Question 740
A production support team notices: A rolling application has three replicas but one node's failure leaves only one ready. Which mechanism BEST fits this evidence?
Explanation
Correct Answer
The decisive distinction is that replica count alone does not guarantee distribution across nodes. Confirm with a targeted check: a node failure leaves the planned number of available replicas.
Incorrect Answers
Committing manifest YAML alone does not change live cluster resources. That can explain why a GitOps deployment repository changes but the cluster is never reconciled, whereas the current evidence supports kubernetes desired replica state and scheduling constraints determine placement.
It would account for a Compose application starts its web container before its database is ready, not the current evidence. For this incident, replica count alone does not guarantee distribution across nodes; the alternative mechanism leaves that distinction unexplained.
Unconditional text appends introduce drift on repeated executions. That can explain why an Ansible playbook runs successfully twice but changes a file each time unnecessarily, whereas the current evidence supports kubernetes desired replica state and scheduling constraints determine placement.





CompTIA Linux+ (XK0-006) Practice Tests