Reputation: 1029
I am planning to test Kubernetes locally, but would like to ask some theoretic questions before.
I created a pipeline in python that takes as input a whole bunch of files from a directory, and created a docker image out of it (this is my Pod)
What I understood from the documentation is that the Kubernetes scheduler will choose automatically the minion to deploy for a given task, my question is, using an 8G memory laptop, is there a 'rule' to follow before creating the minion (specifying the number of minions to deploy) based on the amount of memory available in a machine (regardless if it is a laptop or a cluster) ?
Thanks
Upvotes: 1
Views: 298
Reputation: 312610
You would typically only ever have one minion/host. So if you are deploying your minions on physical hardware, there is a 1:1 mapping between minions and physical hosts.
If you are deploying into a virtual cluster on your laptop, you will want to make sure that each virtual minion has enough memory to run at least a single instance of whatever containers you plan on deploying. "How much is enough?" is a question that only you can answer.
Upvotes: 1