April 2018
Intermediate to advanced
468 pages
14h 34m
English
Most of the work of the scheduler is pretty generic—it figures out which pods need to be scheduled, updates their state, and runs them on the selected node. The custom part is how to map pods to nodes. The Kubernetes team recognized the need for custom scheduling, and the generic scheduler can be configured with different scheduling algorithms.
The main data type is the Scheduler struct that contains a Config struct with lots of properties (this will soon be replaced by a configurator interface):
type Scheduler struct {
config *Config
}
Here is the Config struct:
type Config struct { SchedulerCache schedulercache.Cache Ecache *core.EquivalenceCache NodeLister algorithm.NodeLister Algorithm algorithm.ScheduleAlgorithm GetBinder ...Read now
Unlock full access