Free NCP-MCI-6-10 sample questions

Correct answers: Boot; Home; Data

A manual deployment of a Prism Central VM requires three separate disk image files: the boot disk, the home disk, and the data disk. These files contain the operating system, user data, and metadata/service data for the Prism Central instance, respectively. Swap space is managed within the OS and does not require a separate image file.

Correct answer: Some data exists as a single copy and the cluster cannot tolerate another failure.

With Replication Factor 2, data is written in two copies across different nodes. When one node fails, the data that was on that node now only has one remaining copy on another node in the cluster. The cluster will immediately begin recreating the second copy, but until that process completes, that data is at risk. If another node or disk holding the single remaining copy fails during this time, data loss will occur. The cluster remains fully operational (read/write).

Correct answer: Critical

In Nutanix Prism, 'Critical' is the highest severity level for alerts. It signifies a serious issue that could lead to a production outage, data unavailability, or data loss. Therefore, to automate ticket creation for the most severe issues, the 'Critical' severity level should be used as the trigger in the X-Play playbook.

Correct answer: True

This is True. The Witness is designed to handle this specific split-brain scenario. If the secondary site can reach the Witness but the primary site cannot (due to a primary site failure), the Witness will instruct the secondary site to take ownership of the protected container. This ensures business continuity even during a complete primary site outage.

Correct answer: Use Prism Central to define a VM category for the database VMs and apply a Quality of Service (QoS) policy to guarantee CPU cycles and IOPS.

This is the most comprehensive and flexible solution. By categorizing the VMs and applying a QoS policy in Prism Central, the administrator can set a guaranteed number of CPU cycles (using CPU reservation) and a minimum IOPS value for the database VMs. This prevents resource starvation from 'bully' VMs without statically assigning them to hosts, thus maintaining cluster flexibility. Flash Mode is already active on an all-flash cluster, and reserving storage capacity doesn't guarantee performance. VM-Host affinity is too rigid and violates the 'not pinning VMs' requirement.

Correct answer: The LCM process automates all these steps.

The Life Cycle Manager (LCM) framework is designed to automate the entire upgrade process in a non-disruptive, rolling fashion. When an administrator initiates a firmware update through LCM, the framework automatically handles putting the node into maintenance mode, live-migrating VMs, shutting down the CVM if necessary, applying the updates, restarting services, and exiting maintenance mode before moving to the next node. Manual intervention as described in the other options is not required and circumvents the built-in safety and automation of LCM.

Correct answer: Create a new bridge (e.g., br1), create a new bond (e.g., bond1) with the 1GbE NICs, and add the bond to the new bridge.

The correct and recommended approach for creating a logically and potentially physically separate network is to create a new virtual bridge (e.g., br1). This new bridge will then have its own bond (e.g., bond1) composed of the 1GbE physical NICs. VMs that require access to the backup network can then have a vNIC attached to a virtual network created on br1. This ensures complete separation from the primary br0 network. Adding NICs of different speeds to the same bond is not recommended, and adding a second bond to the same bridge does not provide the desired logical separation.

Correct answer: The test network (VLAN) specified in the recovery plan is not configured or available at the recovery site.

A 'Test Failover' creates a temporary, isolated network bubble for the recovered VMs. The administrator must specify a test network (VLAN) in the recovery plan's network mappings. If this specified VLAN does not exist or is not configured on the hosts at the recovery site, the VMs cannot be attached to a network and will fail to power on. While replication issues can cause failures, a missing test network is a very common and specific cause for this symptom during a test failover.

Correct answers: Oplog Write IOPS; Extent Store Write IOPS

The Nutanix Oplog is a persistent write buffer that resides in the CVM's DRAM and is backed by the SSD tier. 'Oplog Write IOPS' represents the total number of write I/O operations being sent to the Oplog. 'Extent Store Write IOPS' represents the I/O operations that have been de-staged from the Oplog to the persistent SSD storage (the Extent Store). Comparing these two metrics shows how much write I/O is being absorbed by the fast Oplog versus being written to the slower (but still fast) SSD tier.

Correct answer: Snapshots with Redirect-on-Write

Nutanix cloning and snapshot technology is based on a Redirect-on-Write (ROW) algorithm. When a VM is cloned, it is essentially a writable snapshot. The clone initially shares all the same data blocks as the parent image and consumes no extra physical space. Only when new data is written to the clone are new blocks allocated, and metadata pointers are updated to point to these new blocks. This makes cloning extremely fast and space-efficient, which is ideal for VDI deployments.