Starting from vSphere 7.0, software FCoE is deprecated, and in vSphere 8.0 software FCoE profiles are not supported. If you try to apply a host profile from an earlier version to an ESXi 8.0 host, for example to edit the host customization, the operation fails. In the vSphere Client, you see an error such as Host Customizations validation error.

Validation of existing host profiles for ESXi versions 7.x, 6.7.x and 6.5.x fails when only an 8.0 reference host is available in the inventory.

If the peer physical switch of a VMNIC does not support Media Auto Detect, or Media Auto Detect is disabled, and the VMNIC link is set down and then up, the link remains down after upgrade to or installation of ESXi 8.0.

In vSphere 8.0, vCenter Single Sign-On automatically renews a VMCA-generated STS signing certificate. The auto-renewal occurs before the STS signing certificate expires and before triggering the 90-day expiration alarm. However, in long-running upgrade or remediation tasks by using a vSphere Lifecycle Manager image on multiple ESXi hosts in a cluster, vSphere Lifecycle Manager might create a cache of STS certificates internally. In very rare cases, if an STS certificates refresh task starts in parallel with the long-running upgrade or remediation task, the upgrade task might fail as the STS certificates in the internal cache might be different from the refreshed certificates. After the upgrade task fails, some ESXi hosts might remain in maintenance mode.

Some module parameters for the nmlx5_core driver, such as device_rss, drss and rss, are deprecated in ESXi 8.0 and any custom values, different from the default values, are not kept after an upgrade to ESXi 8.0.

When you use the vCenter Server GUI installer or the vCenter Server Appliance Management Interface (VAMI) to restore a vCenter from a file-based backup, the restore workflow might freeze at 90% with an error 401 Unable to authenticate user, even though the task completes successfully in the backend. The issue occurs if the deployed machine has a different time than the NTP server, which requires a time sync. As a result of the time sync, clock skew might fail the running session of the GUI or VAMI.

If an active virtual machine has a PCI passthrough to a DPU at the time of shutdown or reboot of an ESXi host, the host fails with a purple diagnostic screen. The issue is specific for systems with DPUs and only in case of VMs that use PCI passthrough to the DPU.

Although the vSphere Client allows the operation, if you disable IPv6 on an ESXi host with DPUs, you cannot use the DPUs, because the internal communication between the host and the devices depends on IPv6. The issue affects only ESXi hosts with DPUs.

In rare cases, HPE servers with pre-installed Pensando DPU might take more than 10 minutes to reboot in case of a failure of the DPU. As a result, ESXi hosts might fail with a purple diagnostic screen and the default wait time is 10 minutes.

Starting with vSphere 8.0, in both Integrated Dell Remote Access Controller (iDRAC) and HP Integrated Lights Out (ILO), when you have an USB interface enabled, vUSB or vNIC respectively, an additional standard switch vSwitchBMC with uplink vusb0 gets created on the ESXi host. This is expected, in view of the introduction of data processing units (DPUs) on some servers but might cause the VMware Cloud Foundation Bring-Up process to fail.

NVIDIA BlueField DPUs must be in hardware offload mode enabled to allow virtual machines with configured SR-IOV virtual function to power on and operate.

Vendor Device Groups, which enable binding of high-speed networking devices and the GPU, are supported only on VMs with HW version 20 and later, but you are not prevented to configure a VM at HW version earlier than 20 with a Vendor Device Group. Such VMs might not work as expected: for example, fail to power-on.

Moving vSphere plug-ins to a remote plug-in architecture, vSphere 8.0 deprecates support for local plug-ins. If your 8.0 vSphere environment has local plug-ins, some breaking changes for such plug-ins might cause the pre-upgrade check by using VAMI to fail.

In the Pre-Update Check Results screen, you see an error such as:

Warning message: The compatibility of plug-in package(s) %s with the new vCenter Server version cannot be validated. They may not function properly after vCenter Server upgrade.

Resolution: Please contact the plug-in vendor and make sure the package is compatible with the new vCenter Server version.

Although by default when you enable CPU Hot Add to allow the addition of vCPUs to a running virtual machine, virtual NUMA topology is deactivated, if you have a PCI passthrough device assigned to a NUMA node, attempts to remove the device end with an error. In the vSphere Client, you see messages such as Invalid virtual machine configuration. Virtual NUMA cannot be configured when CPU hotadd is enabled.

Pensando Distributed Services Platform (DSC) adapters have 2 high speed ethernet controllers (for example vmnic6 and vmnic7) and one management controller (for example vmnic8):

:~] esxcfg-nics -l

vmnic6 0000:39:00.0 ionic_en_unstable Up 25000Mbps Full 00:ae:cd:09:c9:48 1500 Pensando Systems DSC-25 10/25G 2-port 4G RAM 8G eMMC G1 Services Card, Ethernet Controller

vmnic7 0000:3a:00.0 ionic_en_unstable Up 25000Mbps Full 00:ae:cd:09:c9:49 1500 Pensando Systems DSC-25 10/25G 2-port 4G RAM 8G eMMC G1 Services Card, Ethernet Controller

:~] esxcfg-nics -lS

vmnic8 0000:3b:00.0 ionic_en_unstable Up 1000Mbps Full 00:ae:cd:09:c9:4a 1500 Pensando Systems DSC-25 10/25G 2-port 4G RAM 8G eMMC G1 Services Card, Management Controller

The high-speed ethernet controllers vmnic6 and vmnic7 register first and operate with RSS set to 16 receive queues.

:~] localcli --plugin-dir /usr/lib/vmware/esxcli/int networkinternal nic privstats get -n vmnic6…Num of RSS-Q=16, ntxq_descs=2048, nrxq_descs=1024, log_level=3, vlan_tx_insert=1, vlan_rx_strip=1, geneve_offload=1 }

However, in rare cases, if the management controller vmnic8 registers first with the vSphere Distributed Switch, the high-speed ethernet controllers vmnic6 or vmnic7 uplink might end up operating with RSS set to 1 receive queue.:~] localcli --plugin-dir /usr/lib/vmware/esxcli/int networkinternal nic privstats get -n vmnic6…Num of RSS-Q=1, ntxq_descs=2048, nrxq_descs=1024, log_level=3, vlan_tx_insert=1, vlan_rx_strip=1, geneve_offload=1 }

As a result, you might see slower performance in native mode.

If you deploy a virtual machine from an OVF file or from the Content Library, instead of ESXi automatically selecting the number of cores per socket, the number is pre-set to 1.

If you have the vSphere Distributed Services Engine feature with a Pensando DPU enabled on your ESXi 8.0 system, you cannot set the Maximum Transmission Unit (MTU) on a vSphere Distributed Switch to a value larger than 9174.

When two NICs that use the ntg3 driver of versions 4.1.3 and later are connected directly, not to a physical switch port, link flapping might occur. The issue does not occur on ntg3 drivers of versions earlier than 4.1.3 or the tg3 driver. This issue is not related to the occasional Energy Efficient Ethernet (EEE) link flapping on such NICs. The fix for the EEE issue is to use a ntg3 driver of version 4.1.7 or later, or disable EEE on physical switch ports.

An intermittent timing issue on the ESXi host side might cause NSX installation or upgrade in a vSphere environment with DPUs to fail. In the nsxapi.log file you see logs such as Failed to get SFHC response. MessageType MT_SOFTWARE_STATUS.

If you enable or disable SR-IOV with the icen driver on an ESXi host and configure a transport node in ENS Interrupt mode, some Rx (receive) queues might not work if you do not reboot the host. As a result, some virtual machines might not get DHCP addresses.

Due to hardware limitations, Model 1 Level 2, and Model 2 for Enhanced Network Stack (ENS) mode in vSphere 8.0 is not supported in ConnectX-5 and ConnectX-6 adapter cards.

When you enable LLDP on an ESXi host with a DPU, the host cannot receive LLDP packets.

vCenter Server 8.0 supports VASA API version 4.0. However, after you upgrade your vCenter Server system to version 8.0, the VASA API version might not automatically change to 4.0. You see the issue in 2 cases:

1. If a VASA provider that supports VASA API version 4.0 is registered with a previous version of VMware vCenter, the VASA API version remains unchanged after you upgrade to VMware vCenter 8.0. For example, if you upgrade a VMware vCenter system of version 7.x with a registered VASA provider that supports both VASA API versions 3.5 and 4.0, the VASA API version does not automatically change to 4.0, even though the VASA provider supports VASA API version 4.0. After the upgrade, when you navigate to vCenter Server > Configure > Storage Providers and expand the General tab of the registered VASA provider, you still see VASA API version 3.5.

2. If you register a VASA provider that supports VASA API version 3.5 with a VMware vCenter 8.0 system and upgrade the VASA API version to 4.0, even after the upgrade, you still see VASA API version 3.5.

Migrations to a vSAN datastore by using vSphere Storage vMotion of virtual machines that have at least one snapshot and more than one virtual disk with different storage policy might fail. The issue occurs due to an unauthenticated session of the NFC manager because the Simple Object Access Protocol (SOAP) body exceeds the allowed size.

A rare race condition when assigning a content ID during the update of the CBRC digest file might cause a discrepancy between the content ID in the data disk and the digest disk. As a result, you cannot create virtual machine snapshots. You see an error such as An error occurred while saving the snapshot: A digest operation has failed in the backtrace. The snapshot creation task completes upon retry.

When you have already opened the Utilization view under the Monitor tab for a resource pool or a cluster and then you change the resource pool or cluster, the view might not automatically refresh. For example, when you open the Utilization view of one cluster and then select a different cluster, you might still see the statistics of the first cluster.

In rare occasions, if the vSphere virtual infrastructure is continuously using more than 90% of its hardware capacity, some ESXi hosts might intermittently disconnect from the vCenter Server. Connection typically restores within a few seconds.

Due to a backend issue, you do not see banner notifications for background migration of historical data in the vSphere Client.

If your environment has vCenter Server systems of version 8.0 and 7.x, creating Cloud Native Storage (CNS) block volume by using API is successful, but you might see an error in the vSphere Client, when you navigate to see the CNS volume details. You see an error such as Failed to extract the requested data. Check vSphere Client logs for details. + TypeError: Cannot read properties of null (reading 'cluster'). The issue occurs only if you review volumes managed by the 7.x vCenter Server by using the vSphere Client of an 8.0 vCenter Server.

In rare cases, when requests to the vpxa service take long, for example waiting for access to a slow datastore, the request queue on vpxa might exceed the limit of file descriptors. As a result, ESXi hosts might briefly become unresponsive, and you see a vpxa-zdump.00* file in the /var/core directory. The vpxa logs contain the line Too many open files.

If you use a custom update repository with self-signed certificates that the VMware Certificate Authority (VMCA) does not trust, vCenter Lifecycle Manager fails to download files from such a repository. As a result, vCenter Server upgrade or update operations by using vCenter Lifecycle Manager workflows fail with the error Failed to load the repository manifest data for the configured upgrade.

When you add an existing virtual hard disk to a new virtual machine by using the VMware Host Client, the operation might fail with an error such as The VM configuration was rejected. Please see browser Console. The issue occurs because the VMware Host Client might fail to get some properties, such as the hard disk controller.

Starting from vSphere 8.0, ESXi uses the OpenSSL 3.0 FIPS provider. As part of the FIPS 186-4 requirement, the RSA key size must be at least 2048 bits for any signature generation, and signature generation with SHA1 is not supported.

ESXi 8.0 supports OpenSSL 3.0 and the only TLS protocol enabled by default is TLS 1.2. If you try to enable TLS 1.0 in /UserVars/ESXiVPsDisabledProtocols by using ESXCLI commands, this leads to connection drops.

If you extract an image of an ESXi host deployed from a host profile with enabled stateful install to deploy other ESXi hosts in a vSphere Lifecycle Manager cluster, the operation fails. In the vSphere Client, you see an error such as A general system error occurred: Failed to extract image from the host: no stored copy available for inactive VIB VMW_bootbank_xxx. Extraction of image from host xxx.eng.vmware.com failed.

When you navigate to the Image Depot tab in the vSphere Lifecycle Manager home view or select an add-on from the drop-down list in the New Cluster wizard, you do not see HPE and DELL add-ons for vSphere 8.0. This issue is temporary.

With vSphere 8.0, you can enable vSphere Lifecycle Manager to remediate all hosts that are in maintenance mode in parallel instead of in sequence. However, if a parallel remediation task fails, in the vSphere Client you might not see the correct number of hosts that passed, failed, or skipped the operation, or even not see such counts at all. The issue does not affect the vSphere Lifecycle Manager functionality, but only the reporting in the vSphere Client.

ESXi 8.0 introduces the option to explicitly stage desired state images, which is the process of downloading depot components from the vSphere Lifecycle Manager depot to the ESXi hosts without applying the software and firmware updates immediately. However, staging of images is only supported on an ESXi 8.0 or later hosts. Attempting to stage a vSphere Lifecycle Manager image on ESXi hosts of version earlier than 8.0 results in messages that the staging of such hosts fails, and the hosts are skipped. This is expected behavior and does not indicate any failed functionality as all ESXi 8.0 or later hosts are staged with the specified desired image.

When you start a vSphere Lifecycle Manager remediation on an ESXi hosts with DPUs, the host upgrades and reboots as expected, but after the reboot, before completing the remediation task, you might see an error such as:

A general system error occurred: After host … remediation completed, compliance check reported host as 'non-compliant'. The image on the host does not match the image set for the cluster. Retry the cluster remediation operation.

This is a rare issue, caused by an intermittent timeout of the post-remediation scan on the DPU.

When you look in the VMware Host Client to see the descriptions of the severity event states of an ESXi host, they might differ from the descriptions you see by using Intelligent Platform Management Interface (IPMI) or Lenovo XClarity Controller (XCC). For example, in the VMware Host Client, the description of the severity event state for the PSU Sensors might be Transition to Non-critical from OK, while in the XCC and IPMI, the description is Transition to OK.

In very specific circumstances, when a vSphere vMotion operation on a virtual machine runs in parallel with an operation that sends VMCI datagrams, services that use VMCI datagrams might see unexpected communication or loss of communication. Under the same conditions, the issue can also happen when restoring a memory snapshot, resuming a suspended VM or using CPU Hot Add. As a result, the guest operating system that depends on services communicating over VMCI might become unresponsive. The issue might also affect services that use vSockets over VMCI. This problem does not impact VMware Tools. The issue is specific for VMs on hardware version 20 with a Linux distribution that has specific patches introduced in Linux kernel 5.18 for a VMCI feature, including but not limited to up-to-date versions of RHEL 8.7, Ubuntu 22.04 and 22.10, and SLES15 SP3, and SP4.

If the advanced processor setting Enable IOMMU in this virtual machine is enabled on a virtual machine, and the guest operating system has enabled DMA remapping, the Linux guest operating system might fail to complete the booting process. This issue affects VMs with hardware version 20 and a Linux distribution that has specific patches introduced in Linux kernel 5.18 for a VMCI feature, including but not limited to up-to-date versions of RHEL 8.7, Ubuntu 22.04 and 22.10, and SLES15 SP3, and SP4.