[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: OpenShift Internals


https://github.com/openshift/crankcase/blob/master/gearchanger/mcollective/plugin/lib/gearchanger-mcollective-plugin/gearchanger/mcollective_application_container_proxy.rb #self.find_available_impl will find the node to create the gear on.
https://github.com/openshift/crankcase/blob/master/gearchanger/mcollective/agent/facts/stickshift_facts.rb assembled the facts which are used in the calculations.


On Sep 28, 2012, at 3:56 PM, Gandalf Corvotempesta wrote:

2012/9/29 Krishna Raman <kraman gmail com>

During scaling up, the broker will look at nodes that support the specified gear size and create a new gear on the node with the most capacity.
You cannot choose an explicit node while scaling.

Can you point me to this algoritm in sources?
I would like to customize it by adding a more check that will always try to use as much nodes as possibile for the same app. Because, if I understand you properly, If I have 5 nodes with free space only for 1 gear more, and a sixth node running the main application gear and 5 free "slots", openshift will scale the app on the same node.

This is no good, an app should be scaled over multiple host, if possibile. 

dev mailing list
dev lists openshift redhat com

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]