VxRail: Upgrade Information

Helpful Resources:

• VxRail: How to Run the VxVerify Tool For detailed guidance on using the VxVerify tool to proactively identify issues before upgrading.
• How to upgrade VxRail Cluster from 4.7 Release (YouTube)
• Remote Proactive Support (RPS) General Procedure: VxRail-Upgrade - Customer-Preparation-Guide (Dell Support account is required to view article)
• Dell SolVe Online Portal (Procedure Generator, Dell Support Account required)

Regarding VxRail, what does Remote Proactive Services (RPS) Support?

• • Customers with ProSupport 4 Hour or ProSupport Plus Mission Critical maintenance contracts are eligible for a Dell RPS team to perform VxRail upgrades.
  • Customers with ProSupport Next Business Day maintenance are ineligible but may open a Service Request to request an upgrade subject to a service charge.
  • Standalone VxRail Upgrades with VxRail Managed (Internal) vCenter (See below for more information)
  • Stretched Cluster VxRail Upgrades (See below for more information)
  • VxRail Clusters with Customer Managed (External) vCenter deployments (External vCenter updates are outside Dell Support coverage, see below [No. 4] for more information).

What Type of Upgrades are performed on VxRail Systems?

• Major upgrade = An appliance upgrade which is out of family (example: 7.0.x > 8.0.x)
• Minor upgrade = An appliance upgrade which is in-family (example: 7.0.x > 7.0.x)

Are VxRail Upgrades Disruptive?

VxRail appliance upgrades are considered non-disruptive. The workload VMs are first migrated to other online hosts before placing that host into maintenance mode.
Be aware of the following:

• • Prior to a VxRail upgrade health checks are performed to ensure the system's upgrade readiness.
  • All workload VMs should be configured to allow them to vMotion between VxRail nodes. If a virtual machine (VM) cannot be vMotioned, then planned downtime may be required to allow an upgrade to complete successfully.
    As with any system upgrade, it is recommended to upgrade at times of low I/O and that a maintenance window is scheduled should there be any difficulties.
  • If the system utilization is too high (Memory or CPU usage), then it may be required that some VMs be powered off, to avoid any potential impact to the workload VMs.
  • If the storage utilization is too high, it may be required that you free up space on the vSAN OR migrate some workload VMs off the vSAN onto a different system (if possible).

During a VxRail upgrade, the following components are upgraded (but is not limited to) the following:

• • VxRail Manager
  • VMware ESXi
  • VMware vCenter Server Virtual Appliance (vCSA) (If Internal)
  • VMware Platform Service Controller (PSC) (If Internal)
  • VMware vSAN
  • Hardware-specific upgrades included (Not included on every upgrade but is present within the composite package where applicable).
  • The items below are some of what is upgraded, see the VxRail Support Matrix links to determine component versions between different VxRail code levels.

Dell PowerEdge Platform:

• • PTAgent
  • iSM
  • BIOS
  • iDRAC
  • BOSS

For a full list of upgraded components, see the required VxRail Support Matrix below:

• • VxRail 8.0 Support Matrix
  • VxRail 7.0 Support Matrix
  • VxRail 4.7 Support Matrix
  • VxRail 4.5 Support Matrix

Useful points regarding VxRail appliance upgrades:

• • VxRail upgrades should always be initiated from the VxRail Manager UI/vCenter Plugin.
  • Updating VxRail using the VMware vCenter Update Manager (VUM on vCenter web client) is not supported.
  • VxRail clusters that are deployed with an External vCenter may require the vCenter Server to be updated prior to the VxRail Cluster upgrade (External vCenter upgrades are not supported by Dell and are a user task).
  • For External (User managed) vCenter configuration, it is imperative that you ensure the current version of vCenter is supported on the code level you are upgrading to. See the following article on Dell VxRail: VxRail and Customer-managed VMware vCenter Server (External vCenter) Interoperability Matrix (Dell Support login required to view article)
  • Stretched VxRail clusters running versions (lower than) 4.5.215 or 4.7.300 are not user upgradable, and a Dell Technical Support Engineer (TSE) must perform the upgrade. Upgrades above these code levels are considered user upgradable and the upgrade script automates the Witness update.

How long does or should I expect the VxRail upgrade to take?

The duration of a VxRail appliance upgrade depends on many factors:

• • Number of Hosts
  • Host Hardware Type
  • If DRS is enabled or not
  • System Configuration (stretched or nonstretched and internal or external vCenter)
  • Number of VMs on the system
  • Size of VMs
  • Current Code Level and Target Code Level (major or minor)
  • How many additional updates were included on the destination version (firmware updates and so forth).

As a general guideline, we can use the following as a rough estimate, sometimes, the activity duration can take longer and sometimes it can take less.

Dell PowerEdge hosts:

• • Estimated time per host = 2 hours
  • If DRS is not enabled, add an extra 1 hour to the total time.
  • If the system is a stretched cluster, add an extra 1 hour to the total time.

Self-Upgrade Process Summary

Pre-Upgrade

• • The cluster requires enough resources to operate with one node in maintenance mode at a time. If the cluster is above 90% utilization for any resources, while a node is in maintenance mode, it is recommended to power off nonessential virtual machines to free up memory and CPU. 
  • If the upgrade is a major upgrade, verify the licenses are upgradeable following: Upgrade and Downgrade VMware License Keys via the Broadcom support portal.
    Do not upgrade the licenses yet but verify the availability.
  • Log in to your support account and use Solve to generate the upgrade procedure. During Solve generation, it provides recommended articles to review based on your selections.
    Select Upgrade > Software Upgrade
  • Run VxVerify(Steps also in the Solve upgrade procedure). VxVerify performs various prechecks for issues that can impact an upgrade such as platform service functionality, free space on VxRail Manager, DNS/NTP, database checks, iDRAC checks. VxVerify provides an article number to search on the support site for any errors it finds.

Upgrade

• • Once VxVerify passes, follow the Solve procedure to take snapshots of service machines and upgrade the cluster.

Post-Upgrade

• • Verify ESXi and vCenter licenses are assigned and not showing trial versions. Upgrade and apply licenses as needed.
  • Review cluster health after the upgrade
    Skyline HealthCheck - About the vSAN Skyline Health
    VxVerify post check - VxRail: How to Run the VxVerify Tool

RPS Upgrade Process Summary

Tech Consult (Pre-Check):

• • Perform cluster health checks using scripts and vCenter UI checks (RPS open a support case for any issues).
  • Size the environment to provide an upgrade plan
  • Provide an estimated upgrade duration (this is based on an issue-free upgrade)
  • Provide the Customer with any prerequisites to be actioned.
  • Discuss the upgrade plan and provide the code download link.
  • Answer any user technical questions.

Once the Tech Consult has been completed, the Scheduling Team works with the user to schedule the upgrade based on the upgrade plan. This allows RPS resources to be assigned for the entire upgrade duration and provides handovers where the duration is crossing engineer shifts.

Upgrade:

• • Perform cluster health checks using scripts and vCenter UI checks.
  • Upload the code package.
  • Start the upgrade process.
  • Monitor to completion.

Post Checks:

• • Ensure the VMware vSphere 7.0 & 8.0 license keys are not expired or are running as the 60-day trial version.
  • Review cluster health after the upgrade
    Skyline HealthCheck - About the vSAN Skyline Health
    VxVerify post check - VxRail: How to Run the VxVerify Tool
  • Open a support Service Request for any issues that need investigation.

VxRail Upgrade Procedure

The VxRail is an engineered solution and VxRail Life-Cycle Management (LCM) automates the VxRail upgrade process by loading software and firmware to each node.

• • Configure alarm definitions for email notifications.
  • Take snapshots of service VMs.
  • Upgrade from the current code to a new version.
  • Remove snapshots of service VMs.
  • Change the Mystic Service account password.

VxRail Upgrade Script

A script automates the VxRail Life-Cycle Management (LCM) in the following order.

• • VxRail Manager
  • vCenter Server (if VxRail Managed)
  • ESXi node 1
  • ESXi node 2
  • ESXi node 3
  • ESXi node n

The LCM script starts at node 1, evacuates its VMs, and places the node in Maintenance Mode. All components are upgraded, and the node is rebooted. After the reboot is completed, the node is taken out of Maintenance Mode and VMs migrate back to the node. The script moves to the next node, until all nodes in the cluster have been upgraded.

Watch this video on How to run the VxRail Verify Tool.

Watch this video on How to Upgrade VxRail Cluster from 4.7 Release for VxRail.