ReplicaSets - recap

Question 1

What is a ReplicaSet

Answer 1

• runs multiple instance of a single pod
• provides high availability
• automates bringing up a single pod
• helps with load balancing and scaling

Question 2

What’s the difference between a ReplicaSet and Replication Controller

Answer 2

• Similar purpose
• Replication Controller is the older approach and ReplicaSet is newer

Question 3

How do I create a replication controller

Answer 3

• Create a yaml file with the required fields
• kind: ReplicationController
• spec must include template section

Question 4

What does template do in YAML config

Answer 4

Tells the ReplicationController how to create more pods with those specifications

Question 5

How do I create a ReplicaSet

Answer 5

• Create a YAML definition
• apiVersion must be apps/v1
• kind must be ReplicaSet
• spec > template (same as replication controller)
• replicas: number of desired replicas
• selector: helps ReplicaSet identify what pods fall under its control
• matchLabel: matches labels to the ones on the pod template

Question 6

Why do we need a “selector” in the ReplicaSet YAML definition

Answer 6

Because ReplicaSets can manage pods not created as part of the ReplicaSet (e.g., pods created before the ReplicaSet but which fall under the same selector label)

Question 7

Which labels have to match in the ReplicaSet YAML definition

Answer 7

select > matchLabels > tier
must match
template > metadata > labels > tier

Question 8

How do I scale ReplicaSets

Answer 8

• Update the replicas field
• $ k scale –replicas=<number> -f <yaml></yaml></number>

Question 9

Does running $ k scale <etc> change the YAML definition file</etc>

Answer 9

No. You must also:
* update the yaml
* $ k delete replicaset <name>
* $ k replace 0f <updated></updated></name>