Stateful Set Deployment Failver Scenarios


Just want to know how the stateful set pod failover will happen in kubernetes cluster…
I have one cluster consists of 11 vms and 8 vms are worker nodes . Due to some problem in VM one worker went down … where my stateful set was running with localpath provisioner. My pvc attached to same host since we do not have any cloud storage.

When the VM down, kubernetes will do pod failover to other VM ?

Did replicaset concept applies here ? My pod went to unavailable state since its was not moved to other vm succesfully. Is the idle behaviour?


Hi Gurukiran,

First I am answering the question “Did replica set concept applies here ?”

Statefulset is used to manage stateful application, unlike the resource of type deployment (which creates the replicasets and manages scaling of a set of pods) Statefulset keeps ordering and uniqueness of these set of pods.
Deployment Controller is creating a replicaset to keep " Current state of cluster to desired state based on the deployment manifest file you created". The Statefulset is also a controller but it is not creating any replicaset but creating pod with its own naming convention ( In resource type deployment the pod name will be like --). Every replica of a statefulset have it’s own state. for example a statefulset with two replicas will be creating two pods with its own pvc, ie is two pods and two pvc .


Hi Pradeep,

Thank you very much.

Is there any alternative for stateful sets if node goes down to have the high availability.
My application was recommenced to use stateful set by vendor so we have to go with that. . My application uses raft protocol which is having master and leader concept and partitions. So if any pod missing in the cluster it wont behave as normal sometimes it stop processing requests . To overcome is there any solution ?