Re: Loadbalance and setup plugins

From: Justin Riley <no email>
Date: Fri, 20 Jul 2012 16:14:02 -0400

Hey David,

Sorry to come in so late on this one. In any case Dustin is correct.
The load balancer calls each plugin's on_add_node(), in the order
defined in the config, with the following values:

on_add_node(self, node, nodes, master, user, user_shell, volumes)

The first argument (node) is the node to be added. The next argument
(nodes) is a list of *all* nodes (including new node, existing nodes,
and master), and the third argument (master) is obviously the master
node singled out.

Let me know if you have questions/issues.

HTH,

~Justin

On Fri, Jul 13, 2012 at 10:28:39AM -0700, David Erickson wrote:
> Thanks, ya confirmed.
>
> -D
>
> On 7/13/2012 3:48 AM, Dustin Machi wrote:
> > I believe it runs on_add_node() in this case, take a look at starcluster.clustersetup.DefaultClusterSetup.
> >
> > Dustin
> >
> > On Jul 12, 2012, at 4:41 PM, David Erickson wrote:
> >
> >> HI I am wondering how setup plugins work with load balancing? For
> >> example the signature of the run method for a ClusterSetup based plugin
> >> looks like:
> >>
> >> def run(self, nodes, master, user, user_shell, volumes):
> >>
> >> When loadbalance detects the need and adds a new node to a cluster, do
> >> all existing and the new node show up in the nodes argument above? I
> >> want to ensure that I am not re-initializing anything on my existing nodes.
> >>
> >> Thanks!
> >> David
> >> _______________________________________________
> >> StarCluster mailing list
> >> StarCluster_at_mit.edu
> >> http://mailman.mit.edu/mailman/listinfo/starcluster
> >
> >
>
>
> _______________________________________________
> StarCluster mailing list
> StarCluster_at_mit.edu
> http://mailman.mit.edu/mailman/listinfo/starcluster



Received on Fri Jul 20 2012 - 16:14:06 EDT