PORTWORX-ENTERPRISE-PROFESSIONAL practice question 6 of 15

A Portworx administrator wants to control which nodes will host a KVDB installation. What steps must an administrator take to ensure that KVDB installs on NODE01, NODE03, and NODE05?

Full set: 75 questions
Answer options
A: It is not possible to configure the location of the KVDB prior to installation.
B: Change the following in the 'StorageCluster' spec prior to installation: spec: kvdb: selector: matchNodeName: - NODE01 - NODE03 - NODE05 Correct answer
C: Label NODE01, NODE03, and NODE05 with 'px1/metadata-node=true' prior to installation.
Correct Answer:
B. Change the following in the 'StorageCluster' spec prior to installation: spec: kvdb: selector: matchNodeName: - NODE01 - NODE03 - NODE05
Explanation

To deterministically control which nodes host the internal Portworx Key-Value Database (KVDB), an administrator must define node selection criteria within the StorageCluster custom resource specification. The spec.kvdb section is designed for this purpose. By using a selector, such as nodeAffinity or a nodeSelector, the Portworx Operator can be instructed to schedule the KVDB pods exclusively on the specified nodes. While the exact syntax in the option may be a simplification, the principle of modifying the StorageCluster spec with a node selector is the correct and necessary step to enforce this placement.

Why Incorrect

A: This is incorrect. Portworx provides granular control over the placement of its components, including the internal KVDB, through the StorageCluster specification. C: Simply labeling nodes is insufficient. While labeling is a valid strategy, the StorageCluster spec must also be explicitly configured with a selector that targets those labels. The act of labeling alone does not influence KVDB placement.

References

1. Portworx Enterprise Documentation

"Specify nodes for internal kvdb": This official document details the methods for controlling KVDB placement. It explicitly shows how to use nodeAffinity within the spec.kvdb section to select nodes based on their hostnames (kubernetes.io/hostname)

which is the technically precise implementation of the goal described in the question. It also shows the alternative method of using a selector to target node labels.

Reference Location: Portworx Docs -> Install and Configure -> Install Portworx -> Install on Kubernetes -> Using the Operator -> Create the StorageCluster -> Specify nodes for internal kvdb.

2. Portworx Enterprise Documentation

"CustomResourceDefinition concepts": This document outlines the structure of the StorageCluster CRD. The kvdb field is shown to contain properties like nodeAffinity and selector

confirming that the StorageCluster object is the correct place to define KVDB placement rules.

Reference Location: Portworx Docs -> Reference -> Kubernetes -> CustomResourceDefinition concepts -> StorageCluster -> KvdbSpec.

Scroll to Top