CSCws30219

Cisco UCS M7 servers are susceptible to host lockups triggered by ECC (Error Correcting Code) memory events.

This issue is specifically related to how the server’s BIOS interacts with the EFI (Extensible Firmware Interface) during error conditions. During these events, the BIOS does not correctly retrieve a required variable, causing EFI to return an invalid parameter. As a result, the operating system is unable to continue normal operation and freezes without logging any errors. When this occurs, the host becomes unresponsive and KVM (keyboard, video, mouse) input is not accepted.

Single-bit ECC errors—while not logged due to DDR5 specifications—can trigger this condition, resulting in complete loss of host responsiveness.

X-Series M7: 5.4(0.250048)

C-series M7: 4.3(6.250053)

X-Series M7: 6.0(1.250179)

C-series M7: 6.0(1.250192)

CSCwi71153

On Cisco UCS C-Series M6 servers, BMC reset occurs due to watchdog reset caused by kernel scheduler null pointer dereference.

In addition, kernel panic might also occur in the server.

4.2(3b)

C-Series M6: 6.0(1.250192)

CSCwq61388

On Cisco UCS C-Series M6 servers running CIMC 4.3.5 or later, enabling VMD in BIOS with multiple NVMe drives can cause storage subsystem initialization to exceed 15 minutes, leading to Intersight Server Profile deployment failures.

The likelihood of encountering this issue increases with the number of drives installed.

Disabling VMD or reducing the number of drives can resolve the issue.

4.3(5.250001)

6.0(1.250162)

6.0(1.250174)

CSCwq74490

On the Cisco UCS C245 M8 platform with AMD EPYC 9655 processors, fully loaded NVIDIA L4 GPUs experience random PCIe link width downgrades on different slots when running the Linux Rebooter Stress test. The affected slots change after actions such as an AC power cycle or HUU update; for example, if slots 2 and 6 are downgraded initially, other slots may be downgraded after these actions. This results in GPUs operating at reduced speeds in the OS environment.

4.3(6.250053)

6.0(1.250162)

6.0(1.250174)

CSCwq01863

On Cisco UCS C225 or C245 M8 servers in standalone mode, configuring secure LDAP (LDAPS) on any port other than 636 results in the server attempting a standard LDAP connection instead of a secure one, causing LDAPS authentication to fail.

Always use port 636 for secure LDAP connections. Other ports will not work for LDAPS.

4.3(6.250044)

6.0(1.250162)

6.0(1.250174)

CSCwr24516

On Cisco UCS X210c M8, some servers may not power on or start up if they contain a hardware part (inductor) from a specific vendor. This issue causes a startup failure, and the server will remain off even when you try to turn it on. The problem is caused by a difference in how the new hardware part works, but it can be fully resolved by updating the board to the latest VR (Voltage Regulator) image, which is designed to support this new component.

Update the server with the latest VR image, clear any power on error, and turn the server off and on again. If the server still does not start, contact Cisco TAC for further assistance.

6.0(1.250126)

6.0(1.250162)

6.0(1.250174)

CSCwp91696

On Cisco UCS X215c M8 servers, when an E3.S NVMe drive is added to an empty slot while the server is already running, the drive connects at Gen 4 speed instead of Gen 5. This issue only occurs if the slot was empty during boot and the drive is hot inserted after the operating system has started.

Insert the drive into the slot before powering on the server. Drives added later will only operate at Gen 4 speed until the server is rebooted.

6.0(1.250065)

6.0(1.250162)

6.0(1.250174)

CSCwq01510

On Cisco UCS X215c-M8servers, hot plugging E3.S NVMe drives in slot 8 or slot 9 does not work in ESXi or Windows if those slots were empty when the operating system started. Drives added or removed after boot are not detected by the OS.

There is no workaround. Hot plugging in slots 8 and 9 will not function if the slots are empty during OS boot.

6.0(1.250086)

6.0(1.250162)

6.0(1.250174)

CSCwq93635

On Cisco UCS C240 M6 servers running CIMC 4.3.5 and later firmware, deployments with a large number of NVMe drives and VMD enabled in the BIOS may experience prolonged storage subsystem initialization times, taking up to 50 minutes. This delay can cause the Wait for storage subsystem initialization task to fail in Intersight Server Profile workflows, especially after the Wait for BIOS POST Completion step.

If possible, turn off VMD in the BIOS to avoid the delay. If VMD must stay on, try using fewer NVMe drives during the initial setup, since the more drives you have, the longer the initialization takes.

4.3(6.250077)

6.0(1.250162)

6.0(1.250174)

CSCwq58890

The firmware for the Cisco UCS VIC 15000 series adapters is updated to fix a rare, intermittently seen memory issue and to ensure more robust and consistent operation across diverse memory types and configurations.

X-series: 6.0(1.250120)

B-series: 6.0(1.250126)

C-series: 6.0(1.250127)

X-series: 6.0(1.250122)

B-series: 6.0(1.250127)

C-series: 6.0(1.250131)

CSCwn99720

On Cisco UCS C-Series M8 servers with Intel processors and an R1S CPU configured in a single-CPU setup on the 2-socket platform, Intel SGX remains disabled even after being enabled in the BIOS. This issue does not occur with a dual CPU configuration.

4.3(6.250039)

6.0(1.250127)

4.3(6.250044)