20-OS UPGRADE on a node

--------------------------------------------------------------------------
OS UPGRADE on a node

if a node is down for more than 5 mins then the pods are terminated from that node, kuberenetes considers that node as dead and users wont be able to access those pods.

the time kubernerts waits for a pod to come back online is 5 mins if the pods wonty come back in 5mins then its evicted.

command to set the eviction time - kubectl-controller-manager --pod-eviction-timeout=5m0s

if we want to move all pods on a node to other nodes use command - kubectl drain <node name> - if we use this command , until we remove the restriction no new pods will be scheduled on this node

command to remove restriction - kubectl uncordon <node name>

another command to mark a node unschedullable - kubectl cordon <node name> - this wont effect the alredy running pods , only new pods wont be scheduled on the node 

this cordon and uncordon is for kube scheduler , even if node is cordained we can manually schedule the pods on the node 


--------------------------------------------------------------------------