> For the complete documentation index, see [llms.txt](https://docs.silk.us/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.silk.us/learn/data-management-concepts/non-disruptive-upgrade-ndu.md).

# Non-Disruptive Upgrade (NDU)

Understand how Silk upgrades maintain data availability, what to expect during maintenance, and how to prepare your environment.

Silk non-disruptive upgrades (NDUs) update platform software while maintaining application access to data. Silk Support performs and coordinates upgrades, including readiness checks, upgrade execution, and progress monitoring.&#x20;

The workflow upgrades Silk Flex first, followed by the DataPods. When d.nodes require an update, Silk applies their new images through a separate m.node live migration operation.&#x20;

{% hint style="warning" %}
**Data availability and management availability**

DataPods continue serving application I/O during upgrades. Flex management services are temporarily unavailable during the Flex upgrade, so operations like scaling and provisioning are unavailable during that period.&#x20;
{% endhint %}

## Prepare for an upgrade

Coordinate with Silk Support to confirm the target version, assess readiness, and schedule the upgrade at a suitable time.&#x20;

Before the upgrade, review these requirements:

<table><thead><tr><th width="222.9630126953125">Requirement</th><th>What to verify</th></tr></thead><tbody><tr><td>Cloud CPU quota</td><td>Sufficient quota is available for temporary or replacement resources required during the upgrade. </td></tr><tr><td>Supported configuration</td><td>Your environment meets the cluster and DataPod limits for the target release.</td></tr><tr><td>Flex version</td><td>Flex meets the version requirements for the DataPod upgrade. Silk Support upgrades Flex first when necessary.</td></tr><tr><td>DataPod health</td><td>All DataPod components are ready or healthy. Resolve existing issues before starting.</td></tr><tr><td>Host connectivity</td><td>Hosts follow Silk connectivity best practices and have redundant healthy connections to their DataPods.</td></tr><tr><td>Support access</td><td>Required remote access is available for Silk Support. Confirm the access method with Support before the upgrade. </td></tr></tbody></table>

For configuration limits, see [Silk Platform Configuration Limits](/learn/references/silk-platform-configuration-limits.md).&#x20;

Cloud permissions, network access, and available VM types can also affect upgrade readiness. Confirm the requirements for your cloud environment with Silk Support.&#x20;

## Understand the upgrade workflow

{% stepper %}
{% step %}

### Upgrade Silk Flex

Flex runs on a separate VM outside the DataPod data path. Upgrading Flex doesn't interrupt ongoing DataPod I/O.

Silk Support upgrades Flex to prepare the system components and container images required for the target release. Flex is typically unavailable for management operations for **15-20 minutes**, depending on the environment and upgrade scope.

The upgrade method depends on the cloud:

<table data-view="cards"><thead><tr><th></th><th></th></tr></thead><tbody><tr><td><strong>Azure</strong></td><td><p>A Blue-Green deployment provisions a new Flex VM alongside the existing VM. </p><p></p><p>After validation, an Azure Load Balancer directs traffic to the new instance. </p></td></tr><tr><td><strong>AWS and Google Cloud</strong></td><td>Flex is upgraded in place.</td></tr></tbody></table>

Silk Support normally uses an automated workflow. If a manual Flex upgrade is required, Support needs access to update repositories and container images, along with remote access to the Flex VM. Support creates a snapshot of the Flex VM before proceeding with a manual upgrade.&#x20;
{% endstep %}

{% step %}

### Upgrade DataPods

After Flex is upgraded, the SDP Management Console coordinates the DataPod upgrade. The workflow stages the required artifacts, updates the DataPod system components, and replaces c.nodes with nodes running on the target version.

Before and after each c.node upgrade, Flex checks that every host has more than one healthy connection to the DataPod. If redundant connections are not detected, Flex pauses the NDU until redundant connectivity is restored.

c.node upgrades typically takes **15-25 minutes per c.node**. The DataPod continues serving data, although redundancy may be temporarily reduced while individual c.nodes are replaced. Performance is expected to remain stable with minimal application impact.&#x20;

These timing estimates cover individual upgrade activities; they aren't an estimate for the entire environment. Confirm the expected maintenance duration with Silk Support.&#x20;
{% endstep %}

{% step %}

### Upgrade d.nodes when required

The DataPod NDU downloads updated d.node images, but doesn't apply them as part of the standard c.node upgrade workflow.

When required, d.nodes are upgraded through **m.node live migration** after the DataPod NDU completes. This separate workflow maintains data access while updating components in the data path.

Depending on operational requirements, live migration can follow the NDU immediately or take place in a later maintenance window. Confirm whether your upgrade requires this additional operation and when it should occur with Silk Support.&#x20;
{% endstep %}
{% endstepper %}

## Review Azure upgrade requirements

Azure Flex upgrades require permissions and networking that support provisioning a replacement Flex VM and transitioning access through an Azure Load Balancer.&#x20;

Before the upgrade, confirm with Silk Support that:

* [ ] Flex has the required permissions in the relevant resource groups and virtual network.
* [ ] The Flex subnet has explicit outbound connectivity for upgrade operations, telemetry, and communication with Silk services.
* [ ] The outbound connectivity design and Load Balancer integration are ready for the upgrade.&#x20;

Outbound connectivity can use a NAT Gateway, Azure Firewall, or another approved egress solution. During the Blue-Green upgrade, HTTPS and SSH access to Flex is routed through the Load Balancer.&#x20;

{% hint style="info" %}
**Planning a move to Managed Disk?**\
If you're migrating Azure DataPods from VM Local Disk to Managed Disk (Premium SSD v2), notify Silk Support before initiating NDU so the Azure availability set can be updated.&#x20;
{% endhint %}

Flex also provides an Azure environment validation check for connectivity to Silk Clarity and required Azure services and image resources. Validation runs during initial cluster creation and can be run on demand from the Flex UI.&#x20;

Validation failures don't automatically block Flex upgrades. Review and resolve identified issues with Silk Support, as they may affect upgrade operations or ongoing functionality.&#x20;

## &#x20;Understand full vs. module-based upgrades

Silk supports full NDUs or, when applicable, module-based NDUs (mNDUs).

{% tabs %}
{% tab title="Full NDU" %}
Updates DataPod system components, replaces c.nodes, and prepares node images for the target release.
{% endtab %}

{% tab title="Module-based NDU" %}
Updates selected management components independently without affecting I/O operations.&#x20;
{% endtab %}
{% endtabs %}

Module-based upgrades can reduce overall upgrade time, but some changes require a full NDU. Silk Support can help determine the applicable workflow for the update.&#x20;

{% hint style="warning" %}
mNDU is currently supported only on Azure.
{% endhint %}

## Upgrade completion and recovery

Silk Support configures and monitors full DataPod NDUs through Clarity. Work with Support to confirm completion and any follow-up operations, including m.node live migration.&#x20;

If an issue occurs, Support works to complete the upgrade to the target version. Returning a DataPod to a previous version requires escalation to Silk Engineering and may involve formatting capacity and re-seeding data.&#x20;

{% hint style="info" %}
On Azure, the Blue-Green upgrade process makes it easier to restore the previous Flex instance if needed.
{% endhint %}

## Schedule your upgrade

[Contact Silk Support](https://support.silk.us) to confirm the supported upgrade path, validate your environment, and schedule the upgrade. Review the target release notes beforehand to understand the changes included in the upgrade.

## Related resources

{% content-ref url="/pages/y4QzejE28yq2J0hdbKOc" %}
[Silk Platform Configuration Limits](/learn/references/silk-platform-configuration-limits.md)
{% endcontent-ref %}

{% content-ref url="/pages/jbwHJE9XlrTdPS1lXZgj" %}
[Clusters and DataPods](/learn/products-and-core-concepts/datapod-architecture/clusters-and-datapods.md)
{% endcontent-ref %}

{% content-ref url="/pages/Bdgy6vFNs8TdYaGxD5vF" %}
[Silk Flex Overview](/learn/products-and-core-concepts/silk-product-fundamentals/silk-flex-overview.md)
{% endcontent-ref %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.silk.us/learn/data-management-concepts/non-disruptive-upgrade-ndu.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
