Release Notes for Cisco UCS Software, Release 2.1

CSCur39162

When you run the show platform fwm info hw-stm asic num command on a Fabric Interconnect, the FWM process no longer crashes and reboots the Fabric Interconnect.

2.0(1q)A

2.1(3l)A

CSCuv03557

During Cisco UCS Manager upgrade, FI activation no longer fails with the following error:

2.1(3a)A

2.1(3l)A

CSCuy64856

The Cisco UCS fabric interconnects (FI) are no longer rebooted with the reboot reason FWM hap reset.

2.1(3h)A

2.1(3l)A

CSCva54957

A reboot is no longer triggered without a “user -ack” when modifying a service profile that requires a reboot while shallow association is failing.

2.1(3a)A

2.1(3l)A

CSCuu40978

The syslog is now truncated after it reaches the configured maximum size. It no longer fills up the Fabric Interconnect file system.

2.1(1a)A

2.1(3l)A

CSCuw78008

Web sessions no longer remain on after the Web Session Timeout period specified in the authentication option under user management in the Admin tab.

2.1(3a)A

2.1(3l)A

CSCuy94843

When service profiles remain in the user-ack state for more than 11.5 days, Cisco UCS Manager no longer times out the user-ack state, and waits for an explicit user acknowledgment for all pending activities.

2.1(1a)A

2.1(3l)A

CSCuz20650

When syslog messages are generated continuously, the syslog suspend timer does not recover. Thus, no events are sent to the remote syslog server. This issue is now resolved.

2.1(1a)A

2.1(3l)A

CSCuz86450

The server no longer reboots because the system does not accept user input on the order property of adaptorHostIf.

1.4(1j)A

2.1(3l)A

CSCvc48423

Downloading a bundle more than 1GB in size from a local desktop no longer fails.

2.1(1a)A

2.1(3l)A

CSCuu99255

The “show fabric-interconnect inventory expand” command no longer displays the ethernet port with a role of “Unknown” instead of “Server” on a fabric interconnect with a GEM card.

2.1(1a)A

2.1(3l)A

CSCuj08063

After disassociating the service profile from the server and downgrading the server software bundle to any Cisco UCS Manager release between 2.1(1a) and 2.2(1h), server discovery no longer fails.

2.1(1a)A

2.1(3l)A

CSCug25894

During Cisco 2100 Series IOM boot and chassis reacknowledgment, sysmgr cores are no longer seen.

2.0(4a)A

2.1(3h)A,2.1(3k)A

CSCus11782

After rebooting a Fabric Interconnect (FI) that is operating in the FC end-host mode, all member links of the SAN port channel come up.

2.1(3a)A

2.1(3k)A

CSCus82914

Configuring SPAN on Cisco UCS fabric interconnects for all VLANs including vMotion and storage Ethernet no longer decreases performance.

2.1(3a)A

2.1(3k)A

CSCut55171

During large scale deployments with Cisco UCS 6200 FIs the kernel will not crash anymore on receipt of a corrupted packet in the inband driver. Now the event will be logged in detail to the kernel log.

2.1(3a)A

2.1(3k)A

CSCuv89839

When the fabric interconnect is in switch mode with direct attached storage, and its FC uplinks to the direct attached storage are up, these FC uplinks now allow traffic to pass.

2.1(3a)A

2.1(3k)A

CSCuy25611

When you upgrade C220-M3 server with VIC1225/P81E adapter from release, the option rom will load for the VIC and no longer be stuck in "Loading Cisco VIC Fc driver" state.

2.1(3a)B

2.1(3k)A

CSCuo93591

For a fabric interconnect in end-host mode, the MAC address table aging time no longer gets stuck at 300 regardless of the configuration.

2.1(3a)A

2.1(3k)A

CSCup95855

FSM tasks are no longer stuck in the throttled state in Cisco UCS Manager during Cisco UCS C240 M3 server upgrade.

2.0(5c)A

2.1(3k)A

CSCur01185

The HA policy of Reset is no longer triggering the fabric interconnect to reset.

2.1(3a)A

2.1(3k)A

CSCus32933

Cisco UCS Manager now displays an error message when a WILL_BOOT_FAULT event is raised because of an incorrect CPLD version.

2.1(3a)A

2.1(3k)A

CSCus34689

When using Cisco UCS Manager with C-Series integration, Cisco UCSM GUI no longer displays the following message on hovering between the C-Series servers and FIs:

2.1(3a)A

2.1(3k)A

CSCus56140

Fabric Interconnect failover status in a cluster is no longer stuck in Switchover In Progress when the management interface of the primary fabric interconnect is down for more than the Management Interface Monitoring Period

2.1(3a)A

2.1(3k)A

CSCus73964

When you download an infrastructure software bundle onto a system where the same infrastructure software bundle was previously installed, but was subsequently deleted, the UCS Manager FIs, and IOMs no longer downgrade to that software bundle.

2.1(3e)A

2.1(3k)A

CSCus97608

Faults in Cisco UCS Manager such as “error accessing shared-storage” and timeout/failover warning within the “show cluster extended-state” are no longer displayed when several devices in several chassis are reporting EBUSY within the I2C logs.

2.1(3a)A

2.1(3k)A

CSCux71937

The CPU utilization no longer displays 100 percent for the kernel in the output of the show system internal file /proc/stat command. This occurs when the output contains more than 64 characters (excluding the last two trailing zeros). This could happen if the system had been up for a very long time (more than 200 days, but this time-frame could vary):

The output of this command now indicates the correct CPU utilization levels.

2.1(3a)A

2.1(3k)A

CSCux76128

Firmware Auto Install upgrade validation fails as expected when upgrading to Cisco UCS Manager Release with deprecated hardware. Auto Install can now be initiated by using the force option either through the GUI or the CLI.

2.1(3j)A

2.1(3k)A

CSCut43948

In an Auto Deploy boot deployment, after upgrading the adapter firmware from Release 2.1(1a) to later versions for Cisco UCS VIC adapter M81KR on Cisco servers B230 M1 or C260, ESXi will no longer show a Purple Screen of Death (PSOD) while PXE booting.

2.1(1a)A

2.1(3j)A

CSCus64439

Cisco UCS Manager Mezz logs and VMware vmkernel logs no longer indicate storage latency and numerous FNIC aborts.

2.0(1q)A

2.1(3i)A

CSCug25894

During Cisco 2100 Series IOM boot and chassis reacknowledgment, sysmgr cores are no longer seen.

2.0(4a)A

2.1(3h)A

CSCuq74472

Unnecessary thermal events on IOM stating that the thermal sensor reading is not available no longer occur.

2.1(3c)B

2.1(3h)B

CSCur29264

The security vulnerability identified by Common Vulnerability and Exposures (CVE) CVE-2014-3566 is addressed.

2.1(1a)A

2.1(3h)A

CSCus85186

After activating Cisco Trusted Platform Module (TPM), the enable and active statuses will not remain as disabled and deactivated at the BIOS prompt.

2.0(5g)B

2.1(3h)B

CSCus69458

The heap-based buffer overflow vulnerability in the GNU C library, documented in Common Vulnerability and Exposures (CVE) CVE-2015-0235 is addressed.

1.0(2k)A

2.1(3h)A

CSCur88952

svc_sam_dme core is no longer found while upgrading or downgrading.

1.4(4l)A

2.1(3h)A

CSCur54705

Cisco UCS Manager will no longer send UCS Manager username and password hashes to the configured SYSLOG server every 12 hours.

2.1(1a)A

2.1(3h)A

CSCur77746

Cisco UCS B200 M3 and UCS B22 M3 blade servers running Cisco UCS Manager Release 2.1(3h) no longer allow downgrade of the Board Controller firmware.

2.1(3a)A

2.1(3h)A

CSCut09151
CSCut21914
CSCut08605

Traffic from the host and CIMC will no longer collide on the shared System Management BUS (SMBUS). As a result, certain system failures such as false thermal alarms will not happen.

2.1(3a)B

2.1(3h)B

CSCuq40256

After an FI is reset, priority flow control on the IOM interface to a blade no longer becomes disabled.

2.1(3c)A

2.1(3g)A

CSCuo50049

Cisco UCS Manager no longer experiences HA cluster failover after upgrading from Release 1.4.

2.0(5g)A

2.1(3g)A

CSCuh23872

IOMs connected to a fabric interconnect no longer fail discovery after FI reboot.

2.0(2d)A

2.1(3g)A

CSCuo51708

UCSM autoinstall downgrade no longer fails while reverting firmware during autoinstall process.

2.1(1f)A

2.1(3g)A

CSCur37260

FI upgrade or downgrade no longer fails due to lack of disk space in /mnt/pss.

2.1(3c)A

2.1(3g)A

CSCuq52499

Cisco UCS Manager no longer raises a unnecessary fault on the IOM when the CPU on a blade in the chassis crosses the UNC/UC threshold.

2.1(3b)A

2.1(3g)A

CSCur01379

The security vulnerabilities identified by Common Vulnerability and Exposures (CVE) CVE-2014-7169, CVE-2014-6271, CVE-2014-6277, CVE-2014-7186, CVE-2014-7187, and CVE-2014-6278 are addressed.

2.0(1q)A

2.1(3f)A

CSCuj84257

PXE boot no longer fails to get DHCP address due to DHCP offer packets that contain malformed BootP options.

2.1(2a)B

2.1(3e)B

CSCul69513

Cisco UCS Manager now monitors CRC error counters on the fabric ports on the IOM.

2.1(1a)A

2.1(3e)A

CSCuo34760

Veths/VIFs no longer remain in down state when primary FI gets rebooted and comes back up as subordinate FI in a cluster HA configuration.

2.1(2d)A

2.1(3e)A

CSCul44421

Error no longer encountered when accessing shared-storage during FI reboot, upgrade, or IOM reset.

2.0(5f)B

2.1(3d)B

CSCuo30572

Intel v2 processors no longer cause PSOD with Microsoft Windows 2008 R2 VM guests.

2.1(3a)A

2.1(3d)A

CSCun24381

Customers using the Cisco UCS PowerTool will no longer experience problems when scraping the Java log file for XML parsing of config changes when running Java version 7 update 45.

2.1(3c)A

2.1(3d)A

CSCuo78883

Cisco UCS Manager and KVM users or admins using JRE version 1.7 update >= 40 no longer encounter a pop-up window with the ' Application Blocked by Security Settings' dialog.

2.0(1m)A

2.1(3d)A

CSCun25187

When using third-party certificates with Cisco UCS Central, if a keyring with a certificate chain is configured to use HTTPS communication, the status of Cisco UCS Manager no longer displays in Cisco UCS Central as 'Lost-Visibility'. Certificates can be signed by subordinate Certificate Authority (CA) as well as root CA.

2.1(2a)A

2.1(3d)A

CSCun84897

Changing MTU in a vNIC template used with a global service profile no longer ignores user acknowledgment settings.

2.1(3a)A

2.1(3d)A

CSCud22620

Cisco UCS now has an improved accuracy at identifying degraded DIMMs.

2.1(1a)A

2.1(3c)B

CSCui31011

Cisco UCS Manager no longer reports FI and VIF down faults for Cisco UCS blade servers due to connectivity loss after reboot of subordinate FI.

2.1(2a)A

2.1(2d) only, 2.1(3c) and onwards

CSCui48523

Capability catalog now recognizes the UCS-HDD2TI2F213 2TB Disk.

2.1(3a)A
2.1(3a)T

2.1(3c)A
2.1(3f)T

CSCuj56943

Option added for export of DME and AG logs to a remote server before being deleted locally.

2.1(2a)A

2.1(3c)A

CSCuj78099

FCNS database no longer fails to sync with MDS in Switch mode.

2.1(3a)A

2.1(3c)A

CSCuj92500

HIF ports no longer go down when adapter is receiving multiple FNIC abort messages when generating high IO to target on a RH 6.3 VM.

2.1(1a)B

2.1(3c)B

CSCul00654

HA controller no longer reports INAPPLICABLE state in 'show cluster ext' command output when checking status of a peer (subordinate FI that has become the primary FI) after a swap due to reboot of the previously primary FI.

2.1(1a)A

2.1(3c)A

CSCul38768

Cisco UCS Manager no longer displays below invalid “unresponsive” management service transient fault that was being triggered by a software defect, which gets cleared after a few seconds; this error should only be seen in the case of an actual HA condition failure:

%UCSM-2-MANAGEMENT_SERVICES_UNRESPONSIVE: [F0452][critical][management-services-unresponsive][sys/mgmt-entity-A] Fabric Interconnect A, management services are unresponsive

2.1(2a)A

2.1(2d) and 2.1(3c) only

CSCuj84421

Installing Java 7 update 45 no longer causes UCS Manager GUI failures.

2.1(1f)A

2.1(3b)A

CSCuj61839

Cisco UCS Blade servers running 2.1(3b) firmware on a Cisco M81KR VIC adapter no longer encounter "ASSERT FAILED @ mips/ecpu_pani.c:138" errors.

2.1(2a)B

2.1(3b)B

CSCuj99958

During heavy FC traffic, the server no longer stops responding with an ASSERT FAILED (Exception 2 triggered!) @ mips/ecpu_panic.c:138 error.

2.1(1f)A

2.1(3b)A

CSCul21224

Cisco UCS FI reset no longer occurs due to vlan_mgr hap reset error.

2.0(1s)A

2.1(3b)A

CSCuj10564

Discard TX on a FC trunk port are no longer seen after hot-swapping the Cisco UCS-FI-E16UP expansion module on the Cisco UCS 6248UP FI.

2.1(1f)A

2.1(3b)A

CSCuj32124

During normal operation IOM no longer unexpectedly reboots with CPU reset.

1.4(2b)A

2.1(3b)A

CSCuj42355

When upgrading to 2.1(3b)C, the Cisco UCS C-Series integrated servers no longer lose data connectivity, and VIF paths no longer reflect an Error Unknown state after the FI reboots with the new code.

2.1(2a)A

2.1(3b)A

CSCui41165

Cisco UCS Manager no longer displays “error accessing shared-storage”error or have the following issues:

• Call home fan alerts are sent and cleared immediately
• Errors during IOM boot-up

1.4(2b)A

2.1(3b)A

CSCui82679

FlexFlash storage is no longer disconnected from a Cisco B200 M3 server after booting ESX or ESXi from a FlexFlash card.

2.1(2a)A

2.1(3b)A

CSCui94368

A dcosAG crash is no longer observed on a system running with Call Home enabled.

2.1(1f)B

2.1(3b)A

CSCuh85553

When IPMI is enabled from Cisco UCS Manager, Cipher 0 is no longer used as the default.

2.1(1e)A

2.1(3b)A

CSCul33403

Creating multiple service profiles simultaneously no longer assigns the pool identities in reverse order.

2.1(2a)A

2.1(3b)A

CSCuh49817

The following new microcodes were added to Release 2.1(3a):

• M03106A5_00000019
• M03206C2_0000001A

2.0(1a)B

2.1(3a)B

CSCua50442

BIOS will no longer remap incorrect BMC FRU data for SMBIOS table.

2.0(2q)A

2.1(3a)A

CSCuc66914

A global VLAN will no longer go missing on an FI after rectifying a conflicting FCoE VLAN condition after upgrade from 1.4.1 to 2.0(4a) or later.

2.0(4a)A

2.1(3a)A

CSCud27864

BIOS will no longer hang during POST when memory mapped IO above 4 GB is enabled and CSB-MEZ-QLG-03 is present in the blade.

2.1(1a)B

2.1(3a)A

CSCud89583

Cisco UCS B440 Blade Servers running Citrix XenServer 6.0.2 with E7-4830 and all C states disabled no longer freeze with a "CATERR_N" error.

2.0(3a)A

2.1(3a)A

CSCug41743

The following BIOSes will support 3 DDR-1333 DIMMs per channel and you will no longer see high memory speed error:

• B420M3.2.0.5.0.120720122110

• B420M3.2.0.5a.0.121720121433

• B420M3.2.1.1a.0.121720121615

2.1(1d)B

2.1(3a)A

CSCug51358

FCoE uplinks to the Cisco Nexus 5548 switch will no longer experience an MTS buffer leak between the port manager request high priority and fcoe_mgr due to an FC-Map mismatch between Cisco UCS Manager and the upstream Cisco Nexus 5548 switch.

This FCoE uplinks will no longer flap, and the VFCs will no longer be shown as error disabled.

2.1(1d)A

2.1(3a)A

CSCug62535

BMC no longer logs the following message:

2.0(5a)A

2.1(3a)A

CSCuh39242

The current severity level of Upper Non-critical and Upper Critical CPU thermal faults are no longer incorrectly classified as minor faults.

2.0(2m)A

2.1(3a)A

CSCuh61202

FC storage traffic through an IOM no longer stops when the IOM is reset or reinserted, or the cable between the IOM and FI is removed or reinserted.

Note To avoid this issue, you should first upgrade any Cisco UCS 1240, Cisco UCS 1280, and Cisco M81KR adapter firmware before updating the Cisco UCS infrastructure components (Cisco UCS Manager, IOM, and FI).

2.1(2a)B

2.1(3a)B

CSCuh61543

Cisco UCS Manager will no longer display configuration failure when a service profile with private VLAN and VFC is associated with Cisco UCS C-series C460M2

2.1(1e)A

2.1(3a)A

CSCuh73875

SNMP polling against FI drivers will no longer cause high volume of CPU usuage.

2.1(1a)A

2.1(3a)A

CSCui17731

SFP validation failed error will no longer occur when you insert a supported GBIC transeiver in Fabric Extender Server ports for C-series integration.

2.1(1b)A

2.1(3a)A

CSCui21176

When a PSU is removed from a chassis with 4 PSUs, the power state on the chassis no longer shows “redundancy-failed”.

2.1(2a)A

2.1(3a)A

CSCui37900

When you have the same authentication profile for both the iSCSI initiator and the target, upgrade from 2.0(1t) to 2.1(3a) will no longer have DME crash.

2.1(2a)A

2.1(3a)A

CSCui45873

When the ethpm MTS queues are full, the primary FI no longer reboots with VIM core: mts_acquire_q_space() failing.

2.1(2a)A

2.1(3a)A

CSCui45963

Some of the text and controls are no longer truncated When you create a service profile or service profile template using the wizard.The edit option for storage settings is enabled.

2.1(2a)A

2.1(3a)A

CSCui48112

FC VIF will no longer stay in unpinned state when you connect Cisco UCS C-series C220M3 with Cisco UCS Manager in dual-wire management mode.

2.1(2a)A

2.1(3a)A

CSCui94688

The FI no longer crashes when open file descriptions reaches beyond 10,000 in Cisco UCS Manager, release 2.1.

2.1(1f)B

2.1(3a)A

CSCui31011

Cisco UCS Manager no longer reports FI and VIF down faults for Cisco UCS blade servers due to connectivity loss after reboot of subordinate FI.

2.1(2a)A

2.1(2d) only, 2.1(3c) and onwards

CSCul38768

Cisco UCS Manager no longer displays below invalid “unresponsive” management service transient fault that was being triggered by a software defect, which gets cleared after a few seconds; this error should only be seen in the case of an actual HA condition failure:

2.1(2a)A

2.1(2d) and 2.1(3c) only

CSCuh81555

Board controller activation no longer fails on a limited set of Cisco UCS B200 M3 blade servers, when upgrading to release 2.1(2a).

2.0(4d)B

2.1(2c)B

CSCui06351

Major faults are no longer raised for default keyring certificate status showing as unknown.

2.1(2a)A

2.1(2c)A

CSCui40766

Cisco UCS Manager no longer fails to detect FlexFlash when enabled in a local disk configuration policy.

2.1(2a)A

2.1(2c)A

CSCui62823

VLAN groups are no longer applied incorrectly, which could cause an outage during update.

2.1(1d)A

2.1(2c)A

CSCuc19701

The 2204 IOM no longer reboots when the FI is reset.

2.1(1a)A

2.1(2a)A

CSCuf61116

IOMs no longer crash due to a memory leak in the baseboard management controller (BMC).

2.0(1s)A

2.1(2a)A

CSCuf17523

If the port speed is changed when the port is administratively down, NX-OS no longer reports that the FI port is down, while the interface counters show it is still receiving traffic.

2.1(1a)A

2.1(2a)A

CSCuh28239

During frequent MAC address changes between FIs, you will no longer see a delay in learning MAC addresses, and if the MAC address changes between server ports on the same FI, the MAC address will no longer point to an incorrect destination.

2.1(1b)A

2.1(2a)A

CSCuf01402

Modifying a service profile with two iSCSI vNIC targets defined no longer prompts for a reboot before the second target can be configured.

2.1(1a)A

2.1(2a)A

CSCue47159

In a UCS chassis with one or more empty slots, Cisco UCS Manager no longer shows a critical alert and an "FSM Failed" warning for a slot that has no blade.

2.1(1a)A

2.1(2a)A

CSCue65877

The storage daemon (storaged) running on the blade management controller (BMC) no longer generates multiple core files.

2.1(1a)A

2.1(2a)A

CSCuf57312

FIs running Cisco UCS Manager 2.1(1a) no longer experience a bladeAG reload that results in a core dump.

2.1(1a)A

2.1(2a)A

CSCug13702

After removing the SAN Connectivity Policy from a service profile, the FC zones are deleted and no longer visible using the Cisco UCS Manager GUI or CLI.

2.1(1b)T

2.1(2a)A

CSCug19471

Blades no longer display the discovery icon in Cisco UCS Manager every 2 minutes.

2.0(2q)B

2.1(2a)B

CSCug59101

FI crashes due to HAP reset are no longer triggered by an NTP process crash.

2.0(1s)A

2.1(2a)A

CSCug40776

Running the following commands no longer cause FI reboots due to a memory leak:

• connect nxos
• show vlan
• show run

2.0(3c)A

2.1(2a)A

CSCug20103

The FIs will no longer reset with the following error message:

1.4(1j)A

2.1(2a)A

CSCuh30440

Starting with Release 2.1(2a), Cisco UCS Manager no longer hangs when disassociating a service profile with FC zoning that is running in Fibre Channel switch mode.

For previous versions of Cisco UCS Manager software, use the following workaround:

1. Decommission the affected server, and then recommission it.
The previous FC zone will still be included in the zone database.

2. Create a duplicate service profile to the one that was deleted, using the same name and the same organization.

3. Associate the new service profile to the recommissioned blade.
The previous FC zone is deleted, and a new FC zone is created.

2.1(1d)A

2.1(2a)A

CSCud86528

Beginning with Cisco UCS Manager Release 2.1(2a), if a service profile's “Desired Power State” is in off state, it will be changed to on when the associated physical server is powered on using the reset or other server maintenance actions.

2.0(3b)B

2.1(2a)A

CSCud60153

If a link on which LLDP is configured flaps, it no longer causes memory leaks or an LLDP process crash.

2.0(2t)A

2.1(2a)A

CSCub37558

Cisco UCS Manager now displays the amber and amber blinking LED sensor states of LEDs on the blade, and raises faults in response to the color change.

2.0(3a)A

2.1(2a)A

CSCts11406

Beginning with Cisco UCS Manager Release 2.1(2a), you can delete the decommissioned server on a decommissioned and physically removed rack server from the setup, which allows you to reuse the removed server ID.

Note: If you execute this command when the rack server is decommissioned but physically connected, the rack server is recommissioned and reclaims the server ID.

2.0(2t)A

2.1(2a)A

CSCuc26744

Beginning with Cisco UCS Manager Release 2.1(2a), the GUI does not have the Set Bundle option for the Activate Firmware action. Use the Auto Install feature to activate the firmware on the B/C bundle. This change avoids a firmware activation failure when the CIMC and board controller firmware being activated at the same time. This is a known hardware restriction that could result in the board’s corruption.

2.0(4a)A

2.1(2a)A

CSCuc42488

Starting with Cisco UCS Manager Release 2.1(2a), the FCoE native VLAN is set to 1 where it was previously set to 4049. VLAN 1 can now communicate from upstream to a vNIC on a Cisco UCS blade without failing when the default VLAN configuration is used for the FCoE uplink.

2.1(1a)A

2.1(2a)A

CSCue49383

The default power policy is changed from N+1 to Grid during initial booting when a chassis CMC reboots or is powered on. This change ensures that all power supplies power up. After boot up, the power policy setting you configured is applied.

1.4(1i)B

2.1(2a)B

CSCug26974

For servers with Cisco UCS M71KR-E or Cisco UCS CNA M71KR-Q adapters, changing the ethernet adapter’s policy parameters no longer triggers a reboot on servers with the same policy.

2.1(1d)B

2.1(2a)B

CSCug43293

You no longer receive a "Configuration failed due to mac-derivation-virtualized-port" fault on service profiles when using a vNIC template that has a VM target under a suborganization.

2.1(1d)A

2.1(2a)A

CSCuf31431

Compiling rack server MIBs is now successful when performed on a CISCO-UNIFIED-COMPUTING-TC-MIB.my with 64 bit counters.

2.0(4b)C

2.1(2a)C

CSCtu17983

An ESX boot on blades that use VMware Auto Deploy no longer takes a long time to run.

2.0(1m)A

2.1(2a)B 2.1(2a)C

CSCuf78224

On a Cisco UCS B440-M2 Server with a Cisco UCS CNA M72KR-Q adapter card, VMware Auto Deploy 5.1 no longer hangs during a system boot.

2.0(4b)A

2.1(2a)B

CSCug85569

When performing an autoinstall on a server with the user ack policy, the server now proceeds with a graceful reboot. Some operating systems, such as Microsoft, no longer come up in recovery mode.

2.1(1e)A

2.1(2a)A

CSCuh35570

The fabric interconnect (FI) no longer reboots with a Kernel panic svr_sam_statsAG process error.

2.1(1e)A

2.1(2a)A

CSCue72786

VFC pinning now updates properly on the Cisco UCS M81KR VIC.

2.0(4b)A

2.1(2a)A

CSCud81176

After manually upgrading the CIMC and associating a service profile, the CIMC upgrade no longer fails while the status remains at Activating.

2.1(1a)A

2.1(2a)A

CSCud93569

The secondary FI no longer fails when you upgrade the FI firmware Cisco NX-OS software on Cisco UCS Manager.

2.0(3a)A

2.1(2a)A

CSCti87891

The Cisco UCS Manager shell now supports redirection of the show command output to a remote file system.

2.1(0.407)A

2.1(2a)A

CSCtt38889

The virtual interface on the standby vNIC is now shown as up when the vEth is up.

2.0(1m)A

2.1(2a)B

CSCud55036

With the vNIC template, the VLAN ID is now displayed correctly according to the configured VLAN number, instead of always displaying 1.

2.0(2m)A

2.1(2a)A

CSCue04360

After a boot, the B200 M3 Server no longer hangs after a few days with PECI errors.

2.0(3a)B

2.1(1f)B

CSCuf35678

When VLAN port count optimization (VLAN compression) is enabled on a Cisco UCS 6200 Series Fabric Interconnect, traffic no longer stops if an uplink port channel port goes down.

2.1(1a)A

2.1(1f)A

CSCuf60988

Virtual Fibre Channel ports are no longer error-disabled on one FI when the server is rebooted.

2.0(4a)A

2.1(1f)A

CSCug14669

A Fibre Channel (FC) path loss no longer occurs because the Fibre Channel Forwarder (FCF) MAC address is no longer learned dynamically.

2.1(1b)A

2.1(1f)A

CSCud60746

The system no longer runs out of memory when Call Home is enabled.

2.0(2a)A

2.1(1f)A

CSCug93076
CSCug93221
CSCug98662

The Cisco UCS B200 M3, B22 M3, and B420 M3 Blade Servers no longer experience noncorrectable memory errors during booting.

This patch provides a CIMC update for the voltage regulator. To ensure the voltage regulator is updated successfully, perform the following steps:

1. Update the CIMC image to 2.1(1f).

2. Power off the host.

3. Activate the CIMC.

4. Power on the host.

2.0(5b)B 2.0(5m)B 2.1(1a)B

2.1(1f)B

CSCue49366

Transient faults related to Cisco UCS Manager chassis SEEPROM usage and power capping no longer occur.

2.1(1a)B

2.1(1f)A

CSCue58839

The KVM launch manager now shows all service profiles when launched from a suborganization.

2.1(1a)A

2.1(1f)A

CSCuc87547

Cisco UCS Manager no longer reports PSU failures in the Cisco Nexus 2232 Fabric Extenders configured for Cisco UCS C-Series servers managed by Cisco UCS Manager.

2.0(3a)A

2.1(1f)A

CSCud13423

When the power policy is set to N+1, and an additional PSU is inserted into a slot with power, the new PSU no longer goes into spare mode instead of active mode.

2.0(1a)A

2.1(1f)B

CSCud79598

Renaming a service profile no longer increments the fault count incorrectly.

2.1(1a)A

2.1(1f)A

CSCue46382

Chassis discovery process issues, such as ports on FI-B displaying no object statistics or Cisco UCS Manager reporting incorrect states for ports on both FIs, no longer occur during a Cisco UCS Manager upgrade.

2.0(3c)A

2.1(1f)A

CSCue46600

When a Cisco UCS B440 Blade Server with a more recent Version ID (VID) is inserted in the chassis, Cisco UCS Manager no longer reports the previous VID.

2.0(4b)A

2.1(1f)A

CSCuf03602

Power supply VID data can be obtained by connecting to the IOM and running the show platform software cmcctrl fru psu command.

2.0(1m)A

2.1(1f)A

CSCue48076

If there are more than 21 IP addresses in the ext-mgmt ip pool, adding a subordinate FI to a standalone FI to convert into a cluster no longer causes the console to hang.

2.1(1a)A

2.1(1f)A

CSCug40752

The KVM console now supports Java 1.7 update 17 and Java 1.6 update 43.

2.1(1b)A

2.1(1f)A

CSCuf90470

When Call Home is enabled, Online Insertion and Removal (OIR) or failure of hardware modules in the FI no longer cause the FI to reboot.

2.1(1b)A

2.1(1e)A

CSCuf14193

Upon an upgrade to Cisco UCS Manager 2.1(1d), the FI no longer reboots due to Call Home server HA policy reset when Call Home is enabled.

2.1(1b)A

2.1(1d)A

CSCud56660

Duplicate license IDs no longer cause the LicenseAG process to core.

2.0(4d)A

2.1(1b)

CSCud10237

The eight default port licenses for flexible GEM on the FI are available for use.

2.0(3c)A

2.1(1b)

CSCue38650

The Cisco UCS Manager PSU policy and IOMs are no longer out of synchronization after the IOMs are rebooted.

2.1(1a)A

2.1(1b)

CSCud70368

Cisco UCS Central is now correctly creating the crossdomain.xml file on Cisco UCS Manager member domains.

2.1(1a)A

2.1(1b)

CSCud40412

When regenerating a new key ring or certificate, the new certificate is now successfully published to the FI web server and can be obtained by the HTTP process.

2.1(1a)A

2.1(1b)

CSCud53700

The portAG process no longer crashes while activating the Fabric Interconnect NX-OS software during an upgrade.

2.1(1a)A

2.1(1b)

CSCud59230

Port channels no longer go down after upgrading to Cisco UCS Manager Release 2.1.

2.1(1a)A

2.1(1b)

CSCub22238

The Cisco UCS 6100 Series Fabric Interconnects no longer contain a vulnerability in the Netconf interface.

2.0(1a)A

2.1(1b)

CSCud20253

Power capping on the Cisco UCS B420 Blade Server can now support 32 GB DIMMS.

2.1(1a)A

2.1(1b)

CSCue46817

After upgrading to Release 2.1(1a) or later, a Cisco UCS B440 M1/M2 Blade Server using a RAID key attached to a LSI MegaRAID SAS 9260 no longer generates a "Missing RAID Key" fault alert.

2.1(1a)B

2.1(1b)

CSCud27494

Traffic to a blade server will no longer be dropped and forwarded to another working link when an uplink is shut down either on the fabric interconnect or from the upstream switch.

2.0(4b)A

2.1(1b)

CSCud54919

On blade servers that are not associated with a service profile, CIMC cannot respond from Cisco UCS Manager at the same time.

2.0(1t)B

2.1(1b)

CSCuc38555

Legacy USB support items can no longer be changed in the BIOS setup.

2.1(1a)B

2.1(1b)

CSCud19629

FCoE uplink interface faults are now visible in the Cisco UCS Manager GUI.

2.1(1a)A

2.1(1b)

CSCud20765

When defined through a service profile template, SRIOV virtual functions (VFs) are no longer populated incorrectly in the instantiated service profiles.

2.1(1a)A

2.1(1b)

CSCuc44209

Cisco UCS Manager no longer displays the names for PSUs connected to a Cisco Nexus 2200 Series FEX in reverse order.

1.4(3l)C

2.1(1b)

CSCub51516

DHCP will no longer fail when multiple servers are restarted at the same time.

2.0(1e)A

2.1(1a)A

CSCuc27213

The Cisco UCS B200 M2 Blade Server no longer goes into a continuous reboot loop after upgrading from Release 2.0(1s) to Release 2.0(3a).

2.0(3)A

2.1(1a)A

CSCuc26566

The Cisco UCS 6200 Series Fabric Interconnect no longer reboots without a final confirmation warning after configuration changes.

2.0(4a)A

2.1(1a)

CSCtz07798

A service profile no longer generates configuration failures if the blade it is associated with is removed from the server pool.

2.1(0.208)B

2.1(1a)

CSCuc69455

A core dump caused by a memory leak is no longer seen when multiple VLANs are assigned to a service profile's vNIC.

2.1(0.489)A

2.1(1a)

CSCth96721

There is no longer a 128 character limitation to the number of OUs or the length of the Distinguished Name (DN) when using LDAP authentication with the Active Directory.

2.0(1w)A

2.1(1a)A

CSCtt36593

The svcmonAG process no longer fails and core dumps regularly on a 14 chassis setup.

2.0(1)A

2.1(1a)A

CSCty34034 CSCub48862 CSCub99354 CSCub16754

Discovery, association, or disassociation no longer fails after a BMC firmware update with a message about the VIC adapter.

1.0(2a)

2.1(1a)B

CSCuc00368or CSCuc87155

With RAID1 mode configured with two disks on a Cisco UCS B230 Blade Server, removing and reinserting one disk no longer causes the second disk to be shown as inoperable while a RAID 1 rebuild is in progress.

2.0(3a)A

2.1(1a)A

CSCuc24817

After an FI reboot or FI failover, the vEth is no longer shown as down when Cisco NX-OS shows it as up.

2.0(3c)A

2.1(1a)A

CSCuc59752

The snmptable command no longer fails to return any values.

2.0(2q)A

2.1(1a)A

CSCuc09958 CSCua17646

A Java 1.7 detected error no longer occurs when downgrading from Cisco UCS Manager Release 2.0(3a) and later releases running JRE 1.7 to Cisco UCS Manager Release 2.0(2r) and earlier releases.

2.0(4a)A

2.1(1a)A

CSCtz86513

Registration emails from the SCH portal are now received after new inventory messages are sent from Cisco UCS Manager.

1.4(2b)A

2.1(1a)A

CSCtz76897

While upgrading or discovering the Cisco UCS Manager, when the chassis discovery policy is changed to the set link-aggregation-pref port-channel policy, the FEX no longer goes offline.

2.0(1m)A

2.1(1a)A

CSCtr45130

The Blade Server no longer reboots when activated after upgrading from Cisco UCS Manager Release 1.4(1j) to 1.4(2b).

1.4(2b)A

2.1(1a)A

CSCub64209

FCoE packets are no longer dropped when host-control is enabled in QoS policies assigned to vNICs.

2.0.67 B

2.1(1a)B

CSCtz79579

Cisco UCS Manager no longer reports an incorrect status for faulty disks that fail to power on or link up.

2.0(2.83)B

2.1(1a)B

CSCub34939 CSCty33146

After upgrading Cisco UCS Manager, an SNMP crash no longer reboots both FIs during activation.

1.4(3s)A

2.1(1a)A

CSCuc35326

The Cisco UCS B200 M3, B22 M3, and B420 M3 Blade Servers no longer experience “Server Hardware Not Supported” or discovery errors when upgrading from Release 2.0(2) to Release 2.0(3) or 2.0(4) and the blades are inserted into a Cisco UCS DC chassis.

2.0(3a)A

2.1(1a)A

CSCtw59592

In a server using both a virtualized adapter card and a nonvirtualized card, extraneous NIC ports are no longer generated if there are fewer service profile vNICs than the minimum required physical NIC ports.

2.0(1t)A

2.1(1a)A

CSCtj62296

The minimum power cap that can be set is no longer 3400 W.

1.4(1i)A

2.1(1a)A

CSCtc86297

After a VM restarts, the virtual machine node on the VM tab no longer shows multiple instances of the same VM with one online and one offline.

1.1(1j)A

2.1(1a)A

CSCta66375

Fibre Channel port and server port events now appear on the Fibre Channel port and server port Events tabs.

1.0(1e)A

2.1(1a)A

CSCtu34607

Changing the dynamic vNIC policy to change the number of vNICs no longer causes static vNICs to get reordered on a PCIe bus.

2.0(2m)A

2.1(1a)A

CSCuc72049

When creating an access mode appliance port channel in Cisco UCS Manager, the default VLAN is no longer used instead of the specified VLAN.

2.0(3a)A

2.1(1a)A

CSCtx65534 CSCua31267

Deleting a VLAN in the fabric interconnect no longer causes the vNICs that are carrying that VLAN to flap.

2.0(2q)A

2.1(1a)A

CSCtn87981

Cisco UCS B230 and B440 Blade Servers with Cisco UCS M81KR and 82598KR-CI Adapters no longer fail with an "illegal fru" error.

2.0(1m)

2.1(1a)A

CSCuy64856

The Cisco UCS fabric interconnects (FI) may reboot with a FWM hap reset, which generates a core file.

Use only IGMPv2

-OR-

IGMPv2 and IGMPv3 must not be mixed for the same group. For IGMPv3, ensure that the difference between the source IP addresses for the same group is less than 127.255.255.255.

2.1(3h)A

Resolved in 2.1(3l)A

CSCup88087

In some cases, a B200 M3 server with a single CPU reports faults for a non-existent “CPU 2.”

This issue has no known workaround. The faults can be safely ignored.

2.1(3a)A

CSCup88161

In certain race conditions when running 2.1(1d) firmware, if a memory location which was freed is being used, the 6248 FI kernel crashes with no core.

This issue has no known workaround. If this condition occurs, the system will auto-reboot to a stable condition.

2.1(1d)A

CSCuq53385

In some cases, when attempting to regenerate expired self-signed Cisco UCS Manager certificate, Cisco UCS Manager will return the following regarding regeneration of default Key Ring:

This is accompanied by the following error messages in svc_sam_dme log file in Cisco UCS Manager techsupport log bundle:

To work around this issue, try the following:

1. When possible, use third-party CA signed keyring from internal or external CA; or

2. Contact TAC to resolve the error and to regenerate self-signed certificate.

Note Cisco UCS Manager does not support creating cert-req for default key ring.

2.1(3d)A

CSCup96252

In some cases, when the IOM reboots and FI is working properly, an incorrect fault may be raised due to incorrectly calculated very high delta value, if Cisco UCS Manager is configured to raise a fault on NI Ports

This issue has no known workaround. However, the NI port fault will go away automatically after ten to twenty minutes since this condition creates a false alarm.

2.1(3e)A

CSCus73964

When you download an infrastructure software bundle onto a system where the same infrastructure software bundle was previously installed, but was subsequently deleted, the UCS Manager FIs, and IOMs no longer downgrade to that software bundle.

To work around this issue, allow the downgrade process to complete and then upgrade back to original version.

2.1(3e)A

CSCuq74472

Cisco UCS Manager generates unnecessary thermal events on IOM stating that the thermal sensor reading is not available.

This issue has no known workaround.

2.1(3c)B

Resolved in 2.1(3h)A

CSCuo78883

‘Application Blocked by Security Settings’ error when starting the Cisco UCS Manager GUI or KVM Console application.

Because the Java Code Signing Certificate expired, users on Java 7 update 40 or higher might see the following message:

To fix this issue, you can either temporarily lower your Java security settings to add Cisco UCS Manager as an exception or, if running Java 7 update 51 or higher, you can add the Cisco UCS Manager host IP address to the Exception Site list.

To temporarily lower security settings:

1. Start Java Control Panel. (Location may vary depending on operating system and browser preferences.)

2. Lower security level to Medium.

3. Start Cisco UCS Manager.

4. At the warning message, check the “I accept the risk and want to run this application'” checkbox and click Run.

5. Return to the Java Control Panel and reset your security level.

To add the IP address to the exception site list (for Java 7 version 51 and higher):

1. Start Java Control Panel. (Location may vary depending on operating system and browser preferences.)

2. In Security area, click Edit Site button to add IP address to the list.

If you use HTTPS to access Cisco UCS Manager, ensure that you have the correct prefix.

3. Click OK.

2.1(1a)A

Resolved in 2.1(3d)A.

CSCum95778

Sorting results are incorrect when attempting to sort a listing of service profiles and ORGs.

This issue has no known workaround.

2.1(1b)A

CSCun09113

NTP is not syncing properly after switch to daylight savings time.

This issue has no known workaround.

2.0(4b)A

CSCun14140

System experiences kernel panic or hang when IOMMU is enabled on a UCSB-B200-M3 server.

Disable the IOMMU.

2.1(3b)B

CSCun25187

When using third-party certificates with Cisco UCS Central, if a keyring with a certificate chain is configured to use HTTPS communication, status of Cisco UCS Manager might display in Cisco UCS Central as 'Lost-Visibility.' This occurs if certificates are signed by subordinate CA instead of root CA.

Use third-party certificates that are signed directly by root CA for UCS Central https communication.

2.1(2a)A

CSCun28055

In rare cases, VLAN and port configuration is lost and only mgmt0 configuration remains for an FI that comes back online after an auto-install upgrade of Cisco UCS Manager.

Contact Cisco TAC for further assistance.

2.1(3)B

CSCun39077

Unable to see health status of hard disks in the Vsphere client when viewing the Configuration Tab > Hardware > Health Status window even though hard disks are visible in Cisco UCS Manager.

Install the latest LSI CIM provider, version 00.38.V0.03.

2.1(3a)A

CSCun83328

PCI addresses of static vNICs and vHBAs change after downgrading the Cisco UCS Manager, FI, IOM, and blade from release 2.1 to 2.0(5).

This issue has no known workaround.

2.2(1d)B

CSCun86873

In some cases, during initial discovery or following upgrade-initiated discovery, B230 M2 shows only 8 of 10 cores available.

Reacknowledge the blade.

2.0(5a)B

CSCuq40256

After an FI is reset, priority flow control on the IOM interface to a blade may become disabled.

Either reset the DCE that is affected or reboot the IOM.

2.1(3c)A

CSCuo76415

If a VM changes pinning from one FI to the other FI, the VM MAC address remains in the MAC address table of the original FI.

Either use static pinning with VMware, or manually clear the MAC address on the FI.

2.1(3b)A

CSCuj63448

When upgrading to a catalog that supports new DIMMs, some DIMM information isn’t displayed.

Reacknowledge the blade server.

1.4(4k)A

CSCuj74570

A Cisco UCS B420 M3 blade with a VIC 1240 and a port expander is successfully discovered in a chassis with a 2104XP IOM, even though it is unsupported. When upgrading to the 2204XP IOM, the blade reboots for discovery.

This issue has no known workaround.

The B420 M3 blade with port expander is not supported with the 2104XP IOM.

2.1(2a)B

CSCul38768

Cisco UCS Manager may display below invalid “unresponsive” management service transient fault, which is triggered by a software defect and clears after a few seconds:

This issue has no known workaround.

System will automatically recover assuming there is no actual HA condition failure.

2.1(2a)A

CSCul72408

During upgrade, the following issues occured:

• The IOM backplane ports show admin down in the Cisco UCS Manager GUI.
• The VIF's show non-participating
• The adapter shows DCE interfaces down.

Reboot the IOM in the affected chassis.

2.1(3a)A

CSCuu33864

When upgrading to a Cisco UCS Manager 2.2 release earlier than Release 2.2(6c), the fabric interconnect may become unresponsive at the loader prompt and not boot correctly. Various error messages may also be displayed.

Load a kickstart image either from the bootflash or through tftp. If the drive is still accessible, copy the debug plugin to it and load it. If the drive is no longer accessible, run the 'init system' command to re-initialize the SSD.

If the drive is still accessible it may be

possible to repair the file system corruption from the Linux shell as follows:

• unmount all partitions, /dev/sda3.. /dev/sda9. You might also have to unmount /dev/mtdblock3 prior to being able to unmount /dev/sda7.)
• run 'e2fsck -n -f /dev/sdaX' for X = {3,.., 9}
• Based on the number of errors, run 'e2fsck -y /dev/sdaX', X = {3,.., 9}, to attempt having the system repair the errors.It is recommended to copy critical data back to the SSD because data on disk could still be corrupted even after file system corruption has been fixed. Alternatively, you can run the init-system script to reinitialize the SSD.

Note This can take a while depending on the size of the drive.

2.1(3b)A

Resolved in 2.2(3j)A, 2.2(6c)A

CSCuo40713

The serial number is not displayed correctly when the following disks are used on Cisco UCS M3 servers:

• MZ6ER200HAGM/003DM0B
  (PID: UCS-SD200G0KS2-EP)
• MZ6ER400HAGL/003DM0B
  (PID: UCS-SD400G0KS2-EP)
• MZ6ER800HAGL/003DM0B
  (PID: UCS-SD800G0KS2-EP)

This issue has no known workaround.

When these disks are used, all information is displayed correctly except serial numbers. There is no impact to functionality.

2.1(3c)T

CSCuv03557

During Cisco UCS Manager upgrade, FI activation may fail with the following error:

Pre-Upgrade check failed. Insufficient free space in /var/tmp. Less than required 90%.

Use the following commands to determine if you are experiencing this issue:

If that shows more than 10% usage (2nd last column has usage), either reboot affected FI's or contact TAC for a workaround.

If you are affected, you can check if it is the file in question using the following from NXOS mode:

2.1(3a)A

Resolved in 2.1(3l)A

CSCva54957

A reboot is triggered without a “user -ack” when modifying a service profile that requires a reboot while shallow association is failing.

This issue does not have a workaround.

2.1(3a)A

Resolved in 2.1(3l)A

CSCuw78008

Web sessions may remain on after the Web Session Timeout period specified in the authentication option under user management in Admin tab.

If this issue occurs, kill the sessions using the command.

2.1(3a)A

Resolved in 2.1(3l)A

CSCur77746

On Cisco UCS B200 M3 and UCS B22 M3 blade servers running Cisco UCS Manager Release 2.1(3a) or later releases until Release 2.1(3g), the Board Controller firmware can be downgraded.

Upgrade to Cisco UCS Manager Release 2.1(3h), which prevents the downgrade of the Board Controller firmware.

2.1(3a)A

Resolved in 2.1(3h)A

CSCut09151
CSCut21914
CSCut08605

System Management BUS (SMBUS) traffic will collide on the SMBUS shared by the host and CIMC. As a result, certain system failures such as false thermal alarms may happen.

Avoid running OS applications that access the host SMBUS.

2.1(3a)B

Resolved in 2.1(3h)B

CSCui31011

Cisco UCS Manager reports FI and VIF down faults for Cisco UCS blade servers due to connectivity loss after reboot of subordinate FI.

Preferred:

Use CIMC to reset affected servers, which will force shallow discovery of the server and fix the issue without a reboot.

Alternatives:

• Reset IOM that is connected to the affected FI.

Note Any traffic going through that IOM for all blade servers on that chassis will be impacted unless failover is supported and configured to redirect traffic through the other working IOM/FI.

• Reacknowledge the affected servers.

Note Server will reboot.

2.1(2a)A

Resolved in: 2.1(2d) only, 2.1(3c) and onwards

CSCui87195

FLS cores, with the following message:

Upgrade adapter firmware to release 2.1.3a.

2.0(5c)

CSCui99339

When you upgrade to Release 2.1(3) from 2.1(2) using FW Auto Install to install infrastructure firmware, upgrade fails with an “Upgrade Validation Failed” error.

This issue has no known workaround.

2.1(3a)A

CSCuj84421

Installing Java 7 update 45 causes UCS Manager GUI failures. You may see errors similar to the following:

Downgrade to Java 7 update 40 or below. Previous releases are located on the Oracle Java Archive website.

2.1(1f)A

Resolved in 2.1(3b).

CSCul38768

Cisco UCS Manager may display below invalid “unresponsive” management service transient fault, which is triggered by a software defect and clears after a few seconds:

This issue has no known workaround.

System will automatically recover assuming there is no actual HA condition failure.

2.1(2a)A

CSCul99847

After upgrading to Release 2.1(3a), multiple vEthernet and vFC interfaces stay down with a nonParticipating error state.

This issue has no known workaround.

2.1(3a)A

CSCuo30572

In some cases, Intel v2 processors cause PSOD with Microsoft Windows 2008 R2 VM guests.

To work around this issue, use software MMU or upgrade to Cisco UCS Manager Release 2.1(3d).

2.1(3a)A

Resolved in 2.1(3d)A.

CSCus82914

Configuring SPAN on Cisco UCS fabric interconnects for all VLANS including vMotion and storage Ethernet may decrease performance.

To work around this issue, disable SPAN or remove VLANs with high traffic volumes.

2.1(3a)

Resolved in 2.1(3k)A

CSCuj84274

When Cisco UCS C-Series servers are configured with the IPMI interface enabled, a vulnerability in the Cisco Integrated Management Controller (CIMC) on the Cisco Unified Computing System Series Platforms may allow an unauthenticated, remote attacker to obtain the password hashes residing on the affected device.

The vulnerability is due to the implementation of an insecure authentication protocol. An attacker may exploit this vulnerability by sending a crafted packet to the CIMC of an affected device. An exploit may allow the attacker to receive a response from the CIMC that contains an RKMP message that will allow an attacker to obtain the password hashes for the system, which can then be used in an offline cracking attack.

Disable the IPMI interface on the CIMC running on the UCS system.

2.1(3a)A

Resolved in 3.1(1e)A

CSCun69556

In some rare conditions when rack servers are installed with a cluster-HA Cisco UCS FI configuration, if one FI goes down, the other FI may fail to become primary when rebooted successively.

To recover from this condition, use one of the following:

• Correct issue with first FI so that it is functional and then it can become primary after HA election is completed.
• Reconfigure the functional FI to run in standalone mode if a second functional FI is unavailable.

2.1(2c)A

CSCum15991

If a lsServerExtension file is created when exporting a Cisco UCS Manager configuration for a local service profile, the configuration cannot be imported in Cisco UCS Manager, Release 2.1(3).

Delete the lsServerExtension file from the exported configuration before importing.

2.1(2d)B

CSCuj63448

When upgrading to a catalog that supports new DIMMs, some of the DIMM information is not displayed.

Reacknowledge the blade server.

1.4(4k)A

CSCui17731

When a Cisco SFP 1GB Interface Converter GLC-T is inserted into a Cisco Nexus 2000 Series FEX port, it fails with a "SFP validation failed” error.

Contact Cisco TAC.

2.1(1b)A

Resolved in 2.1(3a).

CSCui21176

When a PSU is removed from a chassis with 4 PSUs, the power state on the chassis may show “redundancy-failed”.

This occurs when the Operability of the removed PSU continues to be displayed as Operable and Power State is displayed as On.

This issue has no known workaround.

2.1(2a)A

Resolved in 2.1(3a).

CSCui37900

When you upgrade to Release 2.1(2a), if a service profile template has the same authentication profile for both the iSCSI initiator and the iSCSI target, the high-availability connection between the FIs may not form, and the show cluster extended-state command displays one of the following errors:

The svc_sam_dme process may also crash.

To avoid this issue, ensure that the authentication profile is different between the iSCSI initiator and the target before upgrading.

If this issue occurs, contact Cisco TAC to revert to the previous code. Fix the authentication profile before resuming the upgrade.

2.1(2a)A

CSCuh61202

FC storage traffic through an IOM stops when the IOM is reset or reinserted, or the cable between the IOM and FI is removed or reinserted.

To avoid being impacted when upgrading from a release prior to 2.1(3a) or 2.2(1b), upgrade the server firmware before performing the corresponding infrastructure upgrade.

This caveat affects FC traffic on the Cisco 1240, Cisco 1280, and Cisco M81KR adapters and is an exception to the normal upgrade procedures found in Cisco UCS Manager upgrade guides.

For more details, please refer to CSCuh61202.

2.1(2a)B

CSCui45873

The primary FI may reboot with VIM core: mts_acquire_q_space() failing. This occurs when the ethpm MTS queues are full.

Reducing the batch size of the VM power cycle may alleviate the issue. For example, if the batch size is 80 VMs, reduce it to 20 VMs.

2.1(2a)A

Resolved in 2.1(3a).

CSCui45963

When you create a service profile or service profile template using the wizard, if you choose Create a Specific Storage Policy in the Local Storage area on the Storage page, some of the text and controls are truncated. This prevents the storage settings from being edited.

When you create a service profile or service profile template using the wizard, skip the storage page and complete the rest of the wizard. After the service profile or service profile template has been created, select the service profile, click the Storage tab, and then click Change Local Disk Configuration Policy in the Actions area to edit the storage settings.

2.1(2a)A

Resolved in 2.1(3a).

CSCui31011

Cisco UCS Manager reports FI and VIF down faults for Cisco UCS blade servers due to connectivity loss after reboot of subordinate FI.

Preferred:

Use CIMC to reset affected servers and force shallow discovery of server and fix the issue without a reboot.

Alternatives:

• Reset IOM connected to FI.

Note Any traffic going through IOM for all blade servers on that chassis are impacted unless failover is supported and configured to redirect traffic through other working IOM/FI.

• Reacknowledge affected servers.

Note Server will reboot.

2.1(2a)A

Resolved in 2.1(2d) only, 2.1(3c) and onwards

CSCui40766

Cisco UCS Manager fails to detect FlexFlash when enabled in a local disk configuration policy.

This issue has no known workaround.

2.1(2a)A

Resolved in 2.1(2c).

CSCui48112

FC VIF stays in unpinned state when you connect Cisco UCS C-series C220M3 with Cisco UCS Manager in dual-wire management mode.

This issue has no known workaround.

2.1(2a)A

Resolved in 2.1(3a).

CSCuh76699

The cmc pwrmgr process might stop working, causing a power cap failure in a Cisco UCS blade or chassis.

Reboot the IOM to restart the cmc pwrmgr process.

2.1(2a)A

CSCuh39242

The current severity level of Upper Noncritical and Upper Critical CPU thermal faults are incorrectly classified as minor faults.

The correct classification for these should be as informational warnings because they do not indicate a problem with the hardware health or performance of the server.

This issue has no known workaround. The faults can be safely ignored.

2.1(1e)B

CSCuh64817

On a VMware VMFex setup, when you remove one of the vNICs on a powered on VM, the change does not show on the Cisco UCS Manager side and it still shows the extra vNIC.

This issue has no known workaround.

2.1(1a)A

CSCuh73875

High CPU usage is seen on a Cisco UCS 6140XP Fabric Interconnect, especially when running the vlan_mgr process. This process starves other lower priority processes such as svc_sam_dme, which can prevent manual failover or configuration synchronization.

Disable SNMP on the FI, or do not poll SNMP on the FI IP or VIP.

2.1(1d)A

Resolved in 2.1(3a).

CSCuh94333

In rare situations, when configuring or unconfiguring FC/FCoE interface, the forwarding process crashes and the FI resets.

This issue has no known workaround.

2.0(2m)A

CSCui24182

In some cases, when Cisco UCS Manager takes more than five seconds to authenticate user credentials and respond to BMC, some blades will timeout with ‘login failed’ message when accessing the KVM.91904

To work around this issue, perform the following steps:

1. Use the KVM.zip file to access the KVM.

2. Wait for the KVM to become accessible again.

3. Use local user credentials defined in Cisco UCS Manager.

2.1(2a)A

CSCug51358

FCoE uplinks to the Cisco Nexus 5548 switch experience an MTS buffer leak between the port manager request high priority and fcoe_mgr due to an FC-Map mismatch between Cisco UCS Manager and the upstream Cisco Nexus 5548 switch. This buffer leak causes the FCoE uplinks to flap, and the VFCs are shown as error disabled.

1. Ensure that the Cisco Nexus 5000 Series switch FC-Map is the same as the Cisco UCS Manager FC-Map.

Note This change must be made on the Cisco Nexus 5000 Series switch.

2. Load the debug plug-in, and change the Cisco UCS Manager FC-Map using the Cisco NX-OS CLI.

These options require an FI reboot if the MTS buffer is leaking.
Disable the uplink ports, change the FC-Map, reboot the FI, and then reenable the uplink ports.

2.1(1d)A

Resolved in 2.1(3a).

CSCug89448

The tech support collection fails on the Cisco UCS Manager GUI. The process starts, but does not complete.

Use the show tech-support command in the Cisco UCS Manager CLI.

2.0(1s)A

CSCuh01579

When the server is rebooted, or the Cisco UCS reset option is used, the Cisco UCS B200 M2 Blade Server displays a USB composite device mounted in Windows. The drive letter assigned to this device varies, which might cause the clustering service to fail.

Disable the USB mass storage controller to prevent the USB composite device from mounting.

Note Disabling the USM mass storage controller also disables virtual CD/DVD ROM functionality.

2.1(1a)A

CSCuh12592

An FI might experience a bladeAG reload that results in a core dump due to lack of memory.

This issue has no known workaround.

2.1(1a)A

CSCuh28274

When installing XenServer 6.0.2 and adding the fNIC driver during the installation, an unrecoverable error occurs.

Install the fNIC driver after the OS is loaded.

2.1(1a)A

CSCug61578

When you remove the management cable from the primary FI, you are not able to view SNMP traps.

This issue has no known workaround and is only seen if the mgmt 0 interface of the primary FI goes down. If the mgmt interface of the secondary goes down, the traps are sent.

2.1(1b)A

CSCug63368

If the DHCP relay agent is installed as the gateway IP address instead of the HSRP IP address during a PXE boot, the PXE boot might fail in vPC environments with specific operating system configurations.

Clear the ARP entry on both vPC peers, and then configure the host to use different IP addresses on the OS from the lease assigned to the adapter.

2.1(1d)B

CSCug62535

BMC continuously logs the following message:

This issue has no known workaround.

2.1(1a)A

Resolved in 2.1(3a).

CSCug41743

With 3 DDR-1333 DIMMs installed per channel, the memory speed is not reduced to 1067 MHz under the following BIOS versions:

• B420M3.2.0.5.0.120720122110
• B420M3.2.0.5a.0.121720121433
• B420M3.2.1.1a.0.121720121615

Downgrade to one of the following BIOS versions:

• B420M3.2.0.4a.0.080920122056
• B420M3.2.1.1.0.100520121529

2.1(1d)B

Resolved in 2.1(3a).

CSCuh28239

In some rare conditions, during frequent MAC address changes between FIs, you might see a delay in learning MAC addresses.

If the MAC address changes between server ports on the same FI, the MAC address might point to an incorrect destination.

The learning delay issue has no known workaround. The problem is resolved automatically in 7-8 seconds when traffic occurs.

If the MAC address points to an incorrect destination, a subsequent frame from the server sourcing MAC address will fix the issue.

2.1(1b)A

Resolved in 2.1(2a).

CSCug59101

FI crashes that are due to a HAP reset are triggered by an NTP process crash.

To prevent FI reboots due to this issue, ensure that NTP server configured is reachable via DNS. If not, use the IPv4 address instead of the hostname to configure the NTP server.

2.0(1s)A

Resolved in 2.1(2a).

CSCug40776

Due to a memory leak that occurs when the NTP server is configured for DNS, but is not reachable via DNS, running the following commands may cause FI reboots:

• connect nxos
• show vlan
• show run

To prevent FI reboots due to this issue, ensure that the NTP server configured is reachable via DNS. If not, use the IPv4 address instead of the hostname to configure the NTP server.

2.0(3c)A

Resolved in 2.1(2a).

CSCug85569

When performing autoinstall on a server with the user ack policy, the server does not proceed with a graceful reboot. Some operating systems, such as Microsoft, may come up in recovery mode.

This issue has no known workaround.

2.1(1b)A

Resolved in 2.1(2a).

CSCug21589

When a fabric failover occurs on a VMware ESXi host with a fabric failover configured, the MAC address of the ESXi host is sent to the uplink switch. In some circumstances, the MAC address of the guest OS is not sent, which causes the network link to go down.

Generate a ping from the corresponding VM to any external hosts.

2.1(1a)T

CSCug26974

For servers with Cisco UCS M71KR-E or UCS CNA M71KR-Q adapters, changing ethernet adapters policy parameters can unnecessarily trigger a reboot on servers with that policy.

Avoid changing any ethernet adapter policy parameters except for Failback Timeout for servers equiped with Cisco UCS M71KR-E or UCS CNA M71KR-Q adapters.

2.1(1d)B

Resolved in 2.1(2a).

CSCuf78224

On a Cisco UCS B440-M2 server with Cisco UCS CNA M72KR-Q adapter card, VMware Auto Deploy 5.1 hangs during boot.

Using ESXi 5.0 gPXE instead of ESXi 5.1 iPXE may work under some conditions, but booting time is slow.

2.0(4b)A

Resolved in 2.1(2a).

CSCuf07670

Latency spikes might be reported on ESXi servers due to the virtual machines losing connections to the storage datastores. This condition occurs when storage datastores are presented through the FCoE uplinks using Cisco UCS and Cisco Nexus 5000 Series switches.

This issue has no known workaround.

2.1(1a)A

CSCud60746

When Call Home is enabled, there is a low chance that the system may run out of memory. This occurs after 365,000 Call Home messages have been sent out, or approximately 4700 inventory messages are generated. This could result in one of the following:

• The primary Fabric Interconnect may reboot with the following reset reason:

• Upgrading the firmware in the primary FI may fail due to the /isan folder filling up.

To reduce the frequency of memory leaks caused by Call Home messages:

1. Reduce the number of Call Home messages by disabling any unnecessary Call Home policies or modifying alert groups and/or profile levels.

2. Reduce inventory message generation.

To completely prevent any memory leaks due to Call Home, disable Call Home.

If the primary FI firmware upgrade fails, contact Cisco TAC to cleanup the /isan directory.

2.0(2a)A

Resolved in 2.1(1f).

CSCuf17523

Cisco NX-OS reports that the FI port is down, but the interface counters show it is still receiving traffic. This condition occurs when the port speed is changed when the port is administratively down.

Administratively enable and administratively disable the port.

2.1(1a)A

Resolved in 2.1(2a).

CSCuf01402

Modifying a service profile with two iSCSI vNIC targets defined prompts for a reboot before the second target can be configured.

To configure the second target, modify any attribute in the service profile. Accept the reboot prompt to configure the targets properly.

2.1(1a)A

Resolved in 2.1(2a).

CSCud59815

When assigning FC ports on the Cisco 6200 Series FI using the slider, all ports are enabled by default. If the number of ports allocated is greater than the license allotment, license faults are generated and ports can go into the grace period.

Disable the FC ports that are not actually going to be used.

2.0(2q)A

CSCud75506

The UUID is translated incorrectly when you upgrade ESXi from version 4.1 or 5.1 on the Cisco UCS B200 M3, B220 M3, or B440 M3 Blade Servers.

This is a display issue only, and does not affect the service profiles associated with the blades.

This issue has no known workaround.

2.0(2r)A

CSCue49366

Symptoms include transient Cisco UCS Manager faults related to a shared Cisco UCS chassis I2C devices, such as a fan module, PSU, or the shared storage located in the Cisco UCS Chassis SEEPROM. These faults may include the terms fan inoperable, PSU, or shared storage.

The detailed or brief tech-support command shows the chassis segment norxack count is high and increasing (hundreds or thousands depending on IOM uptime).

This high rate of i2c access errors shows that the IOM is not backing off of the shared I2C bus for the required amount of time. As a result, one or both of the IOMs are interfering with each other's access to shared I2C resources and neither may be able to get useful work done.

This issue has no known workaround if uninterrupted high-availability service is desired.

If nonredundant operation is tolerable, you could pull one IOM from the chassis.

Powering down the entire chassis for 3 minutes and reapplying power might clear the condition in the short-term.

2.1(1a)B

Resolved in 2.1(2a).

CSCud70315

When a port channel member transitions between up and down states, fabric channel traffic on the port channel is dropped. A SCSI timeout occurs, and the SCSI layer triggers the recovery.

This issue has no known workaround.

Proper SCSI timeout values will help in recovery.

2.0(4a)A

CSCue29352

When you try to change the boot order without checking the option’s "local storage change" and "Reboot on Boot Order Change,” the server is listed in the pending activities list.

Check the Reboot on Boot Order Change check box to trigger a server reboot.

2.0(4d)A

CSCud89583

The Cisco UCS B440 Blade Server that is running Citrix XenServer 6.0.2 with E7-4830 and all C states disabled has a "CATERR_N" error and freezes. The host IP address of the blade is unreachable, and KVM, the front panel dongle, and SOL do not function.

Power cycle the blade.

2.0(3a)A

Resolved in 2.1(3a).

CSCud74915

After you create a new service profile using VM-FEX, a duplicate VIF prevents you from connecting to the original blade server.

1. Disassociate both blades from their service profiles.

2. Reassociate only one blade.

3. Recreate the service profile for the other blade.

2.0(3a)A

CSCuf90470

When Call Home is enabled, Online Insertion and Removal (OIR) or failure of hardware modules (such as fans, power supply, or GEM) in the FI may cause the FI to reboot.

Disable the Call Home function.

2.1(1b)A

Resolved in 2.1(1e).

CSCud27864

BIOS will no longer hang during POST when the memory mapped IO above 4 GB is enabled and CSB-MEZ-QLG-03 is present in the blade.

Disable memory mapped I/O above 4 GB.

2.1(1b)A

Resolved in 2.1(3a).

CSCue47159

In a Cisco UCS chassis with one or more empty slots, Cisco UCS Manager may show a critical alert and a "FSM Failed" warning for a slot that has no blade.

1. Insert a spare blade into the slot that has the alert.

2. After the discovery process has completed successfully, check for the error message if it appears.

3. Decommission the server from Cisco UCS Manager by using the "Server Maintenance" link.

4. After the decommission process has completed successfully, remove the server from the slot.

2.1(1a)A

Resolved in 2.1(2a).

CSCue65877

Under certain conditions when upgrading to Release 2.1(1), the storaged daemon generates core files.

This issue has no known workaround. The storaged daemon is a monitored daemon that restarts automatically. System operation continues and performance is not negatively impacted.

2.1(1a)A

Resolved in 2.1(2a).

CSCud86528