After the HCI upgrade is completed, nodes will restart to have the new version take effect. To use the ECC memory isolation feature, please manually restart the nodes and modify the BIOS options (enable ECC scan and ECC isolation). Note: Forced shutdown is not recommended before the upgrade is completed.

Before upgrading, please use the latest version of aDeploy (refer to Chapter 2.1.1 Upgrade Packages, Documents, and Tools for the download link) to perform the pre-upgrade check and install a pre-upgrade check package.
Please contact a Sangfor technical support representative for pre-upgrade evaluation and check if the following conditions are met:

• The current version is earlier than HCI5.8.6 and needs to be upgraded to a version later than HCI5.8.6.

• The customer uses the platform for high-performance database services and is demanding performance.

• More than ten nodes are involved.

Before upgrading, please check whether the license key is valid.
When graphics cards (including Passthrough and vGPU) are in use, ensure the IOMMU is enabled and restart the node.
If graphics cards are used in versions earlier than 6.2.0, aGPU license key is required after the upgrade. Otherwise, GPU-related features cannot be used.
HCI cannot be upgraded when it provides services as an iSCSI server. Before upgrading, please go to iSCSI clients to disconnect all related connections.
During the pre-upgrade check for the offline upgrade, if "A VM is running on the node. Please shut down the VM and try again." pops up, you can ignore it temporarily. Before an official upgrade, all virtual devices and VM need to be shut down, and a pre-upgrade check should be performed again.
If the current version contains custom packages (Custom-XXX, as shown in the figure below), please get in touch with a Sangfor technical support representative for upgrade evaluation.

Text Description automatically generated

If the current version contains SPs (sp-XXX packages), it can be upgraded directly. If the SPs cause the upgrade failure, please get in touch with Sangfor Technical Support for assistance.
If the upgrade is performed when the storage network is not configured for computing nodes in the cluster, the pre-upgrade check will fail. The upgrade can only be performed after the storage network is configured.
Suppose VDI is required to upgrade because the bad sector check mechanism for disks before the VDI upgrade is inconsistent with the new version. After the upgrade is completed, various alerts may be triggered for virtual storage disks.

{{ $t('index.defaultHeader.logo') }}

Hyper Converged Infrastructure (HCI/aSV)