

On 25/05/12 05:29, Kevin Nuckolls wrote:
> I'm hoping to automate the clustering steps with Chef and I have one
> pivotal question: If I have a cluster with three disc nodes in the cloud
> and one of them goes down and is gone forever, will the other nodes
> eventually give up on treating it as a disc node?

Not automatically. Right now only the node itself can leave the cluster
(reseting it will do). If the node is gone forever as you say, you can
bring up an "impostor" node that looks like the old one (that is, has
the same node name in distributed erlang), `force_cluster' it with the
remaining nodes, and then reseting it. I know, it's ugly, we're working
to make the whole system nicer.

Francesco.