NCI 7.5 Deep Dive – Part One: VM Startup Policies

Nutanix has released NCI 7.5, a major upgrade that brings improvements across the entire platform including AOS, AHV, Flow, DR, Prism, FC, NC2, and the broader NCP ecosystem. It also rolls up all fixes from earlier releases through AOS 7.3.1.2, giving customers a stable, modern foundation to build on.

In this first post of the series, we’ll cover VM Startup Policies which is a powerful new feature that gives administrators fine-grained control over VM boot sequencing during HA events and full cluster restarts.

Smarter Recovery with VM Startup Policies

Most enterprise applications are made up of multiple VMs including databases, middleware, front-end services, identity systems, and more. If these components don’t start in the right sequence after a node or cluster restart, you can easily end up with service failures or extended downtime.

With AOS 7.5 and pc.7.5, Nutanix introduces VM Startup Policies, allowing you to:

• Define VM dependencies and startup order

• Ensure each layer of the application stack comes online at the right time

• Reduce downtime caused by out-of-order service recovery

• Combine multiple policies into a single dependency chain

How It Works

VM Startup Policies use VM categories to define dependency groups. For example:

• Policy 1 Category B depends on Category A

• Policy 2 Category C depends on Category B

AOS builds the sequence automatically:

1. Start Category A

2. Once ready, start Category B

3. Then start Category C

This ensures critical components like databases or domain controllers are available before higher-level services start.

Start Conditions for More Precise Control

Each dependency group includes a start condition to determine when the next tier can begin booting:

1. Power On

The group is considered ready as soon as all VMs are powered on and does not wait for the operating system or applications to become available.

2. Guest Bootup

Using Nutanix Guest Tools (NGT), AOS verifies that the guest OS has fully booted. Prism Central raises alerts if a VM fails to start or if dependent VMs must be skipped.

Timing Parameters

• Delay – Adds a waiting period after the condition is met.

• Timeout – Sets how long AOS waits for NGT to detect Guest Bootup.

These controls help align recovery behavior with application requirements.

Handling Policy Conflicts Automatically

AOS 7.5 resolves situations where policies overlap or conflict with each other.

Start Condition Conflicts: If two policies specify different start conditions for the same dependency, AOS automatically honors the policy created first and marks the other as conflicting.

Dependency Conflicts: If a VM ends up in a circular dependency (for example, appearing in both Category A and Category C), AOS again gives precedence to the first valid rule and flags the conflicting entry.

 To create VM Start Policy, the workflow is simple:

1. Log in to Prism Central

2. Open Infrastructure → Compute → Policies → VM Startup Policies

3. Click Create VM Startup Policy

4. Add a name and description

5. Define dependency groups and categories

6. Choose start conditions, delay, and timeout

7. Save the policy

You can create up to 20 policies, each with six dependency levels and up to five categories per level.

Conclusion

VM Startup Policies in AOS 7.5 offer a smarter, more predictable way to recover multi-tier applications. By defining clear dependencies and start conditions, administrators can ensure that VMs come online in the correct sequence during HA events or cluster restarts, reducing risk, minimizing downtime, and improving operational confidence.

Stay tuned for Part Two, where we’ll explore more new capabilities introduced in NCI 7.5