Mozilla Firefox

146.0.1 (aarch64)

macOS Tahoe 26.2

147.0 (64-bit)

146.0.1 (64-bit)

Windows 11 Enterprise

Apple Safari

Version 26.2 (21623.1.14.11.9)

macOS Tahoe 26.2

Google Chrome

Version 143.0.7499.193 (Official Build) (arm64)

macOS Sequoia 15.7.3 (24G419)

Version 143.0.7499.110 (Official Build) (arm64)

Version 139.0.7258.155

macOS Tahoe 26.2

Version 143.0.7499.193 (Official Build) (64-bit)

Version 143.0.7499.170 (Official Build) (64-bit)

Version 144.0.7559.60 (Official Build) (64-bit)

Windows 11 Enterprise

HTTPS

443

IPMI

623

KVM/vMedia

443

NTP

123

SSH

22

Firmware for OOB (BMC, BIOS, CPLD/EROT, and GPU).

ucs-c880a-m8-4.0.2.260009.tar.gz

Host Upgrade Utility support firmware upgrade of ConnectX7, Bluefield, and OCP.

HUU Version—4.0.2.260022

ucsc-880a-m8-huu-4.0.2.260022.iso

Firmware for OOB (BMC, BIOS, CPLD/EROT, and GPU) and In Band (ConnectX7, Bluefield, and OCP) components.

ucs-c880a-m8-4.0.2.260004.tar.gz

NVIDIA B300 (SXM) GPU

97.10.52.00.17-G540.0216.00.03

97.10.64.00.0C-G540.0216.00.03

Firmware for OOB (BMC, BIOS, CPLD/EROT, and GPU) and In Band (ConnectX7, Bluefield, and OCP) components.

ucs-c880a-m8-4.0.1.260004.tar.gz

NVIDIA B300 (SXM) GPU

97.10.52.00.17-G540.0216.00.03

97.10.64.00.05-G540.0216.00.03

Firmware for OOB (BMC, BIOS, CPLD/EROT, and GPU) and In Band (ConnectX7, Bluefield, and OCP) components.

ucs-c880a-m8-4.0.1.260003.tar.gz

NVIDIA B300 (SXM) GPU

97.10.52.00.17-G540.0216.00.03

CSCwu73066

The Redfish ResetToDefaults command with the ResetType set to PreserveNetwork does not preserve network settings as expected and instead falls back to DHCP. However, performing a reset to factory defaults with network preservation through the web UI works correctly. The issue occurs specifically when using the PreserveNetwork option through Redfish for CIMC reset to factory defaults.

This issue is resolved.

4.0.1.260004

4.0.2.260009

CSCwv31440

Redfish and web UI API requests on the Cisco UCS C880A M8 Rack Server can become unresponsive because HTTPS sessions remain stuck in CLOSE_WAIT. This issue can occur after a BMC restart if certain API activities start before the HSU process is fully ready.

This issue is resolved.

4.0.1.260004

4.0.2.260009

CSCwu10250

The server does not come up after performing a power cycle from the KVM interface. This issues occurs after updating all components using HUU. The usb1 interface fails to appear in the BMC shell ifconfig after BIOS POST completion, and the HGX/GPU inventory is not listed in BMC or HUU following firmware updates to BMC, BIOS, CPLD, or GPU.

This issue is resolved.

4.0.1.260004

4.0.2.260009

CSCwu20901

After the host is powered off from the operating system, a standard power-on request from CIMC or KVM can leave the server in a No Signal state.

This issue is resolved.

4.0.1.260004

4.0.2.260009

CSCwv11700

Duplicate Redis database entries can cause the BMC to enter a continuous reboot loop and make Redfish unavailable.

This issue is resolved.

4.0.2.260004

4.0.2.260009

CSCwv33368

If only the EPP image is updated and the server is left idle, the system can enter decommission mode and the subsequent BIOS upgrade can fail.

This issue is resolved.

4.0.2.260004

4.0.2.260009

CSCwv05350

When the HMC is down or the GPU sled is not seated correctly, Redfish requests on the Cisco UCS C880A M8 Rack Server can take several minutes to return.

This issue is resolved.

4.0.2.260004

4.0.2.260009

CSCwv57839

After repeated host power cycles on the Cisco UCS C880A M8 Rack Server, BIOS-to-BMC communication can fail and the host can stop creating /dev/ipmi0 for KCS access.

This issue is resolved.

4.0.2.260004

4.0.2.260009

CSCws10791

The installer fails to read the ISO from the local media mount (virtual CD/DVD) when installing Ubuntu 24.04 or RHEL 10 through KVM. This results in I/O errors on /dev/sr0 during early boot, causing system to terminate with a protocol error. The local media option is only available through KVM, while remote media is accessible through the Cisco BMC 4.0 UI.

This issue is resolved.

4.0.1.260004

4.0.2.260004

CSCws11095

During host power cycles, multiple components report lower non critical going low asserted and deasserted events in the IPMI SEL logs. These events include temperature sensors, voltages, power supplies, and fan speeds.

This issue is resolved.

4.0.1.260004

4.0.2.260004

CSCwt33362

The server does not complete a proper power cycle when the Power Cycle operation is triggered through the Redfish API while in a Pre-OS environment such as the UEFI shell or BIOS page. Although the Redfish command returns a success response, the system remains stuck with the KVM console showing No Signal and does not reboot as expected. Power Cycle operation works correctly when initiated from the KVM or Web UI.

This issue is resolved.

4.0.1.260004

4.0.2.260004

CSCws97457

After a BIOS firmware update performed while the host is powered off, the host may fail to power on using the Cisco BMC Web UI, displaying the error Performing power action failed. The host remains off and does not recover with standard BMC power actions. Recovery requires a full power cycle through Redfish API. When the BIOS update is done with the host powered on followed by a host power cycle, no power-on failure occurs.

4.0.1.260004

4.0.2.260004

CSCwt15746

The Firmware Inventory page for DGX displays only the BIOS component and the firmware version during Sync Recovery Redfish actions and BMC upgrades. The HGX tab remains empty, and Redfish also shows only the BIOS details. This issue occurs during BMC firmware updates and Sync Recovery operations for BIOS and BMC components.

4.0.1.260004

4.0.2.260004

CSCwt21722

The Graceful Restart option does not function correctly when triggered from the KVM console while the system is in the EFI Shell environment. The KVM console shows No Signal for a long time, and the host does not power cycle as expected. Graceful Restart works properly when initiated through the Redfish interface.

4.0.1.260004

4.0.2.260004

CSCws39920

The CPLD firmware update for MB and DCSCM_0 fails when the force upgrade option is not enabled. This issue occurs when using Redfish API or web GUI interfaces for the upgrade.

This issue is resolved.

4.0.1.260003

4.0.1.260004

CSCws67923

The Cisco UCS C880A-M8 server Web UI does not enforce the Force upgrade option during BMC and BIOS firmware updates.

This issue is resolved.

4.0.1.260003

4.0.1.260004

CSCws32802

BIOS token values set to non-default in the BIOS setup do not persist after a BIOS firmware update on the Cisco UCS C880A M8 Rack Server. This causes BIOS configurations to be lost after the update.

This issue is resolved.

4.0.1.260003

4.0.1.260004

CSCws71001

Remote Media in Cisco BMC 4.0 UI auto-populates available images and supports multiple mount slots for CIFS and NFS protocols. When HTTPS is used as the mount protocol, images are not auto-populated, only one mount slot is available, and manual entry of the image name is required. This causes a different user experience compared to CIFS and NFS mounts.

This issue is resolved.

4.0.1.260003

4.0.1.260004

CSCwv85541

Cisco UCS C880A M8 Rack Server can remain stuck in the Power Off state and may not respond to normal power-on or power-cycle operations.

This issue can be seen during host power-on or when the system shuts down automatically.

Perform a full power cycle by using Cisco BMC 4.0 management controls.

4.0.2.260009

CSCwv75511

Bluefield3 adapters on the Cisco UCS C880A M8 Rack Server may not retain NIC mode settings and can revert to DPU mode after reboot or after a BIOS reset to defaults.

After switching the adapter to NIC mode, an AC power cycle is required before making additional changes.

After switching the Bluefield3 adapter to NIC mode, perform a full AC power cycle before making additional changes.

4.0.2.260009

CSCwv76354

On Cisco UCS C880A M8 Rack Server with B3220 adapters configured for north-south Ethernet traffic, one of the Bluefield3 ports can fail to initialize.

This issue can occur on systems with B3220 adapters configured for north-south Ethernet traffic.

Perform a system power cycle from KVM or Redfish to recover the port.

4.0.2.260009

CSCwv86697

Cisco UCS C880A M8 Rack Server can boot from the backup BIOS SPI image when it does not reach the operating system within the expected time.

This issue can be seen during host reboot when the system does not boot to the operating system within the expected time.

Perform a full power cycle from the Cisco BMC 4.0 UI or through Redfish.

4.0.2.260009

CSCwu20901

After the host is powered off from the operating system, a standard power-on request from CIMC or KVM can leave the server in a No Signal state.

Use Power cycle option to turn on the Host.

4.0.2.260004

CSCws72023

After POST completes and the system stabilizes, some sensors can intermittently report No Reading and the fan can run at a higher duty cycle before recovering. This issue occurs during system power cycle.

It recovers automatically after few minutes of Post complete.

4.0.2.260004

CSCwu10250

The server does not come up after performing a power cycle from the KVM interface. This issues occurs after updating all components using HUU. The usb1 interface fails to appear in the BMC shell ifconfig after BIOS POST completion, and the HGX/GPU inventory is not listed in BMC or HUU following firmware updates to BMC, BIOS, CPLD, or GPU.

Reboot the BMC to recover from this issue.

4.0.2.260004

CSCwt89201

After BIOS defaults are restored under specific ConnectX-8 Option ROM conditions, half of the integrated ConnectX-8 ports can switch to InfiniBand mode while the remaining ports stay in Ethernet mode. This issue occurs when ConnectX-8 Option ROM is enabled in the BIOS Menu, and followed by that if a BIOS restore to default is performed, half of the CX8 ports will be in Ethernet Mode other half in Infiniband mode.

Do not enable Option ROM for ConnectX-8 cards. If the system is impacted, reset the ConnectX-8 port link type back to Ethernet and reboot the system.

4.0.2.260004

CSCwu31617

Requests for a single SEL entry through Redfish can return incomplete details even though the expanded SEL collection returns the full entry data.

Use the web UI for SEL logs, or fetch SEL entries by using the Redfish expand-all query.

4.0.2.260004

CSCws10791

The installer fails to read the ISO from the local media mount (virtual CD/DVD) when installing Ubuntu 24.04 or RHEL 10 through KVM. This results in I/O errors on /dev/sr0 during early boot, causing system to terminate with a protocol error. The local media option is only available through KVM, while remote media is accessible through the Cisco BMC 4.0 UI.

Use the remote media option to mount vMedia is available through the Cisco BMC 4.0 UI.

4.0.1.260004

CSCws11095

During host power cycles, multiple components report lower non critical going low asserted and deasserted events in the IPMI SEL logs. These events include temperature sensors, voltages, power supplies, and fan speeds.

There are no known workarounds.

This issue does not have any functional impact.

4.0.1.260004

CSCws30241

Redfish interface does not display port speed and link status information. Additionally, extra spaces appear in the Serial Number, Part Number, and Model fields of BF-3 adapters.

In the Cisco BMC 4.0 UI, the link speed for BF-3 ports can be found under the Advanced tab in the BIOS settings. This allows you to view and manage the port speed configuration directly from the BIOS interface.

4.0.1.260004

CSCws97457

After a BIOS firmware update performed while the host is powered off, the host may fail to power on using the Cisco BMC Web UI, displaying the error Performing power action failed. The host remains off and does not recover with standard BMC power actions. Recovery requires a full power cycle through Redfish API. When the BIOS update is done with the host powered on followed by a host power cycle, no power-on failure occurs.

Perform the BIOS firmware upgrade while the host is powered on. After the upgrade, perform a host power cycle to activate the new firmware.

4.0.1.260004

CSCwt15746

The Firmware Inventory page for DGX displays only the BIOS component and the firmware version during Sync Recovery Redfish actions and BMC upgrades. The HGX tab remains empty, and Redfish also shows only the BIOS details. This issue occurs during BMC firmware updates and Sync Recovery operations for BIOS and BMC components.

Wait for the upgrade to finish before verifying the components and their versions.

4.0.1.260004

CSCwt33362

The server does not complete a proper power cycle when the Power Cycle operation is triggered through the Redfish API while in a Pre-OS environment such as the UEFI shell or BIOS page. Although the Redfish command returns a success response, the system remains stuck with the KVM console showing No Signal and does not reboot as expected. Power Cycle operation works correctly when initiated from the KVM or Web UI.

Use the Web UI interface to perform Power cycle.

4.0.1.260004

CSCwt21722

The Graceful Restart option does not function correctly when triggered from the KVM console while the system is in the EFI Shell environment. The KVM console shows No Signal for a long time, and the host does not power cycle as expected. Graceful Restart works properly when initiated through the Redfish interface.

Use the web UI, Redfish API, or KVM Power cycle option to perform Graceful restart.

4.0.1.260004

CSCwu73066

The Redfish ResetToDefaults command with the ResetType set to PreserveNetwork does not preserve network settings as expected and instead falls back to DHCP. However, performing a reset to factory defaults with network preservation through the web UI works correctly. The issue occurs specifically when using the PreserveNetwork option through Redfish for CIMC reset to factory defaults.

Use the web UI option to perform a CIMC reset to factory defaults and preserve both network and IPMI configurations.

4.0.1.260004

CSCws11095

During host power cycle on the Cisco UCS C880A M8 Rack Server, multiple components assert and deassert lower non critical going low events, as recorded in the IPMI SEL logs. These events include temperature sensors on various MCUs, voltage sensors on the motherboard and power distribution board, power supply units, and fan speed sensors. This behavior is observed as part of the normal power cycle process and reflects transient state changes in component monitoring during power transitions.

There is no known workaround. These events have no functional impact.

4.0.1.260003

CSCws28323

Cisco UCS C880A M8 Rack Server experiences boot time increases to over 40 minutes when the Option ROM dispatch policy is enabled for all network cards in the BIOS setup.

Disabled option ROM after finishing option configurations.

4.0.1.260003

CSCws32802

BIOS token values set to non-default in the BIOS setup do not persist after a BIOS firmware update on the Cisco UCS C880A M8 Rack Server. This causes BIOS configurations to be lost after the update.

Use Redfish method to configure BIOS token post BIOS upgrade.

4.0.1.260003

CSCws30241

REST API interface for Cisco UCS C880A M8 Rack Server does not display port link speed or link status information.

Refer to Advance tab in BIOS to get Link speed.

4.0.1.260003

CSCws67923

The Cisco UCS C880A-M8 server Web UI does not enforce the Force upgrade option during BMC and BIOS firmware updates.

Select force option during BIOS and BMC update.

4.0.1.260003

CSCws39920

The CPLD firmware update for MB and DCSCM_0 fails when the force upgrade option is not enabled. This issue occurs when using Redfish API or web GUI interfaces for the upgrade.

Choose FORCE option while performing firmware update.

4.0.1.260003

CSCws30241

Redfish interface does not display port speed and link status information. Additionally, extra spaces appear in the Serial Number, Part Number, and Model fields of BF-3 adapters.

In the Cisco BMC 4.0 UI, the link speed for BF-3 ports can be found under the Advanced tab in the BIOS settings. This allows you to view and manage the port speed configuration directly from the BIOS interface.

4.0.1.260004

CSCwu26318

After downgrading to release 4.0.1.260004, Redfish services can stop responding.

Downgrading to release 4.0.1.260004 is not supported. As a workaround, perform a CIMC reset to factory defaults to recover Redfish services.

4.0.2.260004

CSCwt88272

The PSU indexing in the Redfish PowerSupplies collection starts from PSU_0 to PSU_11, while IPMI number PSUs from PSU1 to PSU12. This causes an indexing mismatch across management interfaces. The issue occurs in release 4.0.1.260004 .

Additionally, the Redfish interface may show 13 PSU entries under these conditions.

After upgrading to release 4.0.2.260004, perform a CIMC factory reset with all presever options unchecked. Then, reboot the host to complete the process.

4.0.1.260004

CSCws28323

Server boot time may increase to over several minutes when the OptionRom dispatch policy is enabled for all network cards in the BIOS setup. Some servers may hang at the BIOS POST.

Recovery steps include forcing a power off, loading BIOS defaults, saving, performing a warm reset, and re-enabling the OptionRom. Despite recovery, boot time remains significantly longer than expected.

Disabled option ROM after finishing option configurations.

4.0.1.260004

CSCws43240

The GPU firmware update using the Redfish API force update option completes successfully for most HGX components; however, the HGX_FW_BMC_0 component does not reflect the updated firmware version after the update and server power cycle.

Verify the task completion at 100%, perform a full power cycle, and then check the firmware versions of all HGX components. This ensures the HGX_FW_BMC_0 firmware version updates correctly after the GPU firmware update process.

4.0.1.260004

BMC

BMC Reboot

BIOS

Host Power Cycle

EROT (BMC)

Full Power Cycle

EROT (BIOS)

Full Power Cycle

Motherboard (MB) CPLD

Full Power Cycle / AC Power Cycle

Backplane (BP) CPLD

Full Power Cycle / AC Power Cycle

E1.S CPLD

Full Power Cycle / AC Power Cycle

DCSCM CPLD

Full Power Cycle / AC Power Cycle

GPU

Host Power Cycle