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

Re: Per node pod quotas by pod service type

We generally describe this as pod scheduling affinity and anti affinity (the part where three of one service are deployed to the same node).  Openshift 1.0-1.4 uses cluster wide defaults for this, while 1.5 will have alpha support for per pod level rules to let you say, for instance, "never have more than one X on the same node".   That way, it's up to each app author to define their own rules.

On Mar 3, 2017, at 10:24 AM, Ronan O Keeffe <ronanok donedeal ie> wrote:


Considering the following scenario: One service per pod, 4 services, 4 nodes. So 16 pods across 4 nodes. 

Pods get deployed across the nodes and Openshift by and large does a great job of making sure things are balanced, however sometimes we end up with a scenario where three of one service get deployed on the one node. 

Just wondering can you define a quota to allow only one of each pod type per node? 

For example if you had 4 email pods as part of your overall service you'd want one on each node. 

I can see plenty of ability to set quotas on CPU, RAM, storage etc,  but none on limiting pod type per node. 

users mailing list
users lists openshift redhat com

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