Bootstrapping newly added nodes

[Mark Robson <ma...@gmail.com>]

Hi,

Looking at CASSANDRA-213 I see that there is now a procedure to "Bootstrap"
a newly added node so it will copy the relevant parts of data from other
nodes in the cluster.

This is a Good Thing.

What I do wonder is, why don't we automatically do this when you add a new
node to an existing cluster (when it comes up for the first time) ?

Also, what happens to queries which hit the node while it's getting the
data? Do they return "no data" or do they say "go away, try again", or is it
marked as unhealthy so that the other nodes won't query it?

Mark

Re: Bootstrapping newly added nodes

[Jonathan Ellis <jb...@gmail.com>]

see CASSANDRA-195

On Thu, Aug 13, 2009 at 5:27 AM, Mark Robson<ma...@gmail.com> wrote:
> Hi,
>
> Looking at CASSANDRA-213 I see that there is now a procedure to "Bootstrap"
> a newly added node so it will copy the relevant parts of data from other
> nodes in the cluster.
>
> This is a Good Thing.
>
> What I do wonder is, why don't we automatically do this when you add a new
> node to an existing cluster (when it comes up for the first time) ?
>
> Also, what happens to queries which hit the node while it's getting the
> data? Do they return "no data" or do they say "go away, try again", or is it
> marked as unhealthy so that the other nodes won't query it?
>
> Mark
>