Release Notes for Cisco UCS Manager, Release 2.2

Release Notes for Cisco UCS Manager, Release 2.2

First Published: December 12, 2013

Last Updated: May 24, 2016

This document describes system requirements, new features, resolved caveats, known caveats and workarounds for Cisco UCS

Manager software Release 2.2. This document also includes the following:

•

Current information that became available after the technical documentation was published

•

Related firmware and BIOSes on blade and rack servers and other Cisco Unified Computing System (UCS) components associated with the release

Use this release note as a supplement with the other documents listed in documentation roadmaps: http://www.cisco.com/go/unifiedcomputing/b-series-doc http://www.cisco.com/go/unifiedcomputing/c-series-doc

Contents of the various bundles for this release are described in:

Release Bundle Contents for Cisco UCS Software, Release 2.2

Make sure to review other available documentation on Cisco.com to obtain current information on Cisco UCS Manager.

Contents

This document includes the following sections:

•

Revision History, page 2

•

•

•

•

•

•

•

•

Introduction, page 4

System Requirements, page 4

Updating Cisco UCS Releases, page 5

Internal Dependencies, page 8

Capability Catalog, page 13

New Hardware Features in Release 2.2, page 22

New Software Features in Release 2.2, page 24

Behavior Changes, page 28

Cisco Systems, Inc.

www.cisco.com

Revision History

•

•

•

•

•

Security Fixes, page 31

Resolved Caveats, page 31

Open Caveats, page 60

Known Limitations, page 105

Related Documentation, page 109


Table 1

shows the revision history:

Table 1 Online Change History

Date

December 12, 2013

December 20, 2013

February 9, 2014

February 19, 2014

April 2, 2014

April 10, 2014

May 16, 2014

May 22, 2014

May 24, 2014

May 30, 2014

June 18, 2014

August 1, 2014

August 21, 2014

Description

Created release notes for Cisco UCS Manager, Release 2.2(1b).

Added additional caveats.

Updated release notes for Cisco UCS Software Release 2.2(1c).

Updated release notes for Catalog Release 2.2.1c.T.

Updated release notes for Catalog Release 2.2.1d.T.

Updated release notes for Cisco UCS Software Release 2.2(1d).

Added CSCuo78883 to Open Caveats for release 2.2(1d).


Updated to include additional Open Caveats for Cisco UCS Software Release

2.2(2c).

Modified several Caveat descriptions.

Added CSCul74278 to 2.2(1b) Open Caveats and to 2.2(2c) Resolved Caveats; added ‘Behavior Changes’ section; and corrected three items in Servers list.

Updated release notes for Cisco UCS Software Release 2.2(1e).

Added B22 PID support and updated PID support for other servers in Version

Mapping table for 2.2(2c); added ‘Chassis’ section to Internal Dependencies table; added UCSB-5108-AC2, -DC2, and -HVDC to New Hardware Features section; added note to Upgrade section for CSCud81176 consideration; added CSCuj81245 to 2.2(1b) Open Caveats table and to 2.2(2c) Resolved Caveats table.

September 8, 2014 Updated release notes for Cisco UCS Software Release 2.2(3a).

September 11, 2014 Updated release notes for Cisco UCS Software Release 2.2(2d); updated the

Recommended Minimum Software in Table 4.

October 09, 2014

October 24, 2014

October 31, 2014

November 10, 2014

November 17, 2014

December 4, 2014

Updated release notes for Cisco UCS Software Release 2.2(3b).

Updated release notes for Cisco UCS Software Release 2.2(2e) and 2.2(3c).

Updated release notes for Cisco UCS Software Release 2.2(1f).

Added CSCuo60330 to 2.2(3a) resolved caveats.

Added missing 2.2(3c) adapter to Internal Dependencies table.

Updated release notes for Catalog Release 2.2.3b.T.


2 OL-31127-01

Table 1 Online Change History

Date

December 22, 2014

January 13, 2015

March 4, 2015

March 18, 2015

March 20, 2015

April 15, 2015

May 14, 2015

May 26, 2015

June 5, 2015

June 15, 2015

June 30, 2015

Description

Updated release notes for Cisco UCS Software Release 2.2(1g) and 2.2(3d).

Updated first affected bundle for CSCuq20755.


Updated the New Hardware Features list for Release 2.2(3a).


Updated the Recommended Software Version for C260 and C460 M2 servers.

Updated release notes for Cisco UCS Software Releases 2.2(3g) and 2.2(4b).

Added CSCus10255 to the Resolved Caveats for Cisco UCS Software Release

2.2(4b).

Updated release notes for Cisco UCS Software Release 2.2(5a).

Updated release notes for Cisco UCS Software Release 2.2(1h).

Updated the Capability Catalog and New Hardware Features sections with information about Cisco 12G SAS Modular RAID Controller.

Updated the Resolved Caveats for 2.2(4b).

July 10, 2015 Added CSCuv04436 to the Open Caveats for 2.2(4b) and 2.2(5a). This issue has a software advisory associated with it.


July 17, 2015

July 30, 2015

August 27, 2015



September 11, 2015 Updated the Behavior Changes section for Release 2.2(4b).

Updated the Resolved and Open Caveats for Release 2.2(5c).

Added CSCus81832 to the list of Resolved Caveats for Release 2.2(4b).

September 22, 2015 Added CSCus73196 to the list of Open Caveats for Release 2.2(3d).

September 25, 2015 Updated release notes for Cisco UCS Software Release 2.2(3h).

October 8, 2015 Updated release notes for Cisco UCS Software Release 2.2(6c).

October 17, 2015

November 04, 2015

November 16, 2015

Added CSCuv45173 to the Known Limitations table. This issue has a software advisory associated with it.



Updated release notes for Cisco UCS Catalog Release 2.2.6e.T.

December 14, 2015

December 16, 2015

January 18, 2016


Updated the Behavior Changes section for Release 2.2(6c).

February 02, 2016

February 09, 2016

February 29, 2016

March 31, 2016

April 18, 2016


Updated release notes for Cisco UCS Software Release 2.2(3j).

Updated release notes for Cisco UCS Software Release 2.2(6g).


Updated the New Software Features description for Release 2.2(7b).



OL-31127-01

Introduction

Table 1

Date

May 13, 2016

May 18, 2016

May 24, 2016

Online Change History

Description

Updated release notes for Cisco UCS Software Release 2.2(3k).

Updated release notes for Cisco UCS Software Release 2.2(6i).

Added CSCuz74973 to the Open Caveats for Cisco UCS Software Release 2.2(7b), and updated the workaround for CSCut63966.

Introduction

Cisco UCS™ Manager provides unified, embedded management of all software and hardware components of the Cisco Unified

Computing System™ (Cisco UCS) across multiple chassis, rack servers, and thousands of virtual machines. Cisco UCS

Manager manages Cisco UCS as a single entity through an intuitive GUI, a command-line interface (CLI), or an XML API for comprehensive access to all Cisco UCS Manager functions.

This introductory information is provided in the following sections:

•

•

•

System Requirements, page 4

Updating Cisco UCS Releases, page 5

Hardware and Software Interoperability, page 8

System Requirements

To use Cisco UCS Manager your computer must meet or exceed the following minimum system requirements:

•

•

The Cisco UCS Manager GUI is a Java-based application which requires Java Runtime Environment 1.6 or later.

Cisco UCS Manager uses web start and supports the following web browsers:

•

•

–

–

Microsoft Internet Explorer 9.0 or higher

Mozilla Firefox 7.0 or higher

–

Google Chrome 14.0 or higher

Adobe Flash Player 10 or higher is required for some features

•

Cisco UCS Manager is supported on the following operating systems:

–

Microsoft Windows 7 with minimum 4.0 GB memory

–

Red Hat Enterprise Linux 5.0 or higher with minimum 4.0 GB memory

Cisco UCS Central integration:

–

–

Cisco UCS Manager Release 2.2(1) and 2.2(2) can only be registered with Cisco UCS Central, Release 1.1(1b) or higher.

Cisco UCS Manager Release 2.2(3) and later releases can only be registered with Cisco UCS Central, Release 1.2(1a) or higher.

–

Cisco UCS Manager Release 2.2(7) and later releases require Cisco UCS Central, Release 1.4(1b) or higher.

Note

For more information, see the Feature Support Matrix section of the Cisco UCS Central Installation and Upgrade

Guides .


4 OL-31127-01

Introduction

Updating Cisco UCS Releases

The Cisco UCS Manager A bundle software (Cisco UCS Manager, Cisco NX-OS, IOM firmware) can be mixed with previous

B or C bundle releases on the servers (host firmware (FW), BIOS, CIMC, adapter FW and drivers).

Table 2 lists the mixed A, B, and C bundle versions that are supported:

2.2(3)

2.2(4)

2.2(5)

2.2(6)

2.2(7)

2.0(5)

2.1(1)

2.1(2)

2.1(3)

2.2(1)

2.2(2)

Table 2 Mixed Cisco UCS Releases Supported

Infrastructure Versions (A Bundles)

Host FW

Versions (B or C

Bundles) 2.0(1) 2.0(2) 2.0(3) 2.0(4) 2.0(5) 2.1(1) 2.1(2) 2.1(3) 2.2(1) 2.2(2) 2.2(3) 2.2(4) 2.2(5) 2.2(6) 2.2(7)

2.0(1)

2.0(2)

Yes

—

—

Yes

—

—

—

—

—

—

Yes

Yes

Yes

Yes

Yes

Yes

—

—

—

—

—

—

—

—

—

—

—

—

2.0(3)

2.0(4)

—

—

—

—

Yes

—

—

Yes

—

—

Yes

Yes

Yes

Yes

Yes

Yes

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

Yes

—

—

—

Yes

Yes

—

—

Yes

Yes

Yes

Yes

—

Yes

—

Yes

—

Yes

—

Yes

—

Yes

—

Yes Yes

Yes Yes Yes Yes Yes Yes Yes Yes Yes

— Yes Yes Yes Yes Yes Yes Yes Yes

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

Yes

—

—

—

—

—

—

Yes

Yes

—

—

—

—

—

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes Yes Yes Yes Yes

—

—

—

—

Yes

Yes

Yes

Yes

1

1

1

Yes

Yes

1

Yes

1

Yes

1

Yes

Yes

Yes

Yes

1

1

1

Yes

Yes

Yes

Yes

1

1

1

1. Beginning with Cisco UCS Manager Release 2.2(4), and for M4 servers, a lower version of the infrastructure A bundle will be compatible with the previous version and higher version of B and C server bundles. For example, the Cisco UCS Manager Release 2.2(4)A bundle will be supported with any of the following B bundles for B200-M4 servers: 2.2(3)B, 2.2(4)B, 2.2(5)B, 2.2(6)B, 2.2(7)B.

Note

For M1, M2, M3 servers, only N, N-1 cross-version firmware is supported. For example, for B200 M3 servers, the

2.2(4)A bundle will be supported with 2.1(1)B, 2.1(2)B, 2.1(3)B, 2.2(1)B, 2.2(2)B, 2.2(3)B, and 2.2(4)B bundles).

•

•

If upgrading from a pre-2.2(1b) release and running Management Firmware Pack, refer to open caveat information for

CSCud81176 in

Release Notes for Cisco UCS Software, Release 2.1

.

If an environment includes a mix of servers, refer to CSCuh61202 and CSCus64439 for caveat information.

Note

To avoid this issue, 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.

•

During manual or auto firmware installs, the fabric interconnect may fail to reload after a reboot or power cycle. Refer to

CSCut63966 for caveat information.


OL-31127-01

Introduction

•

•

•

•

•

A workaround for this issue is to modify a script file, which is called before any reboots occur. This modification can limit the chances of experiencing this issue during an upgrade to firmware versions in which this issue is resolved. Contact Cisco

TAC for more information regarding this workaround, and to recover from this issue.

A mix of servers running different B-bundles may be run with a single A-bundle. However, any given server must be running the entire B/C-bundles (with associated drivers). Example: mixing the 2.1(2)B BIOS with the 2.1(3)B CIMC on a server is not supported.

The OS hardware and software interoperability is relative to the B/C-bundle on any given server. To see what OS is supported, see the Hardware and Software Interoperability documentation associated with the B-bundle version.

The A-bundle version must be at or above the same version(s) of any B/C-bundles running on the servers (see

Table 2 ).

This applies for patch levels as well, even though they are not displayed on the table. For example, you can mix 2.1(1f)A with 2.1(1b)B, but you cannot mix 2.1(1b)A with 2.1(1f)B.

Some features introduced in Cisco UCS Release 2.2 require that both the A-bundle version and the B/C-bundle versions be upgraded to the same version. For example, the lower power budget supported for Cisco UCS Release 2.2 is not supported for servers using 2.1 firmware.

In Cisco UCSM Release 2.2 and later releases, the adapter firmware version is different from the Cisco UCSM Release version.

6


OL-31127-01

Introduction

Minimum B/C Bundle Version Requirements for Cisco UCS Manager Features

The following Cisco UCS Manager 2.2 features require the specified minimum B/C bundle version to perform expected operations:

Table 3 Minimum Required B/C Bundle Version

Feature

IPv6 Management Support

B Bundle Version

2.2(1b)B usNIC for Low Latency 2.2(1b)B

Support for Virtual Machine Queue (VMQ) 2.2(1b)B

VM-FEX for Hyper-V Management with

Microsoft SCVMM

2.2(1b)B

Secure Boot

Local Storage Management

2.2(1b)B

2.2(1b)B

Flash Adapters and HDD Firmware

Management

Precision Boot Order Control

2.2(1b)B

2.2(1b)B

Trusted Platform Module (TPM) Inventory 2.2(1b)B

2.2(1b)B DIMM Blacklisting and Correctable Error

Reporting

Netflow monitoring support:

•

•

•

•

•

Cisco UCS VIC 1240 and 1280

Cisco UCS VIC 1225

Cisco UCS VIC 1227

Cisco UCS VIC 1340 and 1380

Cisco UCS VIC 1225T and 1227T

Stateless Offload for Overlay Networks

(NVGRE / VXLAN)

2.2(2c)B

—

—

2.2(3a)B

—

2.2(3a)B

C Bundle Version

2.2(1b)C

2.2(1b)C

2.2(1b)C

2.2(1b)C

2.2(1b)C

2.2(1b)C

2.2(1b)C

2.2(1b)C

2.2(1b)C

2.2(3a)C

—

2.2(2c)C

2.2(3a)C

—

2.2(4b)C

—

Note

Supported only on Cisco UCS

VIC 1340 and VIC 1380 adapters.

BIOS secure boot support

CIMC secure boot support

RDMA over Converged Ethernet (RoCE) support for Microsoft SMB Direct

2.2(1b)B

2.2(1b)B

2.2(4b)B

Note

Supported only on Cisco UCS

VIC 1340 and VIC 1380 adapters.

LLDP support for Fabric Interconnect vEthernet Interfaces

Policy-based port error handling

Advanced Local Storage Configuration

2.2(4b)B

2.2(4b)B

2.2(4b)B

2.2(3a)C

2.2(3a)C

—

2.2(4b)C

2.2(4b)C

2.2(4b)C


OL-31127-01

Internal Dependencies

Table 3 Minimum Required B/C Bundle Version

Feature

Pass-through capability for Gen 3 Fusion IO

Mezzanine cards

Fabric Interconnect traffic evacuation

B Bundle Version

2.2(4b)B

2.2(4b)B

2.2(4b)B Per Fabric Interconnect chassis reacknowledgment

Trusted Platform Module (TPM) configuration support

Consistent Device Naming (CDN) support

2.2(4b)B

2.2(4b)B

C Bundle Version

—

2.2(4b)C

—

2.2(4b)C

2.2(4b)C

Note

CDN is supported only on Windows

2012 R2.

CIFS authentication protocol support for

Scriptable vMedia

Support for Service Profile migration with

UEFI boot mode

Support NVGRE with IPv6 and VMQ

Support usNIC with Intel

®

MPI

2.2(4b)B

2.2(4b)B

2.2(4b)B

2.2(4b)B

2.2(4b)C

2.2(4b)C

—

2.2(4b)C

For detailed instructions for updating the Cisco UCS software and firmware, see the appropriate Upgrading Cisco UCS document for your installation.

Hardware and Software Interoperability

For a complete list of hardware and software interdependencies, see the Hardware and Software Interoperability for UCSM

Managed Servers for a specific Cisco UCS Manager release, here: http://www.cisco.com/c/en/us/support/servers-unified-computing/unified-computing-system/products-technical-reference-list

.html.


Table 4

shows interdependencies between the hardware and versions of Cisco UCS Manager. Server FRU items such as DIMMs are dependent on their server type, and chassis items such as fans and power supplies work with all versions of Cisco UCS

Manager.

Table 4 Internal Dependencies

Recommended Minimum

Software Version

1

Recommended Software

Version Component

Servers

B22 M3 E5-2400

B22 M3 E5-2400 v2

2.1(3g)

2.2(2e)

2.2(7b)

2.2(7b)


8 OL-31127-01


Table 4

Component

B200 M1 and M2

B200 M3 E5-2600

B200 M3 E5-2600 v2

B200 M4

B200 M4 E5-2600 v4

B230 M1 and M2

B250 M1 and M2

B260 M4 E7-2800 v2

B260 M4 E7-4800 v2

B260 M4 E7-8800 v2

B260 M4 E7-4800 v3

B260 M4 E7-8800 v3

B420 M3 E5-4600

B420 M3 E5-4600 v2

B440 M1 and M2

B420 M4 E5-4600 v3

B460 M4 E7-4800 v2

B460 M4 E7-8800 v2

B460 M4 E7-4800 v3

B460 M4 E7-8800 v3

C22 M3 and M3L

C24 M3, M3L, and M3S2

C200 M2 and M2 SFF

C210 M2

C220 M3

2

C220 M4

C220 M4 E5-2600 v4

C240 M3

2

C240 M4

C240 M4 E5-2600 v4

C250 M2

C260 M2

C420 M3

C460 M2

C460 M4 E7-2800 v2

C460 M4 E7-4800 v2

C460 M4 E7-8800 v2

C460 M4 E7-4800 v3

C460 M4 E7-8800 v3

Internal Dependencies (continued)

2.1(3g)

2.1(3g)

2.1(3j)

2.1(3j)

2.1(3g)

2.2(3h)

2.2(7b)

2.1(3g)

2.2(3h)

2.2(7b)

2.1(3j)

2.1(3j)

2.1(3j)

2.1(3j)

2.2(2e)

2.2(2e)

2.2(2e)

2.2(5d)

2.2(5d)

2.1(3g)

2.2(2e)

2.1(3g)

2.2(5d)

2.2(2e)

2.2(2e)

2.2(5d)

2.2(5d)



1

2.2(2e)

2.1(3g)

2.2(2e)

2.2(3h)

2.2(7b)

2.1(3g)

2.1(3g)

2.2(2e)

2.2(2e)

2.2(2e)

2.2(5d)

2.2(5d)


Version

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(6i)

2.2(1h)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(6i)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)


OL-31127-01


Table 4 Internal Dependencies (continued)



1

Component

Adapters

UCS 82598KR-CI

UCS M71KR-E

UCS M71KR-Q

UCS M81KR

UCS NIC M51KR-B

UCS CNA M61KR-I

3

UCS CNA M72KR-Q

UCS CNA M72KR-E

UCS-VIC-M82-8P

UCSB-MLOM-40G-01

UCSB-MLOM-PT-01

UCSB-MLOM-40G-03

UCSB-VIC-M83-8P

UCSC-MLOM-CSC-02

UCSB-MEZ-ELX-03

UCSB-MEZ-QLG-03

UCSC-PCIE-CSC-02

UCSC-F-FIO-1000MP

UCSC-F-FIO-1300MP

UCSC-F-FIO-2600MP

UCSC-F-FIO-5200MP

UCSB-FIO-1600MS

UCSB-FIO-1300MS

UCSC-INVADER-3108

UCSC-NYTRO-200GB

UCSC-MLOM-C10T-02

UCSC-PCIE-C10T-02

UCSC-F-FIO-785M

UCSC-F-FIO-365M

UCSC-F-FIO-1205M

UCSC-F-FIO-3000M

UCSC-F-FIO1000PS

UCSC-F-FIO1300PS

UCSC-F-FIO2600PS

UCSC-F-FIO5200PS

UCSC-F-FIO-6400SS

UCSC-F-FIO-3200SS

Chassis

4

N20-C6508

UCSB-5108-DC

UCSB-5108-AC2

UCSB-5108-DC2

UCSB-5108-HVDC

2.1(3g)

2.1(3g)

2.1(3g)

2.1(3g)

2.2(3a)

2.1(3g)

2.2(3a)

2.2(3a)

2.2(3a)

2.2(4b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)


Version

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

1


OL-31127-01





1

Component

Fabric Interconnect

UCS 6120XP

UCS 6140XP

UCS 6248UP

UCS 6296UP

Fabric Extender or I/OM

UCS 2104

UCS 2208XP

UCS 2204XP

Cisco Nexus 2232PP

Cisco Nexus 2232TM-E

Fabric Interconnect Expansion Modules

N10-E0440

N10-E0600

N10-E0080

N10-E0060

UCS-FI-E16UP

Power Supplies

UCSB-PSU-2500HVDC

UCSC-PSU-930WDC

UCSC-PSU1-770W

UCSC-PSU2-1400

UCSC-PSU2V2-650W

UCSC-PSU2V2-1200W

NVME SSD Drives

UCS-SDHPCIE800GB

UCS-SDHPCIE1600GB

5

10-GB Connections

SFP-10G-SR, SFP-10G-LR

SFP-H10GB-CU1M

SFP-H10GB-CU3M

SFP-H10GB-CU5M

SFP-H10GB-ACU7M

SFP-H10GB-ACU10M

FET-10G

SFP-H10GB-ACU7M=

SFP-H10GB-ACU10M=

SFP-H10GB-CU2M

SFP-H10GB-CU1-5M

SFP-H10GB-CU2-5M

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)


Version

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)


OL-31127-01



Component



1

SFP-10G-AOC1M

SFP-10G-AOC2M

SFP-10G-AOC3M

SFP-10G-AOC5M

SFP-10G-AOC7M

SFP-10G-AOC10M

2.2(7b)

8-GB Connections (FC Expansion Module N10-E0060)

DS-SFP-FC8G-SW

DS-SFP-FC8G-L

2.2(7b)

4-GB Connections (FC Expansion Module N10-E0080)

2.2(7b) DS-SFP-FC4G-SW

DS-SFP-FC4G-LW

1-GB Connections

GLC-T (V03 or higher)

GLC-SX-MM

GLC-LH-SM

2.2(7b)


Version

2.2(7b)

2.2(7b)

2.2(7b)

2.2(7b)

1.

This is the minimum server bundle recommended for this hardware in a mixed firmware configuration, assuming the infrastructure is at the recommended software version.

2.

See the Software Advisory for the minimum firmware level required on the Cisco UCS C220 M3 and Cisco UCS C240 M3.

3.

N20-AI0002, the Cisco UCS 82598KR-CI 10-Gb Ethernet Adapter, is not supported on the B440 server but is still available for other models. We suggest you use the Cisco UCS CNA M61KR-I Intel Converged Network Adapter in place of the Cisco

UCS 82598KR-CI 10-Gb Ethernet Adapter.

4.

Recommended minimum software versions do not take into account mixed environments; to determine the minimum software version for your mixed environment, refer to

“Mixed Cisco UCS Releases Supported” , Table 2 on page 5

.

5.

For 1.6 TB HGST NVME drives, when used with B200-M4 blade servers, the minimum required board controller version is

12, which is bundled with Cisco UCS Manager Release 2.2(7b)B.

1


OL-31127-01

Capability Catalog


The Cisco UCS Manager uses the catalog to update the display and configurability of server components such as newly qualified DIMMs and disk drives. The Cisco UCS Manager Capability Catalog is a single image, but it is also embedded in

Cisco UCS Manager. Cisco UCS Manager 2.2(x) releases work with any 2.2(x) catalog file, but not the 1.x, 2.0 or 2.1 catalog versions. If a server component is not dependent on a specific BIOS version, using it and having it recognized by Cisco UCS

Manager is primarily a function of the catalog version. The catalog is released as a single image in some cases for convenience

purposes in addition to being bundled with UCS infrastructure releases. See Table 5 for details on the mapping of versions to

bundles.

Table 5 Version Mapping

UCS Release

2.2(7b)

Catalog File

ucs-catalog.2.2.7b.T.bin

Adds Support for PID

Cisco UCS B200 M4, C220 M4, and C240 M4 CPUs

•

UCS-CPU-E52699E

Additional Parts Qualified for PID

—

•

UCS-CPU-E52698E

•

•

•

•

•

UCS-CPU-E52697AE

UCS-CPU-E52697E

UCS-CPU-E52695E

UCS-CPU-E52690E

•

•

•

•

UCS-CPU-E52683E

UCS-CPU-E52680E

UCS-CPU-E52667E

UCS-CPU-E52660E

•

•

•

•

UCS-CPU-E52658E

UCS-CPU-E52650E

UCS-CPU-E52650LE

UCS-CPU-E52643E

•

•

•

•

UCS-CPU-E52640E

UCS-CPU-E52637E

UCS-CPU-E52630E

UCS-CPU-E52630LE

•

•

UCS-CPU-E52623E

UCS-CPU-E52620E

UCS-CPU-E52609E

Storage Controller

•

UCSB-LSTOR-PT

PCIe adapter card

•

UCSC-SAS12GHBA


OL-31127-01


Table 5

UCS Release

2.2(7b) continued

Version Mapping (continued)

Catalog File Adds Support for PID

Memory

•

UCS-ML-1X324RU-G

Note

Requires Cisco UCS Manager

Release 2.2(7) or later releases.

•

UCS-ML-1X644RU-G


Memory

•

UCS-ML-1X324RU-A

•

•

•

•

•

UCS-MR-1X081RU-G

UCS-MR-1X162RU-A

UCS-MR-1X162RU-G

UCS-MR-1X322RU-G

UCS-MR-2X162RY-E

Note

Requires Cisco UCS Manager

Release 2.2(7) or later releases.

•

•

•

•

•

UCS-MR-1X081RV-A

UCS-MR-1X161RV-A

UCS-MR-1X322RV-A

UCS-ML-1X324RV-A

UCS-ML-1X644RV-A

Drives

•

UCS-SD400GBK9

•

•

•

•

•

•

•

•

•

•

UCS-HD1T7KL12G

UCS-HD2T7KL12G

UCS-HD4T7KL12G

UCS-HD2T7KL6GA

UCS-HD10T7KL4K

UCS-HD4TBK9

A03-D300GA2

A03-D600GA2

UCS-HDD900GI2F106

UCS-SD16TBKS4-EV

1


OL-31127-01


Table 5

UCS Release

2.2(7b) continued



GPU

•

UCSB-GPU-M6

•

UCSC-GPU-M60

Crypto Card

•

CSB-MEZ-INT8955

NVME Drives:

•

UCS-SDHPCIE800GB

•

UCS-SDHPCIE1600GB

2.2(6i)

2.2(6g)

2.2(6f)

2.2(6e)

—

2.2(6d)


Note

For 1.6 TB HGST NVME drives, when used with

B200-M4 blade servers, the minimum required board controller version is 12, which is bundled with Cisco UCS

Manager Release 2.2(7b)B.

Qlogic network adapters

•

UCSC-PCIE-QNICBT

•

UCSC-PCIE-QNICSFP

—

— ucs-catalog.2.2.6f.T.bin

— ucs-catalog.2.2.6e.T.bin

—

—

—

—

Cisco UCS VIC adapters

•

UCSC-PCIE-C40Q-03

•

—

UCSC-MLOM-C40Q-03

—

—

—

—

•

•

•

•

•

•

•

UCS-SD480GBKS4-EV

UCS-SD16TBKS4-EV

UCS-HD600G10KS4K

UCS-MR-2X162RY-E

UCS-SD480G12S3-EP

UCS-SD120GBKS4-EV

UCS-SD16TB12S3-EP

—

—


OL-31127-01


Table 5

UCS Release

2.2(6c)

2.2(5d)

2.2(5c)

2.2(5b)


Catalog File

ucs-catalog.2.2.6c.T.bin

— ucs-catalog.2.2.5b.T.bin

—

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

—


Additional Memory

•

•

UCS-ML-1X324RU-A

UCS-ML-1X324RU-G

UCS-ML-1X648RU-G

•

•

•

UCS-MR-1X162RY-A

UCS-MR-2X162RX-C

•

UCS-MR-2X162RY-E

Drives

UCS-HD12TB10K12G

UCS-HD1T7K12G

UCS-HD1T7K6GA

UCS-HD2T7K12G

UCS-HD2T7KL12G

UCS-HD300G10K12G

UCS-HD300G15K12G

UCS-HD450G15K12G

UCS-HD4T7KL12G

UCS-HD600G10K12G

UCS-HD600G15K12G

UCS-HD8T7KL4K

UCS-HD900G10K12G

UCS-SD16TB12S3-EP

UCS-SD16TB12S4-EP

UCS-SD16TBKS4-EV

UCS-SD240GBKS4-EV

UCS-SD38TBKS4-EV

UCS-SD400G12S4-EP

UCS-SD480G12S3-EP

UCS-SD480GBKS4-EV

UCS-SD800G12S4-EP

UCS-SD960GBKS4-EV

—

Cisco UCS B420 M4 (UCSB-B420-M4)

•

Memory: UCS-MR-1X648RU-A


1


OL-31127-01


Table 5

UCS Release

2.2(5a)


Catalog File

ucs-catalog.2.2.5a.T.bin



•

CPU: UCS-CPU-E5-4660D,

UCS-CPU-E5-4650D,

UCS-CPU-E5-4640D,

UCS-CPU-E5-4620D,

UCS-CPU-E5-4610D,

UCS-CPU-E5-4669D,

UCS-CPU-E5-4667D,

UCS-CPU-E5-4655D,

UCS-CPU-E5-4627D

•

•

Memory: UCS-MR-1X322RU-A,

UCS-MR-1X162RU-A,

UCS-MR-1X081RU-A

Cisco UCS B260 M4 (UCSB-B260-M4) and


CPU: UCS-CPU-E78890D,

UCS-CPU-E78880D,

UCS-CPU-E78870D,

UCS-CPU-E78860D,

UCS-CPU-E74850D,

UCS-CPU-E74830D,

UCS-CPU-E74820D,

UCS-CPU-E74809D,

UCS-CPU-E78891D,

UCS-CPU-E78893D,

UCS-CPU-E78880LD,

UCS-CPU-E78867D

•

Memory: UCS-MR-2X082RY-E,

UCS-MR-2X162RY-E,

UCS-ML-2X324RY-E,

UCS-ML-2X648RY-E

Cisco UCS C460 M4 (UCSC-C460-M4)

•

•


UCS-CPU-E74820D,

UCS-CPU-E74830D,

UCS-CPU-E74850D,

UCS-CPU-E78860D,

UCS-CPU-E78867D,

UCS-CPU-E78870D,

UCS-CPU-E78880D,

UCS-CPU-E78880LD,

UCS-CPU-E78890D,

UCS-CPU-E78891D,

UCS-CPU-E78893D

Memory: UCSC-MRBD2-12,

UCS-MR-1X081RU-G,

UCS-MR-1X162RU-G,

UCS-MR-1X322RU-G

Storage Controller

•

•

UCSC-MRAIDC460

UCSC-MRAID12G

GPU

•

UCSC-GPU-K80



OL-31127-01


2.2(3k)

2.2(3j)

2.2(3h)

2.2(3g)

2.2(3f)

2.2(3e)

2.2(3d)

—

Table 5

UCS Release

2.2(4c)

2.2(4b)

2.2(3c)

2.2(3b)


Catalog File


1


Additional Cisco UCS B200 M4

•

UCS-MR-1X648RU-A

Additional Cisco UCS B200 M4, C220 M4, and

C240 M4 Memory

•

UCS-MR-1X322RU-A


Additional Cisco UCS C220 M4, and C240 M4

Memory

•

UCS-MR-1X648RU-A

Adapters

•

UCSC-MLOM-C10T-02

•

•

UCSC-PCIE-C10T-02

UCSC-PCIE-Q8362

Drives

—

— — ucs-catalog.2.2.3d.T.bin — ucs-catalog.2.2.3c.T.bin

—

•

•

UCS-HD18TB10KS4K

UCS-HD600G10KS4K

Fabric Extender

•

—

N2K-C2232TM-E-10GE

—

—


1

—

—

—

—

—

—

Additional Cisco UCS B200 M4 CPUs

UCS-CPU-E52640D, UCS-CPU-E52630D,

UCS-CPU-E52630LD,



UCS-CPU-E52650LD

—

—

—

1


OL-31127-01


Table 5

UCS Release

2.2(3a)

2.2(2e)

2.2(2d)


Catalog File

ucs-catalog.2.2.3a.T.bin

1



•


UCS-CPU-E52698D,

UCS-CPU-E52697D,

UCS-CPU-E52695D,

UCS-CPU-E52690D,

UCS-CPU-E52683D,

UCS-CPU-E52680D,

UCS-CPU-E52670D,

UCS-CPU-E52667D,

UCS-CPU-E52660D,

UCS-CPU-E52658D,

UCS-CPU-E52650D,

UCS-CPU-E52643D

•


UCS-MR-1X162RU-A,

UCS-ML-1X324RU-A

—

—

Fusion IO Memory

•

•

UCSB-F-FIO-1300MP

UCSB-F-FIO-1600MS

Cisco UCS C220 M4 (UCSC-C220-M4) and

Cisco UCS C240 M4 (UCSC-C240-M4)

•

CPU: UCS-CPU-E5-2620,

UCS-CPU-E5-2630,

UCS-CPU-E5-2630L,

UCS-CPU-E5-2640,

UCS-CPU-E5-2643,

UCS-CPU-E5-2650,

UCS-CPU-E5-2650L,

UCS-CPU-E5-2660,

UCS-CPU-E5-2665,

UCS-CPU-E5-2670,

UCS-CPU-E5-2680,

UCS-CPU-E5-2690

•


UCS-MR-1X162RU-A,

UCS-ML-1X324RU-A

—

—



OL-31127-01


Table 5

UCS Release

2.2(2c)


Catalog File

ucs-catalog.2.2.2c.T.bin

1



•

CPU: UCS-CPU-E52470B,

UCS-CPU-E52450B,

UCS-CPU-E52440B,

UCS-CPU-E52430LB,

UCS-CPU-E52403B,

UCS-CPU-E52420B,

UCS-CPU-E52407B

•

Memory: UCS-ML-1X324RY-A,

UCS-MR-1X162RY-A,

UCS-MR-1X082RY-A,

UCS-MR-1X041RY-A


2


OL-31127-01


Table 5

UCS Release

2.2(2c)

(cont’d))



Cisco UCS B260-M4 (UCSB-EX-M4-1)

•


UCS-CPU-E78891B,

UCS-CPU-E78880LB,

UCS-CPU-E78857B,

UCS-CPU-E74890B,

UCS-CPU-E74880B,

UCS-CPU-E74870B,

UCS-CPU-E74860B,

UCS-CPU-E74850B,

UCS-CPU-E74830B,

UCS-CPU-E74820B,

UCS-CPU-E74809B,

UCS-CPU-E72890B,

UCS-CPU-E72880B,

UCS-CPU-E72870B,

UCS-CPU-E772850B

•

Memory: UCS-ML-2X324RY-E,

UCS-MR-2X162RY-E,

UCS-MR-2X082RY-E

Cisco UCS B420-M3 (UCSB-B420-M3)

•


UCS-CPU-E54627B,

UCS-CPU-E54610B,

UCS-CPU-E54650B,

UCS-CPU-E54607B,

UCS-CPU-E54620B,

UCS-CPU-E54640B,

UCS-CPU-E54657LB

•

Memory: UCS-ML-1X324RZ-A,

UCS-MR-1X162RZ-A,

UCS-MR-1X082RZ-A

Cisco UCS B460-M4 (UCSB-EX-M4-1)

•

CPU:UCS-CPU-E74809B,

UCS-CPU-E74820B,

UCS-CPU-E74830B,

UCS-CPU-E74850B,

UCS-CPU-E74860B,

UCS-CPU-E74870B,

UCS-CPU-E74880B,

UCS-CPU-E74890B,

UCS-CPU-E78857B,

UCS-CPU-E78880LB,

UCS-CPU-E78891B,

UCS-CPU-E78893B

•

Memory: UCS-ML-2X324RY-E,

UCS-MR-2X082RY-E,

UCS-MR-2X162RY-E



OL-31127-01

New Hardware Features in Release 2.2

Table 5 Version Mapping (continued)

UCS Release

2.2(1h)

2.2(1g)

2.2(1f)

2.2(1e)

2.2(1d)

Catalog File

—

—

— — ucs-catalog.2.2.1d.T.bin — ucs-catalog.2.2.1d.T.bin

1

—


—

—

—

—

2.2(1c)

2.2(1b) ucs-catalog.2.2.1d.T.bin

1

UCS-HD12T10KS2-E

UCS-ML-1X324RY-A

UCS-MR-2X041RY-B

UCS-MR-2X082RY-B ucs-catalog.2.2.1c.T.bin

1

UCS-MR-1X041RY-A

UCS-MR-1X082RY-A

UCS-MR-1X082RZ-A

UCS-MR-2X041RX-C

UCS-MR-2X082RX-C

UCS-MR-2X162RX-C


ucs-catalog.2.2.1b.T.bin

UCS-ML-1X324RZ-A

UCS-SD200G0KS2-EP

UCS-SD400G0KS2-EP

UCS-SD800G0KS2-EP

UCSB-5108-AC2

UCSB-5108-DC2

1.

Available for separate download.

Further details are in the Cisco UCS Manager Configuration Guides .



Release 2.2(7b) adds support for the following:

•

Cisco UCS B200 M4, C220 M4, and C240 M4 shipping with the Intel® Xeon® Processor E5-2600 v4 series CPUs on

Cisco UCS 6100 and 6200 Series fabric interconnects

•

•

•

•

NVIDIA Tesla M6 GPU accelerator for B200 M4 servers

NVIDIA Tesla M60 GPU accelerator for C-Series Servers

UCSB-LSTOR-PT storage controller

SX350 Fusion IO adapters:

–

–

–

–

UCSC-F-S32002

UCSC-F-S13002

UCSC-F-S16002

UCSC-F-S64002


2 OL-31127-01


•

•

•

•

•

•

•

•

•

UCSC-SAS12GHBA PCIe adapter card

UCS-RAID9286CV-8E RAID controller

UCSC-C240-M4SNEBS

PCIe SSD on B200 M4 servers

Trusted Platform Module (TPM) 2.0

QLogic QLE8442 10Gb Dual port 10GBaseT network adapter

QLogic QLE8442 10Gb Dual port SFP+ network adapter

Cisco UCS VIC 1385 and VIC 1387 network adapters

Emulex OCE14102 B network adapter

Release 2.2(5a) adds support for the following:

•

Cisco UCS B420 M4 servers shipping with Intel E5-4600 v3 series CPUs

•

•

•

•

Cisco UCS B260 M4 and B460 M4 servers shipping with Intel E7-4800 v3 series or E7-8800 v3 series CPUs

Cisco UCS C460 M4 servers shipping with Intel E7-4800 v3 series and E7-8800 v3 series CPUs

Cisco 12G SAS Modular RAID Controller (12-port)

Cisco 12G SAS Modular RAID Controller support for Cisco UCS C460 M4 servers

•

NVIDIA Tesla K80 GPU accelerator for C-Series servers


•

•

•

•

•

Cisco UCS VIC 1227T adapter modular LOM (mLOM) for Cisco UCS C220 M4 and C240 M4 servers

Cisco UCS VIC 1225T adapter for C-Series servers

1

Cisco UCS VIC 1340 and VIC 1380 adapters for Cisco UCS B420 M3 server

2

Cisco UCSC-PCIE-Q8362 adapter for all M3 and M4 C-Series servers

Cisco Nexus 2232TM-E 10GE Fabric Extender with support for Cisco UCS VIC 1225T and VIC 1227T adapters.

•

•

•

•

•

•


•

•

Cisco UCS B200 M4 servers shipping with Intel E5-2600 v3 series CPUs

•

•

•

Cisco UCS C220 M4/C240 M4 servers shipping with Intel E5-2600 v3 Series CPUs

Cisco UCS VIC 1227 adapter modular LOM (mLOM) for Cisco UCS C220 M4 & C240M4 servers

Cisco UCS VIC 1340 and VIC 1380 adapters for Cisco UCS B200 M3, B200 M4, B260 M4, and B460 M4 servers

2

Fusion IO ioMemory for B-series and C-series servers

LSI Nytro MegaRaid NMR 8110-4i for C240 M3 server

NVIDIA Tesla K40 GPU accelerator for C-Series Servers

Emulex OCE14102 CNA

FX3S Support

10GB TwinAx cables (2m)

10GB AOC cables (1m, 2m, 3m, 5, 7m, and 10m)

1. Except for Cisco UCS C22 M3, C24 M3 and C420 M3 servers

2. Cisco UCS VIC 1340 and VIC 1380 adapters only support Cisco 6200 Series Fabric Interconnects and later models.


OL-31127-01

New Software Features in Release 2.2

•

•

•

•

•

•

•

Release 2.2(2c) adds support for the following:

B22 M3 servers shipping with Intel E5-2400 v2 CPUs.

B420 M3 servers shipping with Intel E5-4600 v2 CPUs.

B260 M4 servers shipping with two Intel E7-2800 v2 or E7-4800 v2 or E7-8800 v2 CPUs.

B460 M4 servers shipping with four Intel E7-4800 v2 or E7-8800 v2 CPUs.

C22 M3/C24 M3 servers shipping with Intel E5-2400 v2 CPUs.

C460 M4 servers shipping with two or four Intel E7-4800 v2 or E7-8800 v2 CPUs.

Chassis with updated backplane and High Voltage DC support (200v DC - 380v DC) - UCSB-5108-HVDC


•

Chassis with updated backplane UCSB-5108-AC2 or UCSB-5108-DC2



•

Firmware Upgrade Checks the VIF/Interface Status After Fabric Interconnect Reboot—During firmware upgrade, to ensure proper functioning of all services on the fabric interconnect, it is essential to ensure that port configurations and services that go down when the fabric interconnect reboots are re-established after the fabric interconnect comes back up.

Cisco UCS Manager displays any service that is not re-established after the last reboot of a fabric interconnect.

•

•

vNIC Redundancy Pair: Supports two vNICs/vHBAs that are being configured with a common set of parameters through the vNIC/vHBA template pair. This prevents the configuration between two vNICs/vHBAs from going out of sync.

Multiple vNIC/vHBA pairs can be created from the same vNIC/vHBA template pair.

Ability to perform UCS Manager initial configuration without console connectivity based on DHCP lease availability.

•

•

Preserving the following properties during backup or import operations:

–

User-defined labels for Chassis, FEX, Rack Servers, IOMs and Blade Servers

–

Assigned IDs for Chassis, FEX and Rack Servers

Locator LED support for server hard-disks—You can now identify where a specific disk is inserted in a blade or rack server using the local locator LED.

•

•

•

When you add new servers or adapters to an existing Cisco UCS system with a Cisco UCS Manager release that does not support these servers and adapters, discovery of the system may fail. In case of such a failure, the FSM will display an error message that the server or adapter is not supported on the current UCS firmware version. To resolve this issue, do one of the following:

–

Update the Capability Catalog to the latest compatible release

–

Upgrade the Cisco UCS Manager infrastructure firmware to the version required by the new hardware.

Provision to Reset Peer I/O Modules to Factory Defaults—Sometimes, I/O module upgrades can result in failures or

I/O modules can become unreachable from Cisco UCS Manager due to memory leaks. You can now reboot an I/O module that is unreachable through its peer I/O module.

Resetting the I/O module restores the I/O module to factory default settings, deletes all cache files and temporary files, but retains the size-limited OBFL file.

vNIC template CDN Source—Enables you to select the Consistent Naming Device (CDN) Source as the vNIC Name, which in turn can either be customized or derived from the vNIC instance.

2


OL-31127-01


•

•

•

•

•

•

•

•

•

NVMe PCIe SSD Inventory—Cisco UCS Manager discovers, identifies, and displays the inventory of Non-Volatile

Memory Express (NVMe) Peripheral Component Interconnect Express (PCIe) SSD storage devices. You can view the health of the storage devices in the server. NVMe with PCIe SSD storage devices reduce latency, increase input/output operations per second (IOPS), and lower power consumption compared to SAS or SATA SSDs.

PCH SSD Controller Definition—Cisco UCS Manager Platform Controller Hub (PCH) Solid State Drive (SSD)

Controller Definition provides a local storage configuration in storage profiles where you can configure all the disks in a single RAID or in a JBOD disk array.

Host Firmware Package Enhancement—You can now exclude firmware of specific components from a host firmware package either when creating a new host firmware package or when modifying an existing host firmware package. For example, if you do not want to upgrade RAID controller firmware through the host firmware package, you can exclude

RAID controller firmware from the list of firmware package components.

Local disk firmware is excluded from the host firmware package by default.

Provision to suppress VIF down alert—When a blade server that is associated with a service profile is shut down, the

VIF down alert F0283 and F0479 are automatically suppressed.

EFI Shell as a Boot Device—You can create a boot policy with an EFI Shell as the boot device. Booting from an EFI Shell prevents loss of data and provides more options to script, debug, and control various booting scenarios. EFI Shell is supported as a boot device only in the Uefi boot mode.

Multicast Hardware Hash—In a port channel, by default, ingress multicast traffic on any port in the fabric interconnect

(FI) selects a particular link between the IOM and the fabric interconnect to egress the traffic. To reduce potential issues with the bandwidth, and to provide effective load balancing of the ingress multicast traffic, hardware hashing is used for multicast traffic. When multicast hardware hashing is enabled, all links between the IOM and the fabric interconnect in a port channel can be used for multicast traffic.

Reset All Memory Errors—Provides the ability to clear all the errors from the UCS Manager.

CPLD Update of the UCSB-MRAID12G Storage Controller—Starting with Cisco UCS Manager Release 2.2(7b), the infrastructure bundle now supports a CPLD component update. The CPLD update for the UCSB-MRAID12G storage controller can be triggered by doing one of the following:

–

–

Re-acknowledging the server—If the server did not go through a deep-assoc after upgrading to the Cisco UCS

Manager 2.2(7) or later releases.

Updating the associated HFP with a supported B bundle—If the server was associated after upgrading to the Cisco

UCS Manager 2.2(7) or later releases.

VM-FEX Support—Support for ESX VM-FEX added. The support matrix in the Hardware and Software Interoperability

Matrix for this release provides more details.


•

KVM—The Virtual Media > Activate Virtual Devices is available.


•

Server Pack—This feature allows you to support new server platforms

*

on existing infrastructure without requiring a complete firmware upgrade. In this model, new B/C server bundles enabling the new servers will be supported on the previous infrastructure A bundle. For example, Release 2.2(5) B/C server bundles are supported with Release 2.2(4) infrastructure A bundle as highlighted in

Table 2

. New features introduced in the B/C bundles may only become available after upgrading the A bundle to the respective version.

The Server Pack feature provides the additional flexibility of adding new server platforms to active UCS domains without incurring the operational overhead of upgrading firmware across the whole domain.

*

The feature will apply to select server platforms.


OL-31127-01


•

•

•

•

•

•

•

•

•

RDMA over Converged Ethernet (RoCE) for Microsoft SMB Direct—RoCE is a link layer protocol, and hence, it enables communication between any two hosts in the same Ethernet broadcast domain. RoCE delivers superior performance compared to traditional network socket implementations because of lower latency, lower CPU utilization and higher utilization of network bandwidth.

RoCE for Microsoft SMB Direct is supported only on Windows 2012 R2 with Cisco UCS VIC 1340 and 1380 adapters.

LLDP Support for Fabric Interconnect vEthernet Interfaces—You can enable and disable LLDP on a vEthernet interface and retrieve information about LAN uplink neighbors. When using Cisco UCS with Cisco Application Centric

Infrastructure (ACI), LAN uplinks of the Fabric Interconnect are connected to ACI leaf nodes. Enabling LLDP on a vEthernet interface will help the Application Policy Infrastructure Controller (APIC) to identify the servers connected to the Fabric Interconnect by using vCenter.

Policy-Based Port Error Handling—If Cisco UCS Manager detects any errors on active IOM Network Interface (NIF) ports, and if the Error-disable setting is enabled, Cisco UCS Manager automatically disables the respective FI port that is connected to the IOM NIF port that had errors.

When a FI port is error disabled, it is effectively shut down and no traffic is sent or received on that port.

Advanced Local Storage Configuration:

–

Configuration of Storage Profiles and multiple virtual drives—To allow flexibility in defining the number of storage disks, roles and usage of these disks, and other storage parameters, you can create and use storage profiles. A storage profile encapsulates the storage requirements for one or more service profiles. You can also configure multiple virtual drives.

–

–

Configuration of a local LUN or a JBOD as the primary boot device

Support for local storage configuration on multiple storage controllers

–

Support for out-of-band configuration for local storage

Pass-through Capability for Gen 3 Fusion IO Mezzanine Cards—Cisco UCS Manager now supports the pass-through capability of Gen 3 Fusion IO Mezzanine cards for blade servers. Pass-through for Mezzanine cards is a hardware capability that extends the available ports for the Cisco UCS VIC 1340 or VIC 1240 adapter though the Mezzanine slot, and brings the total I/O bandwidth of the VIC 1340 or VIC 1240 adapters to dual 4 x 10 GbE.

Fabric Interconnect Traffic Evacuation—During upgrade, you can evacuate the secondary Fabric Interconnect traffic to ensure that there is no traffic flowing through the Fabric Interconnect from all servers attached to it through an IOM or

FEX. This allows you to ensure that traffic will properly failover to the other Fabric Interconnect before you can proceed with the firmware upgrade.

Per Fabric Interconnect Chassis Reacknowledgment—You can now reacknowledge a chassis per Fabric Interconnect.

This allows you to maintain connectivity to the chassis through the other Fabric Interconnect.

TPM and TXT Configuration through UCSM—The Trusted Platform Module (TPM) is a component that can securely store artifacts, such as passwords, certificates, or encryption keys, which are used to authenticate the server. A TPM can also be used to store platform measurements that help ensure that the platform remains trustworthy. Intel Trusted Execution

Technology (TXT) provides greater protection for information that is used and stored on the server.

This release supports TPM and TXT configuration on Cisco UCS M4 blade and rack-mount servers through Cisco UCS

Manager. TPM is enabled by default and TXT is disabled by default.

Consistent Device Naming Support—Consistent Device Naming (CDN) allows Ethernet interfaces to be named in a consistent manner. This makes Ethernet interface names more persistent when adapter or other configuration changes are made.

Note

CDN is supported only on Windows 2012 R2.

•

•

Fabric Scale Improvements—Each Cisco UCS domain now supports:

–

Up to 8 appliance ports

–

–

320 endpoints

Up to 320 vHBAs per Fabric Interconnect

Scriptable vMedia:


2 OL-31127-01


•

•

•

•

–

–

You can specify the authentication protocol to be used when you select CIFS as the communication protocol with the remote server.

Support for Service Profile Migration with UEFI Boot Mode—When a service profile is migrated from one server to another, the BIOS on the destination server will continue to load the boot loader information and boot in UEFI boot mode.

NVGRE with IPv6 and VMQ—You can enable NVGRE with VMQ and NVGRE with IPv6 on the same vNIC.

usNIC Support with Intel

®

MPI—Intel

®

Message Passing Interface (MPI) is developed for high performance computing.

You can now use the Intel MPI Library version 4 or 5 with Cisco user-space NIC (Cisco usNIC) for a low-latency and high-throughput communication transport.

Other Enhancements:

–

While the infrastructure firmware is being upgraded, an automatic, internal, full state backup file is created.

–

You can configure the file name of the vMedia mount image to use the name of the service profile with which the vMedia policy is associated.

–

–

When creating an SNMP trap, you can use an IPv4 address, or an IPv6 address, or a fully qualified domain name of an IPv4 address as the SNMP host name.

A new tech-support memory option is introduced to create a tech-support file with only server memory information gathered across the Cisco UCS domain.

When creating a server pool qualification policy, you can now specify the storage disk type as HDD, SSD or unspecified.


•

QPI snoop mode via Service Profile

•

•

•

•

•

•

•

•

•


Tiered Port-Licensing for direct-connect C-Series servers

Smart Call Home Enhancements

BIOS/CIMC secure boot support for C-series servers

DIMM blacklisting support for C-series servers

ENIC - DPDK Integration

Stateless Offload for Overlay Networks (NVGRE / VXLAN)

Monitoring of the IOM/FI interfaces

LAN and SAN topology information usNIC support for IP-routable transport

•

•

•

•

•

•

•

•


Scriptable vMedia

NetFlow support

PVLAN Enhancements

Pre-Upgrade Validation Checks

GPU Firmware Management

Wear-Level Monitoring on Flash Adapters

KVM/vMedia Client Enhancements

Cisco VIC Driver Enhancements:

–

Adaptive Interrupt Coalescing (AIC)

–

–

Accelerated Receive Flow Steering (ARFS) netQueue Support


OL-31127-01

Behavior Changes

•

Support for ‘ lacp suspend-individual

’ on Uplink Port-Channel

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•


IPv6 Management support

Cisco Integrated Management Controller (CIMC) In-band Management

Fabric scaling: VLAN, VIFs, IGMP, Network Adapter Endpoints

Uni-Directional Link Detection (UDLD) support

User Space NIC (usNIC) for Low Latency

Support for Virtual Machine Queue (VMQ)

C-Series Servers Direct Connect to FI without FEX

Two-factor Authentication for UCS Manager Logins

VM-FEX for Hyper-V Management with Microsoft SCVMM

Direct KVM Access

Server Firmware Auto Sync

Enhanced Local Storage Management

Flash Adapters and HDD Firmware Management

Precision Boot Order Control

Secure Boot (B-series only)

UEFI Boot Support

FlexFlash (Local SD card) support

Trusted Platform Module (TPM) Inventory

DIMM Blacklisting and Correctable Error Reporting (B-series only)

C-Series Board Controller Firmware Management

Note

If you want to refer to a list of supported OS in this release, check the Hardware and Software Interoperability Matrix for this release.


Release 2.2(7b) introduces the following behavior changes:

•

CSCtx87011—Cisco UCS backup should preserve identities by default.

–

–

Old behavior: By default, Preserve Identities is not selected in the GUI when the backup operation is initiated.

New behavior: By default, Preserve Identities is now selected in the GUI when the backup operation is initiated.

–

Impact: With Preserve Identities selected by default, allocated identities such MAC addresses, UUIDs, WWxNs and server assignments will be preserved in the backup file. This will allow the system be restored with the exact same identity assignments. To perform a backup of the logical configuration that can be re-used in other systems, customers must clear this selection when performing the backup operation.

Release 2.2(6c) introduces the following behavior changes:

•

CSCuu16791—The Firmware Auto Sync Server Policy is now set to No Actions by default.

–

Old behavior: The Firmware Auto Sync Server Policy was set to Auto Acknowledge by default, which attempted to push the default Host Firmware Package into blades/rack-servers during discovery.


2 OL-31127-01


•

–

–

Impact: The customer can now run mixed firmware environments and manage the firmware directly.

During the creation of a host firmware package, Cisco UCS Manager no longer allows checking or clearing individual components of a blade package or rack package.

–

New behavior: Withe the Firmware Auto Sync Server Policy set to No Actions by default, you not longer have to manually clear the server information from the database, and re-attempt the discovery/integration.

–

–

Old behavior: During creation of a host firmware package, if the customer selects the blade package, or rack package, or both, they can check or clear check boxes associated with any components, but they cannot modify the components of these packages. When they click Save Changes, Cisco UCS Manager displays the following error message:

Blade Package and/or Rack package is set. Individual PackImage modification is not allowed for this

Host Firmware Package.

New behavior: Starting with Cisco UCS Manager Release 2.2(6c), during creation of a host firmware package, if the customer selects the blade package, or rack package, or both, all components listed under these packages are grayed out and cannot be selected or cleared. They cannot modify the components of these packages.

Impact: The customer will now not be able to check or clear individual components listed under blade or rack packages, if the blade package, or rack package, or both, are selected during creation of a host firmware package.

Release 2.2(5a) introduces the following behavior changes:

•

CSCut47053—Support all VIC slots (1,2,4,5) for UCSM-managed single wire management of UCS C240 M4 rack-mount servers.

–

Old behavior: In UCSM mode, single wire management of UCS C240 M4 servers was only supported in Riser 1 slot 2.

–

–

New behavior: In UCSM mode, single wire management of UCS C240 M4 servers is now supported in Riser 1 (slots

1 and 2) and Riser 2 (slots 4 and 5).

Impact: The customer can now insert the VIC in Riser 1 (slots 1 and 2) and Riser 2 (slots 4 and 5) for single wire management of UCS C240 M4 rack servers.

Release 2.2(4b) introduces the following behavior changes:

•

Cisco UCS B200 M4, B260 M4, and B460 M4 blade servers no longer support QLogic (UCSB-MEZ-QLG-03) and Emulex

(UCSB-MEZ-ELX-03) mezzanine adapter cards.

–

Old behavior: Cisco UCS B200, B260, and B460 M4 servers supported QLogic (UCSB-MEZ-QLG-03) and Emulex

(UCSB-MEZ-ELX-03) mezzanine adapter cards with Cisco UCS Manager Release 2.2(3).

–

–

New behavior: Starting with Cisco UCSM Release 2.2(4b), Cisco UCS B200 M4, B260 M4, and B460 M4 blade servers no longer support QLogic (UCSB-MEZ-QLG-03) and Emulex (UCSB-MEZ-ELX-03) mezzanine adapter cards.

Impact: These servers will fail to complete the discovery process if upgraded to Cisco UCS Manager Release 2.2(4b) and later releases.

•

•

CSCus73964—Infrastructure bundle download causes environment to downgrade.

–

Old behavior: Manual upgrade or activation of Cisco UCS Manager, Fabric Interconnects, and IOM was allowed even if the startup version of the default infrastructure bundle was set.

–

–

New behavior: If the startup version of the default infrastructure bundle is set, manual upgrade or activation of Cisco

UCS Manager, Fabric Interconnects, and IOM is not allowed.

Impact: To successfully complete a manual upgrade or activation of Cisco UCS Manager, Fabric Interconnects, and

IOM, the customer has to make sure that the startup version is cleared before starting manual upgrade or activation.

This can be done through the Cisco UCS Manager CLI or the Cisco UCS Manager GUI.

CSCuq78417—Change reserved VLAN range from 4030 to 4047.

–

Old behavior: The range of reserved VLAN IDs was from 3968 to 4047.


OL-31127-01


–

–

New behavior: The range of reserved VLAN IDs is now from 4030 to 4047.

Impact: Before downgrading from Cisco UCS Manager Release 2.2(4b), the customer has to make sure that the VLAN

IDs from 3968 to 4030 are not user-configured. VLAN IDs from 3968 to 4030 are system-reserved VLAN IDs in earlier releases.

Release 2.2(3a) introduces the following behavior changes:

•

CSCup09981—Server reboot was required on FlexFlash disable in Local Disk Configuration Policy.

–

Old behavior: Server reboot was required on FlexFlash disable in Local Disk Configuration Policy.

–

New behavior: Server reboot is no longer required when FlexFlash is disabled in Local Disk Configuration Policy.

–

Impact: Customer has to make sure that the FlexFlash SD cards are not being used by the server before disabling.

Release 2.2(2c) introduces the following behavior changes:

•

CSCul74278—Adding or deleting a local storage device in boot policy will cause PNUOS boot irrespective of "Reboot on

Update" flag settings.

–

Old behavior: PNUOS boot occurs only when local device is added and SAN device is removed (or vice versa) in boot policy.

–

–

New behavior: PNUOS boot always occurs after a local storage device is added or removed in boot policy.

Impact: This change allows enabling or disabling of the boot BIOS on local storage controller even in the presence of

SAN boot devices in boot policy.

•

CSCuj81638—The system no longer returns "empty-pool" faults for empty default pools under the root org.

–

Old behavior: When a new system comes online for the first time, "empty-pool " faults get raised for all empty default pools (compute-pool-default, ip-pool-ext-mgmt, ip-pool-iscsi-initiator-pool, mac-pool-default, uuid-pool-default, wwn-pool-default, and wwn-pool-node-default) under the root org. (These fault IDs include F0463, F0464, F0465,

F0466, and F0476.)

–

New behavior: An "empty-pool" fault will no longer be raised for any empty default pool under the root org. Instead, the "empty-pool" fault will only be raised for user-generated empty pools.

–

Impact: This change allows default pools to remain empty when user has pools under a created org.

3


OL-31127-01

Security Fixes

Security Fixes

The following are security fixes in Cisco UCS Manager Release 2.2:

Table 6 Security Fixes in Release 2.2

Release

2.2(3e)A

Defect ID

CSCus69458

2.2(3d)A

2.2(3b)A

2.2.6e

CSCur29264

CSCur01379

CSCuy49897

CVE (s)

CVE-2015-0235

CVE-2014-3566

CVE-2014-7169,

CVE-2014-6271,

CVE-2014-6277,

CVE-2014-7186,

CVE-2014-7187,

CVE-2014-6278

CVE-2008-5161

Description

The heap-based buffer overflow vulnerability in the GNU C library is addressed.

The security vulnerability is addressed.

The security vulnerabilities are addressed.

Output of the commands in OpenSSH version 5.9, show this vulnerability is addressed.

Resolved Caveats

•

•

•

•

•

•

•

•

Resolved caveats are provided in the following release-specific tables:

•

•

•

•

Resolved Caveats in Release 2.2(7b)

Resolved Caveats in Release 2.2(6i)

Resolved Caveats in Release 2.2(6g)

Resolved Caveats in Release 2.2(6f)

•

•

Resolved Caveats in Release 2.2(6e)

Resolved Caveats in Release 2.2(6d)

Resolved Caveats in Release 2.2(6c)


•



Resolved Caveats in Release 2.2(5a)


•

•

•

•

•


Resolved Caveats in Release 2.2(3k)

Resolved Caveats in Release 2.2(3j)

Resolved Caveats in Release 2.2(3h)






OL-31127-01


•

•

•

•

•

•

•

•

•

•

•

•

•



Resolved Caveats in Release 2.2(3a)




Resolved Caveats in Release 2.2(1h)







The following caveats are resolved in Release 2.2 (7b)

Table 7 Resolved Caveats in Release 2.2(7b)

Defect ID Description

CSCud75506 The UUID display on ESXi 5.1 is now consistent with the UUID display on

UCSM when upgrading ESXi to version 5.1 on UCS B200 M3, B22 M3 and

B420 M3 blade servers.

First Bundle

Affected

2.0(2r)B

CSCuw36128 The buffer overflow condition that caused the statsAG mts queue to leak is now fixed.

2.2(1d)A

CSCux05389 After upgrading to release 2.2(7b) and rebooting the subordinate fabric interconnect, occasional VSAN misconfiguration will no longer occur.

CSCux66675 After rebooting a Cisco UCS 6296UP FI, all physical interfaces will no longer connect incorrectly with the Cisco UCS C460 M4 servers.

2.2(3f)A

2.2(6c)A

CSCux85580 Fabric Interconnect cores will no longer be seen on IGMP.

2.2(3b)A

CSCux65310 During blade discovery or CIMC controller reset, a chassis thermal critical fault can be generated due to the time it takes for reconnecting from the IOM to the

CIMC.

However, after several minutes this fault clears on its own.

3.1(1e)B

Resolved in Release

2.2(7b)B

2.2(7b)A

2.2(7b)A

2.2(7b)A

2.2(7b)A

2.2(7b)B

3


OL-31127-01


The following caveats are resolved in Release 2.2 (6i)

Table 8 Resolved Caveats in Release 2.2(6i)



First Bundle

Affected

2.2(1d)A

CSCuw44595 DIMMs with correctable ECC errors are marked Inoperable or Degraded even though correctable errors do not affect normal system operation. This issue is now resolved.

CSCuy01645 DIMM temperature readings are no longer missed when the temperature is 16 degrees Centigrade more than the previous reading.

CSCuj71400 Cisco UCS Manager no longer displays the “FCoE or FC uplink is down on

VSAN X” fault when the member ports for the VSAN are up.

2.2(3a)A

2.2(5b)A

2.2(1a)A

CSCuy62783 In a UCS setup with a VIC13xx adapter on blade servers or rack-mount servers, the server or VIC adapter no longer becomes unresponsive after running IO across network file systems.

2.2(3a)A

CSCuy34161 Any DIMM that is functioning correctly is no longer disabled after the reboot of the UCS B250 M2 blade server.

2.0(5c)B

CSCux58865 DIMM temperature readings are no longer missed when the temperature is 10 degrees Centigrade more than the previous reading.

CSCuu40978 The syslog is now truncated after it reaches the configured maximum size. It no longer fills up the Fabric Interconnect filesystem.

CSCux45723 UCS SNMP memory leaks no longer occur when polling FC interfaces and

SNMP processes (Walk/Get/Get Bulk) that parse any of the following SNMP

MIBs:

2.1(3d)B

2.2(3d)A

2.2(5c)A

•

•

•

•

fcIfNonLipF8Out fcIfTimeOutDiscards fcIfOutDiscards fcIfCreditLoss

•

fcIfTxWtAvgBBCreditTransitionToZero

CSCuq57142 In a port channel universe, the following no longer happen: 2.2(3a)A

•

•

A port channel ID may sometimes not be released after use. This could eventually lead to the universe of port channel IDs being empty, and no port channel IDs being available for use.

After an upgrade, power loss, FI reboot or failover, the empty port channel universe is incorrectly interpreted as a new installation, and repopulated.

This leads to duplicate port channel ID allocation when a server is attached to the FI, or when a server is re-acknowledged.

CSCuw02439 When using Cisco UCS M81KR VIC adapters on a system running Cisco UCS

Manager Release 2.2(2c), the adapters no longer crash and generate core files.

2.2(2c)B

CSCux59298 When using UCS B200 M3 servers with VIC 1240 on a system running Cisco

UCSM Release 2.2(2c), network and SAN no longer lose connectivity.

2.2(2c)B


2.2(6i)A

2.2(6i)A

2.2(6i)A

2.2(6i)A

2.2(6i)A

2.2(6i)B

2.2(6i)B

2.2(6i)A

2.2(6i)A

2.2(6i)A

2.2(6i)B

2.2(6i)B


OL-31127-01



First Bundle

Affected

CSCuy64856 The Cisco UCS fabric interconnects (FI) are no longer rebooted due to fwm hap reset.

2.1(3h)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.

2.2(6e)A

CSCuu40291 When debug logging is enabled, Cisco UCS Manager tech-support showed that syslogd_debug files were present, but the show debug logfile syslogd_debugs

CLI command failed with the following error:

3.1(1e)A

Logfile(syslogd_debugs) does not exist

This issue is now resolved.


2.2(6i)A

2.2(6i)A

2.2(6i)A

The following caveats are resolved in Release 2.2 (6g)

Table 9 Resolved Caveats in Release 2.2(6g)


First Bundle

Affected

CSCux98751 Anytime a new blade is inserted into the chassis, or a blade CIMC controller reboots, a chassis thermal fault is no longer generated because of the time taken to reconnect from the IOM to the CIMC.

2.2(6c)B

CSCuv20324 The FCoE storage no longer becomes unavailable or performs very poorly after a reset of either host-attached switch in the case of standalone systems, or after the reset of an IOM or FI in UCS-managed systems when the IOM or FI sends configuration information to the adapter later than 5 seconds after link-up.

2.2(1b)B

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.

CSCuv97713 After upgrading Cisco UCS Manager, in rare cases, the IOM may core in the sysmgr process leading to IOM reboot. This is now resolved.

2.2(3f)A

2.2(3j)A

CSCux68679 When a UCS B460 M4 server is configured with Fusion IO cards installed in same mezzanine slot of the master and slave blades, actions such as Cisco UCS

Manager upgrade, cluster failover, fabric interconnect reboot no longer trigger server reboot.

2.2(3a)A

CSCux96432 Discovery no longer fails on UCS B420 M4 servers with a 2-CPU configuration and a Fusion IO card in adapter slot 3.

2.2(5b)B

CSCux76128 Firmware Auto Install upgrade validation fails as expected when upgrading to

Cisco UCS Manager Release 2.2(6g) 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


2.2(6g)B

2.2(6g)B

2.2(6g)A

2.2(6g)A

2.2(6g)A

2.2(6g)B

2.2(6g)A

3


OL-31127-01


The following caveats are resolved in Release 2.2 (6f)

Table 10 Resolved Caveats in Release 2.2(6f)


CSCuv46749 When using Cisco B200 M4 blade servers with the UCSB-MRAID12G storage controller, the following random, incorrect transient alerts or faults are no longer reported:

•

Critical Fault [F1004] Controller Inoperable, Reason: Device reported corrupt data

First Bundle

Affected

2.2(3g)B

•

Critical Fault [F1004] Controller Inoperable, Reason: Device non-responsive

CSCux40478 When installing a Cisco B200 M3 blade server with a UCSB-MLOM-40G-01

40G VIC and either a UCSB-F-FIO-1300MP or a UCSB-F-1600MS ioMemory

PCIe flash Mezzanine card, in a chassis with an N20-I6584 IOM, blade discovery no longer fails because of an “Invalid adapter-iocard combination” error.

CSCux47667 Service profile association no longer fails if the service profile was previously associated to a different server with LUN deployed.

2.2(5b)B

2.2(6e)A


2.2(6f)B

2.2(6f)B

2.2(6f)A

The following caveats are resolved in Release 2.2 (6e)

Table 11 Resolved Caveats in Release 2.2(6e)


CSCuw44524 The server reboot issues related to clear CMOS BIOS operation in Cisco UCS

Manager Release 2.2(5a), 2.2(5b), 2.2(5c) or 2.2(6c) for E7 v2 processors on the

C460 M4, B260 M4, and B460 M4 servers are resolved.

CSCuw23829 When inserting a HDD into the final disk slot of a C240 M4 server with 8 supported slots for disks, Cisco UCS Manager no longer displays a disk in the

9th slot.

CSCux10203 When you decommission a C-Series server after it was discovered in direct attached configuration, and the direct attached switch port was changed from

Ethernet to FC mode, the following error message no longer appears:

"Warning : if_index 0x1a01a000[Ethx/x] does not exists in VLAN database#ERROR"

First Bundle

Affected

2.2(5a)B

2.2(4b)A

2.2(3h)A

CSCuw84010 Integrated C-Series rack servers with Seagate drives no longer fail association when using the host firmware policy in a service profile.

CSCuw59409 The DME crash issue that you may experience when you upgrade a directly connected C-Series rack server without decommissioning to 2.2(5c), and connect the server to FI port, is resolved.

2.2(3h)A

2.2(3h)A


2.2(6e)B

2.2(6e)A

2.2(3j)A,

2.2(6e)A

2.2(3j)A,

2.2(6e)A

2.2(3j)A,

2.2(6e)A


OL-31127-01



First Bundle

Affected

CSCuv55823 Faults no longer occur when incompatible combinations of CIMC firmware and

UCS Manager firmware are installed at the same time.

2.2(1a)A

CSCuv32417 Cisco B200 M4 and B420 M4 blade servers that run on UEFI OS will not reboot unexpectedly.

2.2(4b)B


2.2(6e)A

2.2(6e)B

Note

BIOS POST not being set as complete, which causes the next shallow discovery to reboot the server, has been resolved. Shallow discovery can occur because of different events including Cisco UCS Manager cluster failover, CIMC reset, or chassis reacknowledgment.

The following caveats are resolved in Release 2.2 (6d)

Table 12 Resolved Caveats in Release 2.2(6d)


CSCuv45173 When you upgrade C-series server firmware for C220-M4, C240-M4 to Cisco

UCS Manager 2.2(6d), you will no longer see the following critical alarm:

Board controller upgraded, manual a/c power cycle required on server x.

First Bundle

Affected

2.2(6d)


2.2(6d)C

The following caveats are resolved in Release 2.2 (6c)

Table 13 Resolved Caveats in Release 2.2(6c)


First Bundle

Affected

CSCuw03000 PXE boot fails with BSOD when RoCE is enabled.

CSCut78943 When making changes to vNICs or vHBAs that will be provisioned on Cisco

Virtual Interface Cards (VICs) 1340 and 1380 adapters, a warning on the placement order impact appears.

2.2(5a)B

2.2(3c)A

CSCuu33864 When upgrading to UCS Manager 2.2(6c) or later, the FI boots successfully without the possibility of causing a file system corruption on the SSD if the FI is equipped with a Unigen SSD.

2.2(3b)A

CSCum50468 After upgrading the UCS Manager to version 2.2(6c) or higher, false faults for

Fabric VSAN Membership Down are cleared.

2.2(1b)A

CSCut37134 A Cisco UCS B200 M4 server running ESXi 5.5 no longer crashes when you boot locally.

CSCur79257 During an FI bootup, the RACE condition no longer causes UCS Manager configure "interface vethernet 0” to switch.

2.2(3a)A

2.2(3a)A

CSCup95855 FSM tasks are no longer stuck in the throttled state in Cisco UCS Manager during Cisco UCS C240 M3 server upgrade.

CSCuv51214 Messages log are no longer overfilled with "BMC is suspecting that palo is in the boot block. Leaving I2C bus alone" messages.

2.2(1d)A

2.2(3g)B


2.2(6c)B

2.2(6c)A

2.2(3j)A,

2.2(6c)A

2.2(6c)A

2.2(6c)A

2.2(6c)A

2.2(3j)A,

2.2(6c)A

2.2(6c)B


3 OL-31127-01



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. The value got stuck at 300 regardless of the configuration. This was the issue which was resolved.

This value can be changed through UCSM GUI or CLI.

First Bundle

Affected

2.2(1c)A

CSCuu68351 The updated BIOS in UCS M3 B-Series Servers is compatible when running the

Snow Inventory client on Windows OS, and does not hang.

2.0(1m)B

CSCut88909 The KVM option, Virtual Media > Activate Virtual Devices is now available.

2.2(3b)A

CSCuu55899 When the VLAN port-count optimization is enabled, and an uplink Ethernet is in a port channel, traffic is now able to flow to the available members in the port channel when there is a link flap on one of its members.

2.2(3b)A

CSCuq63868 When creating a vNIC in a LAN connectivity policy, the show configuration command no longer generates a software error.

CSCut35123 After high-availability failover, chassis-seeprom local IO failures are no longer triggered

.

2.2(2a)A

2.2(4b)A


2.2(6c)A

2.2(6c)B

2.2(6c)A

2.2(6c)A

2.2(6c)A

2.2(6c)A

The following caveats are resolved in Release 2.2(5d):



CSCuw50417 After upgrading the BIOS of Cisco UCS B200 M3 servers to Release 2.2(5d), the OS no longer displays the following error message if the OS boot time is exceptionally slow:

Initializing Power Management ...

Power: 2568: No supported CPU power management technology detected

"Intel Enhanced SpeedStep is supported but disabled in BIOS"

First Bundle

Affected

2.2(4b)B


2.2(5d)B

CSCuv20999 When Cisco UCS Manager is managed by Cisco UCS Central, and there is a global service profile associated with a blade server, global VLANs can now be successfully assigned to a particular uplink interface through Cisco UCS

Manager (VLAN Uplink Manager > VLANs > VLAN Manager) without an error message being displayed.

CSCuw13170 When auto-deploying the installation of ESX on Ivy Bridge-EX platforms, some platforms that enforce ACS violations strictly will no longer generate a non-maskable interrupt (NMI) and crash the host OS.

2.2(4b)A

2.2(5c)B

2.2(5d)A

2.2(5d)B

OL-31127-01



The following caveats are resolved in Release 2.2(5c):



First Bundle

Affected

CSCuu89283 You can create LUNs with certain non-default virtual drive policy attributes, such as read-policy, write-policy, and cache-policy, which are managed through storage profiles for rack servers with a UCSC-MRAID12G storage controller.

2.2(4b)A

You can modify such virtual drive policy attributes for these LUNs as well without causing service profile association to fail.

CSCuv00089 After upgrading the Fabric Interconnect (FI) to Cisco UCS Manager Release

2.2(5c), systems that have OS-based NIC teaming configured as Active-Standby no longer experience network connectivity issues.

2.2(4b)A


2.2(5c)A

2.2(5c)A

Note

Network connectivity issues still occur on systems that have OS-based

NIC teaming configured as Active-Standby when you upgrade the FI to

Cisco UCS Manager Release 2.2(4).

CSCuv49345 When a server is discovered from only one Fabric Interconnect (FI), and the

DME primary instance is the other FI, disassociation of the associated service profile no longer fails.

2.2(1a)A

CSCuu60867 Cisco UCS C240 M3 servers with dual RAID controllers can now be discovered on SimpliVity solutions with OmniStack cards.

2.2(3f)A

CSCuv13545 If Cisco UCS Manager is registered with Cisco UCS Central, and the service profile refers to a global host firmware pack policy, UCSM no longer downloads images during shallow discovery.

2.2(4b)A

CSCuv28540 When you run Cisco UCS Manager Releases between 2.2(1) and 2.2(3) with catalog version 2.2(5b)T, the VIC PCI slot ID is now populated correctly.

Hence, SAN boot no longer fails on servers where the boot vHBAs are placed on a VIC.

2.2(4b)T

Note

This issue still occurs when you run UCSM Releases between 2.2(1) and

2.2(3) with catalog versions 2.2(4b)T and 2.2(5a)T.

CSCuu15250 Toggling the locator LED for a Fabric Interconnect in UCSM no longer fails.

2.2(1d)A

2.2(5c)A

2.2(5c)A

2.2(5c)A

2.2(5b)T

2.2(5c)A

3


OL-31127-01


The following caveats are resolved in Release 2.2(5b):



CSCuv04436 Cisco UCS B200 M4 servers with the following CPUs will no longer experience performance degradation:

First Bundle

Affected

2.2(4b)B

•

•

•

•

•

E5-2667 v3

E5-2643 v3

E5-2640 v3

E5-2637 v3

•

•

•

E5-2630 v3

E5-2630L v3

E5-2623 v3

E5-2620 v3

•

E5-2609 v3

CSCuv29668 When using Cisco UCS 6100 Series Fabric Interconnects with Cisco UCS 2100

IOMs, blade servers with two Cisco UCS M81KR VIC adapters no longer fail discovery after updating Cisco UCS Manager to Release 2.2.(5b) and later releases.

2.2(4b)A

CSCus11782 After rebooting a Cisco UCS 6248UP Fabric Interconnect (FI) that is operating in the FC end-host mode, all member links of the SAN port channel come up.

2.2(1d)A

CSCuu35687 In Cisco UCS B420 M4 servers with UCSB-MRAID12G and

UCSB-LSTOR-PT RAID controllers, the Fault/Locator LEDs for disks 3 and 4 are no longer swapped.

2.2(5a)B

CSCuu58282 In Cisco UCS B420 M4 servers with UCSB-MRAID12G and

UCSB-LSTOR-PT RAID controllers, in the rare event that Online Controller

Reset (OCR) is triggered during normal I/Os on JBOD drives, excessive Fast

Path IO failures are no longer seen after the controller is reset.

2.2(5a)B


2.2(5b)B

2.2(4c)B

2.2(5b)A

2.2(3j)A,

2.2(5b)A

2.2(5b)B

2.2(5b)B

The following caveats are resolved in Release 2.2(5a):

Table 17 Resolved Caveats in Release 2.2(5a)


CSCuu53920 In Cisco UCS B260 M4 and B 460 M4 servers, LUN creation and service profile association no longer fails.

First Bundle

Affected

2.2(4b)A

CSCuu37369 Storage profiles can now be created within a sub-organization in Cisco UCS

Manager.

CSCuu52001 In Cisco UCS C240 M3 rack mount servers with external LSI RAID controllers, service profile association no longer fails.

2.2(4b)A

2.2(4b)A


2.2(5a)A

2.2(5a)A

2.2(5a)A


OL-31127-01



First Bundle

Affected

CSCut54652 During service profile association, firmware upgrade on third-party Converged

Network Adapters (CNA) no longer fails with the following error:

2.2(4b)A


2.2(5a)A

Unable to find VNIC Device

CSCuu42945 Service profiles created from service profile templates with local storage profiles no longer fail association with server pools.

CSCuu65128 Continuous reboots of Cisco UCS 6200 series Fabric Interconnect are no longer observed after configuring more than 64 vNICs (excluding dynamic vNICs) on a single server.

2.2(4b)A

2.2(4b)A

2.2(5a)A

2.2(5a)A




First Bundle

Affected

CSCuv04436 Cisco UCS B200 M4 servers with the following CPUs will no longer experience performance degradation:

2.2(4b)B

•

•

•

•

•

E5-2667 v3

E5-2643 v3

E5-2640 v3

E5-2637 v3

•

•

•

•

E5-2630 v3

E5-2630L v3

E5-2623 v3

E5-2620 v3

E5-2609 v3


2.2(4c)B




First Bundle

Affected

CSCus81832 Using a Cisco UCS VIC 1340 or VIC 1380 adapter with a Sandy Bridge CPU in a Cisco UCS B420 or UCS B200 M3 server no longer causes the server to hang.

Red Hat Enterprise Linux 6.5 or 7.0 now installs successfully.

2.2(3c)B

2.1(3e)A CSCus18092 IOMs that are connected to a Fabric Interconnect no longer reboot due to a software-controlled reset.

CSCus10255 The number of licenses used by ports is now displayed correctly.

2.2(3a)A


2.2(4b)B

2.2(4b)A

2.2(4b)A

2.2(6c)


4 OL-31127-01



CSCup88161 Cisco UCS 6248 Fabric Interconnect crashes with no cores are no longer observed when new blade servers are being added to the environment.

CSCus72177 After changing the IP address of the KVM pool and Fabric Interconnect from one subnet to another, KVM launches successfully.

CSCus40519 Changing the Fibre Channel trunking mode when the Fibre Channel port-channel is active no longer impacts the Fabric Interconnect expansion module.

First Bundle

Affected

2.2 (1d)A

2.2(3d)A

2.2(3c)A

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

CSCup99955 In a scale Cisco UCS environment, the Call Home process no longer crashes and reboots the Fabric Interconnect due to a memory leak.

2.2(1b)A

CSCut04941 Cisco UCS clusters scheduled for an update from Cisco UCS Central will conform to the user acknowledgment setting.

CSCur79565 On a Cisco UCS B-Series server running Cisco UCS Manager Release 2.2(3b) or later versions, Netflow data is now sent to the collector.

CSCut03052 IP QoS core may happen in a VM-FEX scale set up, when you reboot Fabric

Interconnects back to back.

2.2(1b)A

2.2(3b)A

2.2(1b)A


2.2(4b)A

2.2(4b)A

2.2(4b)A

2.2(4b)A

2.2(4b)A

2.2(4b)A

2.2(4b)A

2.2(4b)A

The following caveats are resolved in Release 2.2(3k)

Table 20 Resolved Caveats in Release 2.2(3k)


CSCuv03557 During Cisco UCS Manager upgrade, FI activation no longer fails with the following error:

Pre-Upgrade check failed. Insufficient free space in /var/tmp.

Less than required 90%.

First Bundle

Affected

2.1(3a)A

CSCuv46749 When using Cisco B200 M4 blade servers with the UCSB-MRAID12G storage controller, the following random, incorrect transient alerts or faults are no longer reported:

•

Critical Fault [F1004] Controller Inoperable, Reason: Device reported

•

Critical Fault [F1004] Controller Inoperable, Reason: Device non-responsive

2.2(3g)B


2.2(3k)A

2.2(3k)B

CSCuv53399

After IOM firmware upgrade, or after an IOM reset while running effected firmware, Cisco UCS Manager will no longer show Major fault F0481 indicating the IOM encountered a POST failure.

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.

CSCuv97713 After upgrading Cisco UCS Manager, in rare cases, the IOM may core in the sysmgr process leading to IOM reboot. This is now resolved.

2.2(3a)A

2.2(3f)A

2.2(3j)A

2.2(3k)A

2.2(3k)A

2.2(3k)A


OL-31127-01



CSCuj71400 Cisco UCS Manager no longer displays the ’FCoE or FC uplink is down on

VSAN X' fault when the member ports for the VSAN are up.

First Bundle

Affected

2.2(1a)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. The value got stuck at 300 regardless of the configuration. This issue in now resolved. This value can be changed through UCSM GUI or CLI.

2.2(1c)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:

2.2(1b)A


2.2(1d)A


2.2(3a)A

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

CSCus59926 You can see the following issues with Call Home messages, although Smart Call

Home registers successfully:

•

Call Home messages are not generated.

2.2(3a)A

•

Test inventory is not sent.



2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

CSCus76125 Global service profile now resolves the LAN ping group even when the ping group is created in Cisco UCS Manager.

2.2(3d)A

CSCuv20324 The FCoE storage no longer becomes unavailable or performs very poorly after a reset of either host-attached switch in the case of standalone systems, or after

2.2(1b)B the reset of an IOM or FI in UCS-managed systems when the IOM or FI sends configuration information to the adapter later than 5 seconds after link-up.

CSCux68679 When a UCS B460 M4 server is configured with Fusion IO cards installed in same mezzanine slot of the master and slave blades, actions such as Cisco UCS

Manager upgrade, cluster failover, fabric interconnect reboot no longer trigger server reboot.

CSCuq57142 In a port channel universe, the following no longer happen:

2.2(3a)A

2.2(3a)A

•

•


After an upgrade, power loss, FI reboot or failover, the empty port channel universe is incorrectly interpreted as a new installation, and repopulated.

This leads to duplicate port channel ID allocation when a server is attached to the FI, or when a server is re-acknowledged.

2.2(3k)A

2.2(3k)A


4 OL-31127-01



CSCux45723 UCS SNMP memory leaks no longer occur when polling FC interfaces and

SNMP processes (Walk/Get/Get Bulk) that parse any of the following SNMP

MIBs:

•

fcIfNonLipF8Out

•

•

•

•

fcIfTimeOutDiscards fcIfOutDiscards fcIfCreditLoss fcIfTxWtAvgBBCreditTransitionToZero

First Bundle

Affected

2.2(5c)A


2.2(3k)A

CSCux59298 When using UCS B200 M3 servers with VIC 1240 on a system running Cisco

UCSM Release 2.2(2c), network and SAN no longer lose connectivity.

CSCuy01645 DIMM temperature readings are no longer missed when the temperature is 16 degrees Centigrade more than the previous reading.

2.2(2c)B

2.2(5b)A

CSCuy64856 The Cisco UCS fabric interconnects (FI) are no longer rebooted due to fwm hap reset.

2.1(3h)A

CSCuu40291 When debug logging is enabled, Cisco UCS Manager tech-support showed that syslogd_debug files were present, but the show debug logfile syslogd_debugs

CLI command failed with the following error:

Logfile(syslogd_debugs) does not exist

3.1(1e)A


CSCuu40978 The syslog is now truncated after it reaches the configured maximum size. It no longer fills up the Fabric Interconnect filesystem.

CSCuw02439 When using Cisco UCS M81KR VIC adapters on a system running Cisco UCS

Manager Release 2.2(2c), the adapters no longer crash and generate core files.

2.2(3d)A

2.2(2c)B


CSCux58865 DIMM temperature readings are no longer missed when the temperature is 10 degrees Centigrade more than the previous reading.

2.2(3a)A

2.1(3d)B

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.

CSCuy34161 Any DIMM that is functioning correctly is no longer disabled after the reboot of the UCS B250 M2 blade server.

2.2(6e)A

2.0(5c)B

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

2.2(3k)A

OL-31127-01



The following caveats are resolved in Release 2.2(3j)

Table 21 Resolved Caveats in Release 2.2(3j)


First Bundle

Affected

CSCus55944 While updating a service profile template, the service profile no longer reboots without a warning when you create and add a VLAN to a rack server that is associated with an inherited service profile.

2.2(3a)A

CSCuw78688 The failure to load configuration that caused system outage when upgrading

Cisco UCS Manager to 2.2(5c) is resolved.

2.2(3a)A

CSCus11782 After rebooting a Cisco UCS 6248UP Fabric Interconnect (FI) that is operating in the FC end-host mode, all member links of the SAN port channel come up.

2.2(1d)A

CSCus85186 After activating Cisco Trusted Platform Module (TPM), the enable and active statuses now remain as disabled and deactivated.

2.2(1d)B


2.2(3j)A

2.2(3j)A

2.2(3j)A,

2.2(5b)A

2.2(1h)B,

2.2(3j)B

CSCuv06504 The svc_sam_dme process is no longer crashing during a steady state of operation when a database (db) corruption is detected primarily on the subordinate fabric interconnect.

CSCuv72975 When upgrading the Cisco UCS Manager infrastructure bundle and using

VIC-1340 or VIC-1380, the backplane port of the IOM and VIC is no longer down.

CSCuw59409 The DME crash issue that you may experience when you upgrade a directly connected C-Series rack server without decommissioning, and connect the server to FI port, is resolved.

CSCuw84010 Integrated C-Series rack servers with Seagate drives no longer fail association when using the host firmware policy in a service profile.

2.2(3d)A 2.2(3j)A,

2.2(3a)B

2.2(3h)A

2.2(3h)A

CSCux10203 When you decommission a C-Series server after it was discovered in direct attached configuration, and the direct attached switch port was changed from

Ethernet to FC mode, the following error message no longer appears:


CSCup95855 FSM tasks are no longer stuck in the throttled state in Cisco UCS Manager during Cisco UCS C240 M3 server upgrade.

CSCur01185 The HA policy of Reset is no longer triggering the Cisco UCS 6296UP fabric interconnect to reset.

CSCus32933 Cisco UCS Manager now displays an error message when a

WILL_BOOT_FAULT event is raised because of an incorrect CPLD version.

2.2(3h)A

2.2(1d)A

2.2(1d)A

2.2(2c)B

CSCus56140 Cisco UCS Manager now performs a successful failover if the primary out-of-band management port is detected as down for more than the

Management Interface Monitoring default 5 minutes.

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.

1.4(2b)A

2.1(3e)A

2.2(6c)A

2.2(3j)B

2.2(3j)A,

2.2(6e)A

2.2(3j)A,

2.2(6e)A

2.2(3j)A,

2.2(6e)A

2.2(3j)A,

2.2(6c)A

2.2(3j)A

2.2(3j)B

2.2(3j)A

2.2(3j)A

4


OL-31127-01



CSCus93431 Adding other storage, such as a local disk or a USB to a SAN boot policy no longer deletes the SAN boot policy from Cisco UCS Manager.

First Bundle

Affected

2.2(3c)A

CSCut10525 The FlexFlash firmware version is now updating and the

FFCH_ERROR_OLD_FIRMWARE_RUNNING is not longer displaying in the fault summary after you update the UCS B200 M4 server firmware using

2.2(4b).

CSCut28278 After upgrading the infrastructure and the Cisco UCS Manager image, the subordinate fabric interconnect no longer fails the pre-upgrade check because of insufficient free space in the /var/sysmgr directory.

2.2(3a)B

2.2(1d)A

CSCuv31912 UCS Manager iptables are no longer duplicating rules in the FORWARD table. 2.2(3d)A


2.2(3j)A,

2.2(6c)A

2.2(3j)B

2.2(3j)A

2.2(3j)A,

2.2(6c)A

2.2(3j)A,

2.2(6f)A

CSCux71937 The CPU utilization always displayed 100 percent for the kernel in the output of the show system resources command. This could happen if the system had been up for a very long time (more than 200 days, but this time-frame could vary):

2.2(3b)A

FI(nx-os)# show system resources

Load average: 1 minute: 0.50 5 minutes: 0.71 15 minutes: 1.04

Processes : 563 total, 3 running

CPU states : 0.0% user, 100.0% kernel, 0.0% idle

Memory usage: 3490164K total, 3140304K used, 349860K free

This has now been resolved. The output of this command now indicates the correct CPU utilization levels.

CSCuu33864 When upgrading to UCS Manager 2.2(3j) or later, the fabric interconnect boots successfully without the possibility of causing a file system corruption on the

SSD if the fabric interconnect is equipped with a Unigen SSD.

2.2(3b)A 2.2(3j)A,

2.2(6c)A

The following caveats are resolved in Release 2.2(3h):

Table 22 Resolved Caveats in Release 2.2(3h)


CSCus64439 Cisco UCS Manager Mezz logs and VMware vmkernel logs no longer indicate storage latency and numerous FNIC aborts.

CSCuu15465 If the version of board controller on a UCS B200 M4 server is higher than the version in the Host Firmware Pack, Cisco UCS Manager no longer tries to downgrade firmware of board controllers on the servers.

First Bundle

Affected

2.2(1d)B

2.2(3b)A

While running Cisco UCS Manager Releases 2.2(3b) to 2.2(3g), you can use catalog version 2.2(4b)T or later versions to get this fix.


2.2(3h)B

2.2(3h)A


OL-31127-01


The following caveats are resolved in Release 2.2(3g):



First Bundle

Affected

CSCus74206 After disconnecting and re-connecting a cable between the FI and IOM, the VFC interfaces on the chassis will not flap and cause any disruption in the FC traffic.

2.2(1d)A

CSCur66094 Servers will not be stuck in discovery state. The locked CIMC connection issue causing bladeAG network thread lock up is resolved.

2.2(2c)A

CSCur96296 FEX will not be online and offline in quick succession under scaled setups.

CSCus91342 You will not be prompted to reboot the blade server when adding a vLAN to a vNIC template.

2.2(3a)A

2.2(1b)A

2.2(3e)A CSCut28626 When you import a backup file with multiple vLANs, the B200 M2 server association will not be stuck at 86%.

CSCut45598 6296 FI will not reboot with kernel panic.

CSCut54264 Server connectivity issue relative to non-reuse of connection slots is resolved.

2.2(3c)A

2.2(1c)B


2.2(3g)A

2.2(3g)A

2.2(3g)A

2.2(3g)A

2.2(3g)A

2.2(3g)A

2.2(3g)B

Note

The logs will show MCSERVER_LIMIT_REACHEDIN whereas netstat will show that there are only a few connections.

CSCur19358 The firmware update on K2 NVIDIA adapter will not fail with the error message

“FI unreachable”.

2.2(2e)

CSCut13146 Chassis Discovery and service profile association no longer fail during downgrade to a UCSM version 2.2(2x) or earlier, when "isolated Vlans" are allowed on appliance ports or vNIC service profiles.

The Cisco B200M4 blade server will not reboot unexpectedly. CSCut61527

&

CSCus42584

Note

Storage Controller FRU information mismatch that causes the next shallow discovery to be converted to a deep discovery has been resolved. Shallow discovery can occur through different activities such as replacing IOM cables.

2.2(1g)A

2.2(3a) &

2.2(1b)A

CSCus89837 The excessive logging of the error “error:Invalid argument” in chassis pwrmgr tech support log will not happen anymore.

2.1(3b)A

CSCut45721 FI running BladeAG will no longer need rebooting due to low kernel memory 2.1(1a)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.

2.2(3b)B

CSCut02769

dbsync

between Cisco UCS Fabric Interconnects will share only the required files and will not fail when copying the files from peer FI /opt/db/flash.

CSCur21496 After a successful firmware update, the NVIDIA GPU cards no longer display the incorrect version number.

2.2(3b)A

2.2(2e)A

2.2(3g)A

2.2(3g)A

2.2(3g)A

2.2(3g)A

2.2(3g)A

2.2(3g)A

2.2(3g)A

2.2(3g)A

4


OL-31127-01



CSCur83866 All overlapping IP addresses are appropriately removed during deletion and deployed at creation.

First Bundle

Affected

2.2(1b)A

CSCut69252 Service profile disassociation no longer fails with the flexflash scrub policy with no SD cards on C220 M4 servers.

2.2(3c)A


2.2(3g)A

2.2(3g)A

The following caveats are resolved in Release 2.2(3f):



CSCus61659 After an IOM is reset, priority flow control on the IOM interface to a blade no longer becomes disabled.

First Bundle

Affected

2.1(3c)A

CSCuq95926 Fabric Interconnect failover status is no longer stuck in Switchover In Progress when the management cable is unplugged.

2.2(2c)A

CSCut11027 When using a maintenance policy with user acknowledgment, a server with a third-party adapter will not reboot without asking for user acknowledgment after a SAN boot change in the boot policy.

2.2(3c)A

CSCuq51890 Entity physical description returns correct getnext values during SNMPwalk on

Entity MIB.

2.2(2c)A


2.2(3f)A

2.2(3f)A

2.2(3f)A

2.2(3f)A

The following caveats are resolved in Release 2.2(3e):



CSCus69458 The heap-based buffer overflow vulnerability in the GNU C library, documented in Common Vulnerability and Exposures (CVE) CVE-2015-0235 is addressed.

First Bundle

Affected

1.0(1e)

CSCur70034 Service Profiles with a common VLAN name configured on one of the vNICs no longer fail with the following error: “Incorrect VLAN configuration on one of the VNICs".

CSCuh78503 iSCSI reboot loop no longer fails with Oprom initialize error 1.

2.2(3c)A

2.1(2a)B

2.2(2c)A CSCur83235 KVM login no longer fails in UCS Manager Release 2.2(3e) over Cisco

AnyConnect VPN.

CSCus55072 When C460-M4 servers are unassociated, firmware upgrade FSM no longer restarts after the BIOS POST FSM stage.

2.2(3a)A

CSCus51733 Toshiba HDD firmware version 5706 is now packaged in UCS Manager.

Downgrades to version 5705 can change the status of the Toshiba HDD to

Unconfigured Bad and should be avoided.

2.2(3a)B

CSCur99740 Fabric Interconnect image upgrade no longer fails when you run the

reload all command

from the Fabric Interconnect.

2.2(3d)A


2.2(3e)A

2.2(3e)A

2.2(3e)B

2.2(3e)A

2.2(3e)A

2.2(3e)B

2.2(3e)A


OL-31127-01





CSCur32075 The following new microcode for the Haswell-EP processor was added to release 2.2(3d):

First Bundle

Affected

2.2(3a)B

•

M6F306F2_0000002B

CSCur42086 Blade AG no longer cores constantly and prevents assigning a service profile.

2.2(1d)A

CSCur77716 Call Home Config FSM

error no longer occurs when upgrading from 2.2(1f).

2.2(3c)A

CSCuq17289 ‘Failed to allocate exchange’ messages are no longer observed when vNICS with different COS values are used.

2.2(3c)A

CSCuq74472 Unnecessary thermal events on IOM stating that the thermal sensor reading is not available no longer occur.

CSCur05013 Messages.log file on a Cisco UCS B200 M2 blade no longer fills with messages including ‘364:BMC lost control of the bus.[0xd 0x0]’.

2.1(3c)A

2.2(2c)B

CSCur29264 The security vulnerability identified by Common Vulnerability and Exposures

(CVE) CVE-2014-3566 is addressed.

2.2(3a)A

CSCur38408 Fabric interconnect HA failover no longer causes the attached Cisco UCS C460

M4 servers to reboot when a PCIe card is present in slot 2 and riser 2 is missing.

2.2(2c)A

CSCur48661 Global service profile association no longer fails with a ‘failed to copy images’ error.

CSCun23603 UCS Manager now properly upgrades the power sequencer software.

2.2(3b)A

2.2(3c)A

CSCur37260 FI upgrade or downgrade no longer fails due to lack of disk space in /mnt/pss.

2.1(3b)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


2.2(3d)B

2.2(3d)A

2.2(3d)A

2.2(3d)A

2.2(3d)A

2.2(3d)B

2.2(3d)A

2.2(3d)A

2.2(3d)A

2.2(3d)A

2.2(3d)A

2.2(3d)A




First Bundle

Affected

CSCuq35685 IGMPv2 traffic no longer causes MAC flapping on upstream switches when the

FI has multiple uplinks connecting to different switches.

2.2(1b)A

CSCuq46105 UCS Manager now uses power sequencer version 3.0 to address the rare issue of an unexpected reboot to the Cisco UCS 6248 FI.

2.2(1c)A

CSCuq63086 Cisco B200 M3 no longer fails with a ‘Controller 1 on server x/y is inoperable’ error.

2.2(1e)B


2.2(3c)A

2.2(3c)A

2.2(3c)B

4


OL-31127-01


Table 27 Resolved Caveats in Release 2.2(3c) (continued)


CSCur08722 Firmware upgrade FSM no longer fails on the Cisco UCS C240 M3 server with the following error:

First Bundle

Affected

2.2(3a)A

IVY Bridge Processor is installed on this server, cannot downgrade to unsupported CIMC version.

CSCup45930 When upgrading the Cisco UCS B230 M2 server from Cisco UCS Manager

Release 2.2(1x) to Cisco UCS Manager Release 2.2(2x) or from Cisco UCS

Manager Release 2.2(2x) to Cisco UCS Manager Release 2.2(3x), or when upgrading the Cisco UCS C460 M4 server from Cisco UCS Manager Release

2.2(2x) to Cisco UCS Manager Release 2.2(3x), association no longer fails.

2.2(1e)A


2.2(3c)A

2.2(3c)A




CSCuq63592 Blade association no longer fails on servers with Cisco UCS VIC 1340 or VIC

1380 adapters when some vHBAs manually configured in the service profile on

Host Port 1 or Host Port 2 while other vHBAs are placed systematically by setting "Desired Host Port" to ‘

ANY

.’

First Bundle

Affected

2.2(3a)A

CSCuq64931 Blades with Cisco UCS VIC 1340 or VIC 1380 adapters no longer fail to PXE boot / iSCSI boot/ SAN boot when SRIOV vNICs (PF and VF) are placed on first host port and PXE vNIC/ iSCSI vNIC / SAN boot vHBA are placed on second host port.

CSCuq70724 DIMM I2C controller is automatically reset if there are invalid I2C transactions that prevent the chassis from obtaining temperature readings from blades (which can cause fan speeds to spin at 100%).

2.2(3a)A

2.2(2d)B

CSCuq81262 Cisco UCS Manager no longer becomes unresponsive (after seven days when

VM-related data collection begins) if anonymous reporting (AR) is enabled when upgrading to Cisco UCS Manager release 2.2(3b) in an environment specifically configured with VMs that use VM-FEX on any hypervisor.

CSCuq92477 Infra bundle upgrades from Cisco UCS Manager release 2.2(3a) no longer fail if performed from Cisco UCS Central.

2.2(3a)A

2.2(3a)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


2.2(3b)A

2.2(3b)A

2.2(3b)B

2.2(3b)A

2.2(3b)A

2.2(3b)A

OL-31127-01



The following caveats are resolved in Release 2.2(3a):

Table 29 Resolved Caveats in Release 2.2(3a)


CSCuo60330 IOMs are no longer unreachable after upgrading the FI when upgrading to Cisco

UCS Manager, Release 2.2(2) or later.

First Bundle

Affected

2.2(1d)A

CSCue93518 Cisco UCS Manager now displays an alert when users are close to reaching the

VLAN port count limit.

CSCug73260 The UCS Manager DME process no longer crashes on the subordinate FI during an upgrade or power outage.

CSCui12868 Some conditions that result in the chassis displaying a thermal sensor reading error due to third-party applications accessing the PCH SMBus or CPU SMBus controller have been fixed.

2.1(1a)A

2.1(1a)A

2.0(4b)B


2.2(3a)A

2.2(3a)A

2.2(3a)A

2.2(3a)B

Note

An additional enhancement addressing this same symptom will be included in the upcoming 2.2(3b) patch.

CSCui53234 Call Home alerts now include alert customization and service profile information. 2.0(3a)A

CSCul69513 Interface (error) counters for the fabric ports on the IOM can now be monitored using the UCS Manager GUI and CLI.

CSCum48026 Local vMedia mounts in progress can now be cancelled by the user instead of waiting for a timeout.

2.1(1a)A

2.2(2c)B

CSCun00368 A fault is now raised if the FI has less memory than expected.

CSCun00720 The UCS 6296 FI no longer fails to boot after a software upgrade, reload, or power cycle.

2.2(3a)A

2.2(3a)A

2.2(3a)B

2.1(3b)A 2.2(3a)A

2.0(2m)A 2.2(3a)A

2.2(3a)A

2.2(3a)B

CSCun66874 DIMM blacklisting no longer prevents the host from accessing the DIMM bus.

CSCun84951 MCA and MCE errors are now logged when a Cisco UCS blade server reboots due to a CPU CATERR fault.

2.2(2c)A

2.2(1b)B

CSCuo09991 The primary FI and secondary FI in a HA cluster no longer fail to sync after rebooting.

CSCuo34760 vEthernet interfaces and VIFs no longer remain in down state if the primary FI in a HA cluster is rebooted and comes up as the secondary FI.

2.2(1b)A

2.1(2d)A

2.2(3a)A

2.2(3a)A

CSCuo48978 Cisco UCS Manager now prevents downgrading the BIOS to an unsupported version on Cisco UCS C22, C24, C220, and C240 M3 servers with E5-2600 or

E5-2400 v2 series CPUs.

CSCuo51708 Cisco UCS Manager downgrade no longer fails during autoinstall.

CSCuo76425 Ingress CRC errors are no longer seen over TwinA.x cables with rapid link-flaps on the 6248 and 6296 platforms

2.1(2a)C

2.2(1d)A

2.2(1c)A

CSCuo80898 When upgrading an environment from Release 2.1(1a) to newer releases, the SAN port-channels that have member ports with different max speed will not be configured and a fault is raised to indicate this.

2.1(2a)A

2.2(3a)C

2.2(3a)A

2.2(3a)A

2.2(3a)A

5


OL-31127-01


Table 29 Resolved Caveats in Release 2.2(3a) (continued)


CSCup05019 Status of FlexFlash controller no longer shows as ‘

Disconnected Partition From

Host

” and SD cards are now visible in the local disks option after inserting those cards into a blade with Flexflash enabled in local storage policy.

First Bundle

Affected

2.2(1d)A

CSCup47504 Cisco UCS VIC 1240 adapter no longer fails to complete the PXE boot process due to missing DHCP offer on a Cisco UCS 6200 FI using an ASA as the DHCP relay.

2.2(1d)B


2.2(3a)A

2.2(3a)B

2.2(3a)A CSCup61601 FI no longer crashes upon executing the show platform software fcoe_mgr info global

command when there are too many noncontiguous VLANs configured.

2.2(1d)A

CSCup61947 MAC learning no longer fails after adding 1,000 or more VLANs with a large PV count with 120 or more virtual interfaces each with 1,000 VLANs.

CSCup82677 Line rate is no longer limited to 10GB even when QoS line rate is set to 20GB or

40GB.

2.2(1d)A

2.1(3b)A

CSCup88087 Cisco UCS B200 M3 server with a single CPU no longer reports faults for a second

CPU not present in the hardware.

2.1(3a)A

2.2(3a)A

2.2(3a)A

2.2(3a)A




CSCur01379 The security vulnerabilities identified by Common Vulnerability and Exposures



CSCur16493 The UCSC-GPU-VGXK2 GPU adapter firmware has been updated to version

80.04.F5.00.03_2055.0552.01.08.

CSCup45930 When upgrading the Cisco UCS B230 M2 server from Cisco UCS Manager

Release 2.2(1x) to Cisco UCS Manager Release 2.2(2x) or from Cisco UCS

Manager Release 2.2(2x) to Cisco UCS Manager Release 2.2(3x), or when upgrading the Cisco UCS C460 M4 server from Cisco UCS Manager Release

2.2(2x) to Cisco UCS Manager Release 2.2(3x), association no longer fails.

First Bundle

Affected

2.0(1q)A

2.2(2a)B

2.2(1e)A


2.2(2e)A

2.2(2e)B

2.2(2e)A




CSCui12868 Some conditions that result in the chassis displaying a thermal sensor reading error due to third-party applications accessing the PCH SMBus or CPU SMBus controller have been fixed.

First Bundle

Affected

2.0(4b)B

CSCun25692 M71KR-Q no longer runs out of heap and causes SAN boot failure while allocating memory to critical data structure.

2.2(1b)B


2.2(2d)B

2.2(2d)B


OL-31127-01


Table 31 Resolved Caveats in Release 2.2(2d) (continued)


CSCun79973 vNIC no longer hangs when QoS policy on a vNIC is changed while a Netflow session with transmit monitoring is active on that vNIC.

First Bundle

Affected

2.2(2c)A

CSCuo34760 Veths/VIFs no longer remain in down state on a Primary FI that gets rebooted and comes back up as subordinate FI.

2.1(2d)A


2.2(2d)A

2.2(2d)A

CSCuo79500 A vNIC with a usNIC connection policy no longer falsely shows as a Dynamic vNIC after a second vNIC with a usNIC policy is added.

CSCuo98011 Service dcosag no longer crashes during Infra firmware upgrade

(ucs-k9-bundle-infra.2.2.2c.A.bin) if upgrade is initiated from Cisco UCS Central, when Cisco UCS Manager is registered to Cisco UCS Central using IPv4 address and image bundles are not present in Cisco UCS Manager.

CSCup07488 ECC sensors no longer read or report invalid “Upper Non-Recoverable” data when there are existing failed PECI transactions on a blade.

2.2(1c)A

2.2(2c)A

2.2(2c)B

2.2(2d)A

2.2(2d)A

2.2(2d)B

CSCup61601 FCOE_MGR no longer crashes when FI is configured with too many noncontiguous VLANs.

CSCup61947 MAC learning no longer fails after adding 1000 or more VLANs with a large PV count with 120 or more virtual interfaces each with 1000 VLANs.

2.2(1d)A

2.2(1d)A

CSCup82677 A Cisco UCS system with ESXi OS is no longer limited to 10Gb speed when QoS is configured with line-rate of 20Gb or 40Gb.

2.2(1d)A

2.2(2d)A

2.2(2d)A

2.2(2d)A




CSCtf73879 Cisco UCS Manager (GUI or CLI) no longer fails to report the local disk failures, faults, alarms, status, or disk errors/error codes from MegaRAID controller.

First Bundle

Affected

1.4(1)


2.2(2c)

CSCuh89441 RHEL6.3 SAN boot on M73KR-E no longer fails.

CSCuh92027 The cluster no longer gets stuck in "switchover in progress" mode when unplugging the primary FI's management port.

2.2(1b)

2.1(1a)A

2.2(2c)

2.2(2c)A

2.2(2c)A CSCuh99542 New SNMP user configured via Cisco UCS Manager now shows up in UCS NX-OS

CLI.

2.1(1e)A

CSCuj74570 B420 M3 with UCSB-MLOM-PT-01 installed no longer reboots for discovery when first IOM is upgraded to 2204XP.

2.1(2a)B 2.2(2c)B

CSCuj81245 OS kernel and adapter logs no longer report multiple FNIC aborts when user starts a heavy write operation to SAN disk.

CSCuj81385 GUI no longer returns "PSUs have failed, please check input" fault.

2.2(1b)B

CSCuj84543 IOM no longer incorrectly reports that grid redundancy is lost while all PSUs are in "Active PS" state.

CSCul03508 Kernel memory no longer runs out and causes FI to reboot when running CA AIM tool with Cisco UCS Manager.

2.1(3a)B

2.1(3a)A

2.1(1d)A

2.2(2c)B

2.2(2c)B

2.2(2c)A

2.2(2c)A


5 OL-31127-01




CSCul15865 Remote user logged in as admin with valid locales is now able to create, modify, and delete in other organizations apart from the locales downloaded so long as locales restriction is not applied to the admin profile.

CSCul44421 Error no longer encountered when accessing shared-storage during FI reboot, upgrade, or IOM reset.

First Bundle

Affected

2.2(1b)A

2.1(2d)A

CSCul74278 The server no longer fails to boot from the local disk in cases where the boot policy was initially configured only for SAN devices and is later modified to add local disk or local HDD device.

2.2(1b)A


2.2(2c)A

2.2(2c)A

2.2(2c)A

CSCul87625 FI no longer reports PSU failure when there is no real outage.

CSCul95341 Server no longer becomes unreachable after vNIC failover/failback.

2.1(3a)A

2.2(1b)A

CSCum10869 IO throttle count option on the FC adapter policy is now 256 by default and range changed to ‘256 to 1024’ to remove confusion caused by default value of 16, which was interpreted by driver to mean 2048.

2.1(1f)A

CSCum18805 Error message no longer received when pinging from peer FI's local-mgmt.

2.2(1b)

2.2(2c)A

2.2(2c)A

2.2(2c)A

2.2(2c)A

CSCum19106 HDDs are displayed appropriately now in both Cisco UCS Manager and in BIOS for B200 M3.

2.2(1b)B

CSCum19132 httpd_cimc.sh crash no longer seen when changing switch mode from End-Host to

Switch mode and vice versa.

2.2(1b)A

2.2(1b) CSCum20965 Service profile now correctly disassociates from the blade when selecting

“Disassociate Service Profile” from Cisco UCS Manager GUI.

CSCum51302 User no longer receives ' switchport trunk allowed vlan none

' configuration error on vEthernet configuration for VLAN change.

2.2(1b)A

CSCum57954 Ping requests no longer timeout when configuring a B200 M3 with Cisco UCS

VIC 1240 and Cisco UCS VIC 1280 with VM-FEX on the ESXi DVS while vMotioning a VM between two blades.

CSCum60793 Cisco UCS Manager no longer reports

‘mount suspend success’

to operations manager errors when UCS becomes unreachable or unregistered with Cisco UCS

Central.

2.2(1b)

2.1(2c)

CSCum63448 A Cisco UCS compute blade no longer requests a reboot for minor changes when a component firmware in an "Activating" state after an upgrade is completed.

CSCum82385 Allowed VLANs in vNIC now matches the allowed VLANs in vEthernet in NX-OS and vNICs no longer get updated when the service-profile is renamed.

2.1(2c)A

2.1(3a)A

CSCum84239 No longer have to be logged in as admin to be able to connect adapter.

CSCum84618 Installation of Win2008R2 no longer fails to reload after loading the disk.sys driver, preventing final installation steps to be completed.

2.2(1b)A

2.1(3a)B

CSCum87525 The httpd.sh process on the subordinate FI no longer crashes after upgrading to

2.2(2c) from 2.1(3a) via auto install.

CSCum95778 Incorrect sorting results are no longer displayed when attempting to sort a listing of service profiles and ORGs.

2.2(1b)A

2.2(1b)A

2.2(2c)B

2.2(2c)A

2.2(2c)A

2.2(2c)A

2.2(2c)A

2.2(2c)A

2.2(2c)A

2.2(2c)A

2.2(2c)A

2.2(2c)B

2.2(2c)A

2.2(2c)A


OL-31127-01




CSCum98771 User no longer receives config failure message ‘

Flexflash controller error, probably not supported, not equipped or inoperable

’ when Cisco UCS

Manager is updated from 2.1(2a) or above to 2.2(1) with FlexFlash State enabled.

First Bundle

Affected

2.2(1b)A

CSCun13327 A Cisco UCS system with a high number of blade or rack-servers no longer experiences a high number of messages collected in the bladeAG logs, causing logs to overflow and prevent user from collecting necessary data for troubleshooting purposes.

CSCun14140 UCSB-B200-M3 Kernel no longer panics or hangs.

2.1(3a)A

2.1(3b)B

2.2(1b)A CSCun19372 User is no longer prevented from adding interface from both fabrics to the VLAN when launching the LAN Uplink Manager from the Equipment > Fiber

Interconnects > Fabric Interconnect A (or B) path.

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.

2.2(1b)A


2.2(2c)A

2.2(2c)A

2.2(2c)B

2.2(2c)A

2.2(2c)A

2.2(2c)A CSCun25187 Cisco UCS Manager no longer becomes Lost-Visibility in Cisco UCS Central when keyring with certificate chain is configured for https communication if third-party certificates are signed by subordinate CA instead of root CA.

CSCun42280 Messages log no longer overfilled with "BMC is suspecting that palo is in boot block. Leaving I2C bus alone." messages.

2.1(2a)A

2.1(2a)B

CSCun43320 Linux PXE server no longer fails to boot from gPXE server.

CSCun46196 A Cisco UCS running 2.2(2c) that has a M1 or M2 blade no longer reports a minor fault related to Inband configuration when there is actually no Inband configuration intended for that blade.

2.1(2a)B

2.2(1b)

2.2(2c)B

2.2(2c)A

2.2(2c)A

2.2(2c)B CSCun94906 The Cisco UCS Manager DME no longer crashes when a SP is referring to a SP template and Cisco UCS Manager is connected to Cisco UCS Central during policy ownership conflict checks.

2.2(1b)B

CSCun95096 During association, the service profile no longer fails with an error stating that the

RAID configuration is invalid due to an attempt to mix different drive technologies simply because one drive has an incorrect capability catalog technology field (set to "HDD" when it should be set to "SSD").

2.2(1b)T

CSCuo06886 Cisco UCS Manager no longer populates the "Created at" field with a strange number in the wrong format instead of the date when exporting events in CSV format.

CSCuo09527 A blade running on a 2-socket B420 M3 no longer boots to the BIOS screen after an update if the boot order is configured in the Cisco UCS Manager and the BIOS is set to strict mode.

CSCuo12965 KVM console no longer fails to launch when user tries to launch from the

Equipment tab in Cisco UCS Manager or from the KVM Launch Manager due to erroneous ‘No management IP address set’ error message when server has no management IP addresses assigned but the associated service profile does.

2.2(1b)A

2.2(1b)A

2.2(1b)A

CSCuo23630 User no longer receives “There are uncommitted changes” prompt after applying a port license to FI and, then, trying to move to another screen.

2.2(1c)A

2.2(2c)T

2.2(2c)A

2.2(2c)A

2.2(2c)A

2.2(2c)A


5 OL-31127-01




CSCuo30572 Intel v2 processors no longer cause PSOD with Microsoft Windows 2008 R2 VM guests.

CSCuo36008 User no longer receives the ‘ peer connectivity: disconnected

’ warning after re-seating the IOM(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.

First Bundle

Affected

2.1(3a)A


2.2(2c)A

2.2(1c)A 2.2(2c)A

2.0(1m)A 2.2(2c)A

The following caveats are resolved in Release 2.2(1h):

Table 33 Resolved Caveats in Release 2.2(1h)


First Bundle

Affected

CSCut63966 Switch will not stop at loader prompt upon reboot due to incorrect boot variables or /opt corruption.

2.2(1b)A

CSCur88952 svc_sam_dme core is no longer found after upgrading from or downgrading to

Cisco UCS Manager Release 2.1(1b).

2.1(3g)A

CSCus83447 Cisco UCS Fabric Interconnect reload or switchover due to a leap second update no longer occurs.

2.2(1b)A

CSCus85186 After activating Cisco Trusted Platform Module (TPM), the enable and active statuses will not remain as disabled and deactivated.

2.2(1d)B


2.2(1h)A,

2.2(3g)A

2.2(1h)A,

2.2(3f)A

2.2(1h)A,

2.2(3e)A

2.2(1h)B,

2.2(3j)B,

2.2(4b)B

2.2(1h)A,

2.2(3g)A

CSCut03086 In a VM-FEX scale set up with over 100 servers, when you use auto install infrastructure firmware bundle and UCS Manager FI upgrades, DME core will not happen.

CSCur54705 Cisco UCS Manager will no longer send UCS Manager username and password hashes to the configured SYSLOG server every 12 hours.

2.2(2c)A

2.1(1a)A

CSCut09151,

CSCut21914 and

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.

CSCuo50049 Cisco UCS Manager will not experience HA cluster failover after upgrading from Release 1.4.

2.1(3a)B

2.0(5g)A

2.2(1h)A,

2.2(3e)A

2.2(1h)A,

2.2(3f)B

2.2(1h)A,

2.2(3d)A

OL-31127-01



The following caveats are resolved in Release 2.2(1g):



CSCuo51708 UCSM autoinstall downgrade no longer fails while reverting firmware during autoinstall process.

First Bundle

Affected

2.1(1f)A

CSCuq70724 DIMM I2C controller is automatically reset if there are invalid I2C transactions that prevent the chassis from obtaining temperature readings from blades (which can cause fan speeds to spin at 100%).

2.1(1a)B

Resolved in

Release

2.2(1g)A

2.2(1g)B

CSCuq20755 Race conditions no longer occur due to DME crashing and restarting.

CSCuq74472 Unnecessary thermal events on IOM stating that the thermal sensor reading is not available no longer occur.

2.0(1q)A

2.1(3c)A

CSCur37260 FI upgrade or downgrade no longer fails due to lack of disk space in /mnt/pss.

2.1(3b)A

2.2(1g)A

2.2(1g)A

2.2(1g)A

The following caveats are resolved in Release 2.2(1f):


First

Affected

Bundle Defect ID Description

CSCuq35685 IGMPv2 traffic no longer causes MAC flapping on upstream switches when the FI has multiple uplinks connecting to different switches.

2.2(1b)A

2.0(1q)A CSCur01379 The security vulnerabilities identified by Common Vulnerability and Exposures



CSCuo34760 vEthernet interfaces and VIFs no longer remain in down state if the primary FI in a HA cluster is rebooted and comes up as the secondary FI.

CSCuq63086 Cisco B200 M3 no longer fails with a ‘Controller 1 on server x/y is inoperable’ error.


2.2(1f)A

2.1(2d)A 2.2(1f)A

2.2(1e)B

2.2(1f)A

2.2(1f)B




CSCup61947 MAC learning no longer fails after adding 1000 or more VLANs with a large PV count with 120 or more virtual interfaces each with 1000 VLANs.

CSCul44421 Error no longer encountered when accessing shared-storage during FI reboot, upgrade, or IOM reset.

First

Affected

Bundle

2.2(1d)A

2.0(5f)B


2.2(1e)A

2.2(2d)A

2.2(1e)B

5


OL-31127-01


Table 36 Resolved Caveats in Release 2.2(1e) (continued)


CSCum98771 User no longer receives config failure message ‘

Flexflash controller error, probably not supported, not equipped or inoperable

’ when Cisco UCS

Manager is updated from 2.1(2a) or above to 2.2(1) with FlexFlash State enabled.

First

Affected

Bundle

2.2(1b)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.

CSCun25187 Cisco UCS Manager no longer becomes Lost-Visibility in Cisco UCS Central when keyring with certificate chain is configured for https communication if third-party certificates are signed by subordinate CA instead of root CA.

2.2(1b)A

2.1(2a)A

CSCun25692 M71KR-Q no longer runs out of heap and causes SAN boot failure while allocating memory to critical data structure.

CSCun83328 vNIC PCI addresses no longer get changed when downgrading from release 2.2(1e) to 2.0(5).

2.2(1b)B

2.2(1d)B

CSCun84897 Server reboots are no longer triggered when vNIC template is used with global service profile and customer changes MTU setting.

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.

2.1(3a)A

2.0(1m)A

CSCuo79500 Cisco UCS Manager no longer falsely displays information that indicates a vNIC with a usNIC connection policy was automatically changed to dynamic vNIC when a customer adds a second vNIC with a usNIC policy.

2.2(1c)A


2.2(1e)A

2.2(1e)A

2.2(1e)A

2.2(1e)B

2.2(2d)B

2.2(1e)B

2.2(1e)A

2.2(1e)A

2.2(1e)A




CSCui31011 VIFs and vNICs on the subordinate no longer fail to come up after FI failover.

CSCul38768 “Management services are unresponsive” fault no longer displayed in Cisco UCS

Manager unless there is actually an HA condition failure.

CSCul96021 Frequent crashing of dcosAG process no longer occurs after Cisco UCS Manager upgrade.

First Bundle

Affected

2.1(2a)A

2.1(1e)A

2.2(1b)A

CSCum02561 Server no longer reboots after infra-only upgrade that is followed by a service profile change.

2.2(1b)A

CSCum04646 FI no longer reboots while doing SNMP Walk with LDAP configuration enabled.

2.2(1b)A

CSCum09954 Erroneous minor faults “FCoE” or “FC uplink is down” are no longer displayed in

Cisco UCS Manager when FC uplink interfaces are admin shut.

2.2(1b)A

CSCum75266 FIs no longer go to bash prompt during upgrade to 2.2 release.

CSCum82888 Access to Cisco UCS Manager, FIs, and Virtual IP is no longer blocked after upgrade or reboot if default keyring was deleted.

2.2(1b)A

2.2(1b)A


2.2(1d)A

2.2(1d)A

2.2(1d)A

2.2(1d)A

2.2(1d)A

2.2(1d)A

2.2(1d)A

2.2(1d)A


OL-31127-01


Table 37 Resolved Caveats in Release 2.2(1d) (continued)


CSCun01514 Boot Order change via CLI or XML API no longer fails between SAN & other devices.

First Bundle

Affected

2.2(1b)A

CSCun19289 The mgmt0 interface on an FI no longer drops traffic coming from blades behind that FI.

CSCun21077 Blades running ESXi OS will no longer hit a speed cap of 10GB when running on hardware that allows for higher speeds.

2.2(1b)B

2.2(1b)A


2.2(1d)A

2.2(1d)B

2.2(1d)A

CSCun59192 Secure properties, such as authentication passwords and community strings, are no longer reset when importing a backup configuration that was created while being registered to Cisco UCS Central with admin policies managed globally.

2.2(1c)A 2.2(1d)A




CSCum43435 The Cisco UCS VIC 1280 adapter on a Cisco UCS B440 M2 blade no longer hangs during PXE boot.

First Bundle

Affected

2.2(1b)B


2.2(1c)B

2.2(1c)B CSCul44120 Cisco UCS blade servers performing a PXE boot from a Citrix PVS system no longer fail after the initial bootloader initialization with a “No API found” error.

CSCum51025 Cisco UCS Manager domains registered in Cisco UCS Central no longer fail during scheduled backups.

2.1(2a)B

2.2(1b)A 2.2(1c)A

CSCum25003 Associated local and global service profile inventory information is no longer sent to Cisco UCS Central when no changes are made.

2.2(1b)A 2.2(1c)A




CSCug93342 After rebooting the FI during an upgrade, the FI no longer boots to the Loader prompt.

First Bundle

Affected

2.1(3a)A


2.2(1b)A

Note

To ensure the new power sequencer firmware is installed properly, you must power cycle the FI after the upgrade is completed.

CSCuc19701 The 2204 IOM no longer reboots when the FI is reset.

CSCuf90470 When Call Home is enabled, Online Insertion and Removal (OIR) or failure of hardware modules in the FI no longer causes the FI to reboot.

2.1(1a)A

2.1(1b)A

2.2(1b)A

2.2(1b)A

5


OL-31127-01


Table 39 Resolved Caveats in Release 2.2(1b) (continued)


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.

CSCuj32124 During normal operation IOM no longer unexpectedly reboots with CPU reset.

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

First Bundle

Affected

2.1(2a)A

CSCuj84421 Installing Java 7 update 45 no longer causes UCS Manager GUI failures.

CSCuc38783 Firmware upgrade no longer fails on BRCM 57712 adapter.

CSCuj77813 Associating service profile to a C-series server integrated with Cisco UCS Manager no longer fails.

CSCui43905 When you try to create an iSCSI NIC on a service profile, UCS Manager no longer displays any warning messages.

2.1(1f)A

2.1(1a)A

2.1(3a)C

2.1(2a)B

1.4(2b)A

1.4(2b)A


2.2(1b)A

2.2(1b)A

2.2(1b)A

2.2(1b)A

2.2(1b)A

2.2(1b)A

2.2(1b)A

•

Errors during IOM boot-up

CSCuj78099 FC traffic between a Cisco UCS FI and a Cisco MDS switch is no longer disrupted when the FI is in switch mode.

CSCuj61839 With the 2.2(1b)B bundle, Cisco UCS Blade servers on a Cisco M81KR VIC adapter no longer encounter "ASSERT FAILED @ mips/ecpu_pani.c:138" errors.

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(3a)A 2.2(1b)A

2.1(2a)A 2.2(1b)A

2.1(1f)A 2.2(1b)A

CSCuf77316 Windows 2012 installed on a FlexFlash card no longer fails Microsoft certification. 2.1(2a)B

CSCui87195 The FLS process on Cisco CNA M72KR-E no longer cores with the following message:

2.0(5c)C

2.2(1b)A

2.2(1b)A

130820-19:06:33.645547 fls.fc vnic 15: Local port down for lif 4.

CSCuj99958 During heavy FC traffic, the server no longer stops responding with an

ASSERT

FAILED (Exception 2 triggered!) @ mips/ecpu_panic.c:138

error.

CSCug89448 The tech support collection no longer fails on the Cisco UCS Manager GUI.

CSCug63368 PXE boot no longer fails in vPC environments if the DHCP relay agent is installed as the gateway IP address during a PXE boot instead of the HSRP IP address.

2.1(1f)A 2.2(1b)A

2.0(1s)A 2.2(1b)A

2.1(1d)B 2.2(1b)A

2.1(2a)A 2.2(1b)A CSCui82679 FlexFlash storage is no longer disconnected from a Cisco B200 M3 server after booting ESX or ESXi from a FlexFlash card.

CSCuc52981 Downloaded license files for FIs continue to be displayed in the downloads area after installation.

CSCti39470 RAID 50 and RAID 60 configurations can now be created and managed by Cisco

UCS Manager.

2.0(3a)A

1.4(1i)A

2.2(1b)A

2.2(1b)A


OL-31127-01

Open Caveats

Open Caveats

Open caveats are provided in the following release-specific tables:

Open caveats may be listed in association with the release in which they were first noticed or in the release identified as the first affected. Users should review open caveats in all releases to avoid overlooking a defect that may impact their release.

•

•

•

•

•

Open Caveats in Release 2.2(7b)

Open Caveats in Release 2.2(6e)

Open Caveats in Release 2.2(6c)


•

•

•

•

Open Caveats in Release 2.2(5a)


Open Caveats in Release 2.2(3j)

Open Caveats in Release 2.2(3h)

•

•

•

•

Open Caveats in Release 2.2(3g)

Open Caveats in Release 2.2(3f)

Open Caveats in Release 2.2(3d)


•

•

•

•





•

•

•

•



Open Caveats in Release 2.2(1f)


•

•

•





6


OL-31127-01

Open Caveats

The following caveats are open in Release 2.2(7b)

Table 40 Open Caveats in Release 2.2(7b)

Defect ID Symptom

CSCuz74973 When you use a B200 M4 server with a

UCSB-MRAID12G SAS RAID controller and a

CPLD firmware version earlier than version

05D, the B200 M4 server powers off unexpectedly. The OBFL displays the following log message:

[platform_power_state_irq_handler]:18:VDD_

PWR_GOOD: Deasserted

Workaround

If this issue occurs, upgrade the infrastructure and server firmware to Cisco

UCS Manager Release 2.2(7b) or later releases.

Note

Both infrastructure and server firmware must be upgraded.

First Bundle

Affected

2.2(7b)B

If the server firmware is upgraded to a fixed release earlier than the infrastructure firmware, do one of the following after the infrastructure firmware is upgraded to the fixed release.

•

•

Re-acknowledge the server

Decommission and then acknowledge the server

This is required for the correct server firmware component to upgrade and prevent the issue.

To avoid this issue, ensure that the server has

BIOS and CIMC versions from the same C

Series bundle.

2.2(7b)C CSCuy74364 When the CIMC version in the server is earlier than the 2.2(7) C Series bundle CIMC version, or, when the BIOS version in the server is later than the 2.2(6) C Series BIOS version, the host reboot is triggered along with a BIOS upgrade even for any unrelated service profile modification.

CSCux63909 FC abort is observed while running uplink port flap test on uplink Ethernet port-channel from an

FI to a Cisco Nexus 7000 switch.

No known workaround.

2.2(7b)A

OL-31127-01


Open Caveats

The following caveats are open in Release 2.2(6e)

Table 41 Open Caveats in Release 2.2(6e)


CSCuz20650 When syslog messages are generated continuously, the syslog suspend timer does not recover. Thus, no events are sent to the remote syslog server.

Workaround

There is no known workaround.

CSCux47667 Service profile association to a server fails if the service profile was previously associated to a different server, and a LUN was deployed.

First Bundle

Affected

2.2(6e)A

Resolved in

2.2(3k)A,

2.2(6i)A

2.2(6e)A Delete the LUN configuration of the service profile before associating to a different server.

Resolved in

2.2(6f)A

The following caveats are open in Release 2.2(6c)

Table 42 Open Caveats in Release 2.2(6c)

Defect ID Symptom Workaround

First Bundle

Affected

CSCux66675 Rebooting a Cisco UCS 6296UP FI causes all physical interfaces to connect incorrectly with the Cisco UCS C460 M4 servers.

Reacknowledge the servers to resolve this issue.

CSCux98751 Anytime a new blade is inserted into the chassis, or a blade CIMC controller reboots, a chassis thermal fault may get generated because of the long time taken to reconnect from the IOM to the

CIMC. This issue is caused by added time for

BIOS Power Capping and SMBUS protection abilities.

This fault will clear by itself, and has no known workaround.

2.2(6c)A

Resolved in

2.2(7b)A

2.2(6c)B

CSCuw44524 Server reboot with 'Uncorrectable error ECC errors' may occur when you initiate a 'clear

CMOS' operation in Cisco UCS Manager

Release 2.2(5a), 2.2(5b), 2.2(5c) or 2.2(6c) for

E7 v2 processors on the C460 M4, B260 M4, and

B460 M4 servers.

CSCuv45173 When you upgrade C-series server firmware for

C220-M4, C240-M4 to Cisco UCS Manager

2.2(6c), you will see the following critical alarm:

Board controller upgraded, manual a/c power cycle required on server x.

Use the HUU or IMC interfaces to upgrade the BIOS again to any version, activate the

BIOS, then reboot the system.

2.2(6c)A

Resolved in

2.2(6e)A

The alarm seen is mistakenly categorized as a critical alarm. It does not impact the functionality of the server and can be ignored.

2.2(6d)C

CSCuw13352 If you set the Processor C6 Report to enabled in the BIOS policy, the system hangs during the OS boot process.

CSCuu78398 All blades running with the Storage Controller

Firmware that does not support 4K native drives are showing the wrong virtual drive block size.

When configuring the Processor C6 Report in the BIOS policy, accept the default value of disabled.

This issue has no known workaround.

Resolved in

2.2(6g)B

2.2(6c)A

2.2(6c)A


6 OL-31127-01

Open Caveats


CSCuv18630 Cisco UCS B420 M3 servers with the firmware for UCSB-MRAID12G installed reports a nonfunctional Toshiba hard drive as Missing in

Cisco UCS Manager.

Workaround


First Bundle

Affected

2.2(6c)A

CSCuv25760 On CPU 4830 and 4850, the P-state is not changing when running the linpack benchmark on Redhat, which results in the frequency not increasing.


2.2(6c)A

CSCuv27475 The UCS B-Series Blade Servers include a version of Unified Extensible Firmware

Interface (UEFI) that is affected by the vulnerabilities identified by the Common

Vulnerability and Exposures (CVE) IDs:

CVE-2014-8274.


2.2(6c)A

CSCuv27587 The import configuration fails when call home is not configured while performing a backup operation.

Configure call home before taking a backup. 2.2(6c)A

Note

Call home is configured in the system where UCS Manager imports the configuration.

CSCuv38136 The platform might hang if there are any errors during the boot loader phase when in UEFI boot mode and using a Fibre Channel (FC) boot.

Error conditions that require re-discovery/login for the FC LUN path during the BIOS boot phase, until the OS takes over, are currently not handled by VIC Oprom UEFI Driver. This results in a hang when the OS is being loaded by the

BIOS using a VIC UEFI driver.

CSCuv59487 SPAN does not capture broadcast packets when you add vNIC as source.

Shutdown and reboot the server.

Use the HIF port where the Virtual Network

Interface (vEth) is bound as the SPAN source.

2.2(6c)A

2.2(6c)A

CSCuv62162 The Power Technology token in UCS Manager

BIOS policy has no effect on relevant CPU power performance option, such as EIST, Turbo

Mode, P-States, and C-States.

Each setup option that is meant to be controlled by the Power Technology token should be set manually in UCS Manager.

2.2(6c)A


OL-31127-01

Open Caveats


CSCuv76884 The Adaptor policies of dynamic vNICs created for a Global Service Profile (GSP) vNIC on UCS

Manager are incorrect when there are two

Dynamic vNIC policies with the same name created in different organization levels.

Workaround

Change the Global Service Profile label to update the service profile vNICs.

Note

This might cause the server to reboot.

CSCuv88433 UCS Manager displays a different sets of I2C data based on whether the data is from the master

IOM or the slave IOM that is connected to the primary fabric.

There is no workaround for this issue.

First Bundle

Affected

2.2(6c)A

2.2(6c)A

Note

If UCS Manager primary is connected to the master IOM, the correct set of 12C data is collected.

CSCuw10623 SBR firmware updates for servers with

LSI9271-8i adapter appear to have a loss of datastore in the ESXi host.


2.2(6c)A

The following caveats are open in Release 2.2(5c)



CSCux45723 UCS SNMP memory leaks occur when polling

FC interfaces and SNMP processes

(Walk/Get/Get Bulk) that parse any of the following SNMP MIBs:

•

fcIfNonLipF8Out

•

•

fcIfTimeOutDiscards fcIfOutDiscards

Workaround

•

•

fcIfCreditLoss fcIfTxWtAvgBBCreditTransitionToZero

CSCuw13170 When auto-deploying the installation of ESX on

Ivy Bridge-EX platforms, some platforms that enforce ACS violations strictly will generate a non-maskable interrupt (NMI), and the host OS may crash.


First Bundle

Affected

If SNMP memory leaks occur, avoid polling the SNMP MIBs on FC interfaces.

2.2(5c)A

Resolved in

2.2(3k)A,

2.2(6i)A

2.2(5c)B

Resolved in

2.2(5d)B

6


OL-31127-01

Open Caveats

The following caveats are open in Release 2.2(5b)



CSCuy01645 DIMM temperature readings are reported as NA when the temperature is 16 degree Celsius more than the previous reading. This results in missed temperature readings, which in turn results in

Cisco UCS Manager generating thermal alerts.

Workaround


CSCux96432 Discovery fails on UCS B420 M4 servers with a

2-CPU configuration and a Fusion IO card in adapter slot 3.

To resolve this issue, update the BIOS to version 2.2.6d.0 or later versions.

CSCux40478 When installing a Cisco B200 M3 blade server with a UCSB-MLOM-40G-01 40G VIC and either a UCSB-F-FIO-1300MP or a

UCSB-F-1600MS ioMemory PCIe flash

Mezzanine card, in a chassis with an N20-I6584

IOM, blade discovery fails with an “Invalid adapter-iocard combination” error.

When this issue occurs, remove the flash

Mezzanine card to be able to continue installation.

First Bundle

Affected

2.2(5b)A

Resolved in

2.2(3k)A,

2.2(6i)A

2.2(5b)B

Resolved in

2.2(6g)B

2.2(5b)B

Resolved in

2.2(6f)B

OL-31127-01


Open Caveats

The following caveats are open in Release 2.2(5a)

Table 45 Open Caveats in Release 2.2(5a)


CSCuw44524 When using Cisco UCS Manager Release

2.2(5a), 2.2(5b), 2.2(5c) or 2.2(6c) for E7 v2 processors on the C460 M4, B260 M4, and B460

M4 servers, while you perform a clear CMOS

BIOS operation, the following error may occur, which eventually causes the server to reboot or crash:

"System Software event: Memory sensor,

Uncorrectable ECC error, DIMM socket 1,

Channel A, Memory Riser 2, Processor

Socket 1. was asserted"

Workaround

When this issue occurs, downgrade to Cisco

UCS Manager Release 2.2(4c).

Note

This issue will not be seen when using versions prior to Cisco UCS

Manager release 2.2(5a), since support for both E7 v2 and v3 processors is available from Cisco

UCS Manager release 2.2(5a).

First Bundle

Affected

2.2(5a)B

2.2(5a)C

Note

This issue is specific to Cisco UCS

Manager B and C server bundles.

Resolved in 2.2(6e)

CSCuv04436 Cisco UCS B200 M4 servers with the following

CPUs may experience performance degradation:

•

E5-2667 v3

There is no workaround for this issue. See the Software Advisory.

A patch will be released in July, 2015.

2.2(4b)B

•

•

•

E5-2643 v3

E5-2640 v3

E5-2637 v3

E5-2630 v3

•

•

•

E5-2630L v3

E5-2623 v3

CSCuu38206 If you disable and enable TXT several times, the state of TXT in the Cisco UCS Manager BIOS policy and the actual tboot state may get out of sync.

CSCuu58282

•

•

E5-2620 v3

E5-2609 v3

In Cisco UCS B420 M4 servers with

UCSB-MRAID12G and UCSB-LSTOR-PT

RAID controllers, in the rare event that Online

Controller Reset (OCR) is triggered during normal I/Os on JBOD drives, excessive Fast Path

IO failures may be seen after the controller is reset.

When this occurs, do the following:

•

•

Clear the CMOS

Re-apply the Cisco UCS Manager BIOS policy

When this occurs, reboot the host.

2.2(5a)B

2.2(5a)B

Resolved in

2.2(5b)B

6


OL-31127-01

Open Caveats


CSCuu35687 In Cisco UCS B420 M4 servers with

UCSB-MRAID12G and UCSB-LSTOR-PT

RAID controllers, the Fault/Locator LEDs for disks 3 and 4 are swapped.

CSCuu49909 If you disable TXT in the Cisco UCS Manager

BIOS policy after installing ESX with TXT enabled, ESX becomes unresponsive and displays the following message:

Workaround


When this issue occurs, do one of the following:

•

•

Re-enable TXT

Re-flash BIOS

First Bundle

Affected

2.2(5a)B

Resolved in

2.2(5b)B

2.2(5a)B

CSCuu61885

CSCuu71563

Relocating modules and starting up the kernel....

When UCS C460 M4 servers with Intel

®

Haswell processors are in UCSM mode, if the boot order policy is configured to boot from SAN, and the network adapter connected to it is installed on any PCIe slot other than PCIe slot 4, the BIOS may not show the SAN boot device in the boot order, and the system may fail to boot from SAN.

As a result of SAN boot failure, the system may become unresponsive at the end of the BIOS

POST, or attempt to boot from other boot devices configured in the boot order policy. If the system is booted from some other boot device, for example, Local HDD, the SAN LUN will still be accessible and usable.

When this issue occurs, do the following:

1.

Install the network adapter on PCIe slot

4

2.2(5a)C

2.

3.

4.

Re-acknowledge the server

Re-configure the boot order policy

If the issue still persists, try rebooting the system a few times.

The following caveats are open in Release 2.2(4b):



CSCuv32417 Cisco B200 M4 and B420 M4 blade servers that run on UEFI OS may reboot unexpectedly when

BIOS POST is not marked as complete.

Workaround


CSCuw23829 When inserting a HDD into the final disk slot of a C240 M4 server with 8 supported slots for disks, Cisco UCS Manager may displays a additional disk in the 9th slot, which is non-existent.

This error has no workaround. It does not affect any functionality. All the HDD operations can still be performed on the

HDD shown on the final slot. However, any operation or configuration attempted on the invalid HDD will result in an error.

First Bundle

Affected

2.2(4b)B

Resolved in

2.2(6e)B

2.2(4b)A

Resolved in

2.2(6e)A


OL-31127-01

Open Caveats


CSCuv20999 When Cisco UCS Manager is managed by Cisco

UCS Central, and there is a global service profile associated with a blade server, global VLANs cannot be assigned to a particular uplink interface through Cisco UCS Manager (VLAN

Uplink Manager > VLANs > VLAN

Manager), and the following error message is displayed:

Workaround

When this error occurs, do the following:

1.

2.

3.

Disassociate all the service profiles that would use the VLANs.

Manually configure disjoint layer 2 on

Cisco UCS Manager.

Associate the service profiles back to the blade servers.

First Bundle

Affected

2.2(4b)A

Resolved in

2.2(5d)A

Global Service Profile[xxxxx] cannot be modified from UCS domain. Please make the changes from UCS Central that you are registered with.

CSCuw50417 After upgrading the BIOS of Cisco UCS B200

M3 servers to any release between 2.2(4b) and

2.2(5c), the OS may randomly report the following messages in the /var/log/boot.gz file of the OS if the OS boot time is exceptionally slow:

When this error occurs, it disables the OS

PM capability, but has no known workaround to resolve it.

2.2(4b)B

Resolved in

2.2(5d)B

Initializing Power Management ...

Power: 2568: No supported CPU power management technology detected

"Intel Enhanced SpeedStep is supported but disabled in BIOS"

CSCuu89283 Service profile association fails and displays the following message:

There is no known workaround for this issue. 2.2(4b)A

Controller does not support non-default virtual drive properties

Resolved in

2.2(5c)A

This happens when you modify virtual drive policy attributes, such as read-policy, write-policy, and cache-policy, for a rack server with a UCSC-MRAID12G storage controller that has LUNS managed through storage profiles.

6


OL-31127-01

Open Caveats


CSCuv00089 After upgrading the Fabric Interconnect to Cisco

UCS Manager Release 2.2(4) or 2.2(5), systems that have OS-based NIC teaming configured as

Active-Standby will experience network connectivity issues. This is because the MAC addresses of teamed interfaces are learned statically on both FI server ports.

Workaround


•

Change NIC teaming to Active-Active configuration.

•

Note

Network connectivity issues still occur when you upgrade the FI to Cisco UCS

Manager Release 2.2(4), although it is resolved when you upgrade to Release

2.2(5).

•

Disable the slave or backup interface from OS. There will be no redundancy with this option, and it can restore network connectivity until you can implement fabric failover configuration.

Instead of OS-based NIC teaming, use a single vNIC in the service profile and enable fabric failover for it.

First Bundle

Affected

2.2(4b)A

Resolved in

2.2(5c)A

Note

This configuration change requires server reboot.

CSCuv13545 If Cisco UCS Manager is registered with Cisco

UCS Central, and the service profile refers to a global host firmware pack policy, UCSM downloads images instead of skipping it during shallow discovery.

CSCuv29668 When using Cisco UCS 6100 Series Fabric

Interconnects with Cisco UCS 2100 IOMs, some blade servers with two Cisco UCS M81KR VIC adapters fail discovery after updating Cisco UCS

Manager to Release 2.2.(4b) and later releases, or to Release 2.2(5a). The second adapter causes the discovery failure.

UCS Manager could display several errors including the following:

There is no known workaround for this issue. 2.2(4b)A

Resolved in

2.2(5c)A

Downgrade UCS Manager to a release earlier than Release 2.2(4b).

2.2(4b)A

Resolved in

2.2(5b)A sendsamedmeadapterinfo: identify failed adapter: repo lookup failed

OL-31127-01


Open Caveats


CSCuv04436 Cisco UCS B200 M4 servers with the following

CPUs may experience performance degradation:

•

E5-2667 v3

Workaround

There is no workaround for this issue. See the Software Advisory.

A patch will be released in July, 2015.

•

•

•

•

E5-2643 v3

E5-2640 v3

E5-2637 v3

E5-2630 v3

First Bundle

Affected

2.2(4b)B

•

•

E5-2630L v3

E5-2623 v3

•

•

E5-2620 v3

E5-2609 v3

CSCut54652 If the rack server is using UCSC-PCIE-C10T-02 or UCSC-MLOM-C10T-02 adapters, during service profile association, firmware upgrade on any third-party Converged Network Adapter

(CNA) fails with the following error:

When this occurs, do the following:

1.

Remove all vHBAs and vNICs from the service profile that is placed on

UCSC-MLOM-C10T-02 or

UCSC-PCIE-C10T-02.

2.2(4b)A

Resolved in

2.2(5a)A

Unable to find VNIC Device

Note

Re-association triggers automatically and the firmware upgrade on third-party CNAs will succeed.

2.

To the service profile, add the vHBAs and VNICs that were removed in step 1.


CSCuu53920 In a Cisco UCS B460 M4 server, service profile association fails when virtual drive creation is attempted by using a storage profile on a slave

RAID controller.

CSCut53878 When port profiles do not have any VLANs and both Fabric Interconnects are rebooted, STP cores are observed.

CSCuu52001 In a C-Series rack server with an external RAID controller, service profile association fails at

78% with the following error message:

2.2(4b)A

Resolved in

2.2(5a)A

To avoid this, during port profile creation, ensure that each port profile has at least one

VLAN.

2.2(4b)A

Remove the external RAID controller, and the service profile association will complete successfully.

2.2(4b)C

Resolved in

2.2(5a)A cannot support multiple scsi controllers

7


OL-31127-01

Open Caveats


CSCuu37369 A storage profile can be created in the main organization within Cisco UCS Manager, but it cannot be created through a sub-organization.

CSCuu42945 Service profiles created from service profile templates with local storage profiles fail association with the following error:

Workaround

Use the Cisco UCS Manager CLI to create a storage profile in a sub-organization:

1.

scope org

First Bundle

Affected

2.2(4b)A

Resolved in

2.2(5a)A

2.

3.

create storage-profile name

commit-buffer

You can also create a storage profile within the sub-organization service profile during service profile creation.

When this occurs, modify the description of service profile template.

2.2(4b)A

Resolved in

2.2(5a)A

Server unavailable, there are not enough resources overall, incomplete

LUN configuration.

CSCuu65128 If you configure more than 64 vNICs (excluding dynamic vNICs) on a single server, the Fabric

Interconnect reboots continuously due to an

LLDP core.

Do not configure more than 64 static vNICs on a single server (excluding dynamic vNICs)

2.2(4b)A

Resolved in

2.2(5a)A

When this occurs, reacknowledge the chassis to start discovering the servers.

2.2(4b)A CSCus73395 If the discovery policy is configured as

Platform-Max, and all links are not connected and acknowledged, servers are not discovered on first installation with import all backup configuration.

CSCuo62019 When adding or deleting multiple VLANs for a large number of port profiles as part of a single operation, the DME might restart.

CSCut72773 Installation of ESXi 5.5 U2 fails through PXE boot on Cisco UCS M4 servers in UEFI boot mode.

When this occurs, apply each addition or deletion as separate operations instead of as a bulk operation.


Alternate options include:

•

Use the boot script from ESXi 6.0

2.2(1b)A

2.2(4b)B

CSCut35123 After high availability failover, you may see chassis-seeprom local IO failure with the following message:

•

Use the legacy BIOS mode for PXE installation

When this occurs, reset the power supply unit.

2.2(4b)A

Resolved in

2.2(6c)A

Warning: there are pending SEEPROM errors on one or more devices, failover may not complete.

CSCus71522 Changes made to the user-defined flow record definitions for NetFlow are not updated and propagated to the vNIC that is already using the existing flow record definition.

When this occurs, remove the policy from the interface and reapply.

2.2(4b)A


OL-31127-01

Open Caveats


CSCut96983 For LUNs that are created by using the local disk configuration policy, the option to add UEFI boot parameters is not enabled. Hence, local LUN boot fails for UEFI after service profile migration.

Workaround

For UEFI boot from local LUN, use LUNs that are created by using storage profiles.

In Cisco UCS Manager, specify the vNIC/vHBA placement order manually using the following steps:

First Bundle

Affected

2.2(4b)A

CSCuo89748 Under rare conditions in a scale deployment, the

Cisco UCS Manager GUI may crash with the following error message:

Fatal Error: Event sequencing is skewed, Would you like to login again or exit?

When this issue occurs, exit the GUI and log in to it again.

You may continue to see the error message until the underlying triggering condition is resolved.

2.2(2a)A

CSCup43526 On a scale Cisco UCS system, creation and download of the technical support file from the

Cisco UCS Manager GUI times out with a timeout message.

CSCut78943 When a service profile is associated to a server with a Cisco UCS VIC 1340 or VIC 1380 adapter, the actual order of vNICs and vHBAs does not reflect the configured order.

Even though the timeout message appears in the Cisco UCS Manager GUI, the technical support file is created and saved locally.

2.2(1b)A

4.

2.2(3c)A

Resolved in

2.2(6c)A

5.

Expand Service_Profile_Name > vNICs.

In the Work pane, click the Network tab.

6.

7.

8.

In the Actions area, click Modify

vNIC/vHBA Placement.

In the Modify vNIC/vHBA Placement dialog box that appears, do the following:

a.

Choose Specify Manually from the

Select Placement drop-down list.

b.

Explicitly assign the vNICs and vHBAs to the vCons.

c.

In the list of Specific Virtual

Network Interfaces, select the vNICs or vHBAs that you have newly assigned to the vCons, and set

Admin Host Port for these vNICs or vHBAs as 2.

Save the configuration.

The following caveats are open in Release 2.2(3j):

7


OL-31127-01

Open Caveats

Table 47 Open Caveats in Release 2.2(3j)


CSCuv97713 After upgrading Cisco UCS Manager, in rare cases, the IOM may core in the sysmgr process leading to IOM reboot.

Workaround


The following caveats are open in Release 2.2(3h):

Table 48 Open Caveats in Release 2.2(3h)


CSCux10203 When you decommission a C-Series server after it was discovered in direct attached configuration, and the direct attached switch port was changed from Ethernet to FC mode, the following error message may appear:


When this issue occurs, switch the ports back to Ethernet mode from FC mode and allow the server decommission to complete.

CSCuw59409 When you upgrade a directly connected C-Series rack server without decommissioning, and connect the server to FI port, you may experience a DME crash issue.

Contact TAC to decommission the rack server from the back end.

2.2(3h)A

Resolved in

2.2(3j)A,

2.2(6e)A

2.2(3h)A

Resolved in

2.2(3j)A,

2.2(6e)A

2.2(3h)A CSCuw84010 Integrated C-Series rack servers with Seagate drives fail association when using the host firmware policy in a service profile.

Do one of the following:

•

Remove the C-Series rack server from

UCSM integration and manually updating the server with HUU.

Resolved in

2.2(3j)A,

2.2(6e)A

•

•

Remove the host firmware package from the service profile.

Remove the Seagate disks from the

C-Series servers and place them in

B-Series servers.

First Bundle

Affected

The following caveats are open in Release 2.2(3g):

First Bundle

Affected

2.2(3j)A

Resolved in

2.2(3k)A,

2.2(6g)A

OL-31127-01


Open Caveats

Table 49 Open Caveats in Release 2.2(3g)


CSCuv46749 When using Cisco B200 M4 blade servers with the UCSB-MRAID12G storage controller, the following random transient alerts or faults are incorrectly reported, and these alerts or faults get cleared in 30 to 40 secs:

•

Critical Fault [F1004] Controller

Inoperable, Reason: Device reported corrupt data

Workaround


•

Critical Fault [F1004] Controller

Inoperable, Reason: Device non-responsive

First Bundle

Affected

2.2(3g)B

Resolved in

2.2(6f)B

Note

When a Storage Controller device is non-responsive, the virtual drives and local disks for this controller are also reported as Inoperable as a result.

CSCuv51214 Messages log overfilled with "BMC is suspecting that palo is in boot block. Leaving

I2C bus alone." messages.


CSCuv28540 When you run Cisco UCS Manager Releases between 2.2(1) and 2.2(3) with catalog versions

2.2(4b)T or 2.2(5a)T, SAN boot fails on servers where the boot vHBAs are placed on a VIC. This is because the VIC PCI slot ID displays N/A in the UCSM inventory.

Downgrade the catalog version to the corresponding Cisco UCS Manager release and re-acknowledge the rack server.

2.2(3g)

Resolved in

2.2 (6c)

2.2(3g)A

Resolved in

2.2(5c)A with catalog version

2.2(5b)T

Note

This issue is resolved when you use catalog version 2.2(5b)T, but still occurs with catalog versions 2.2(4b)T and

2.2(5a)T.

The following caveats are open in Release 2.2(3f):

7


OL-31127-01

Open Caveats

Table 50 Open Caveats in Release 2.2(3f)


CSCux05389 Occasional VSAN misconfiguration occurs after upgrading to release 2.2(3f) and rebooting the subordinate fabric interconnect.

Workaround


•

•

First Bundle

Affected

2.2(3f)A

Change the VSAN on the HBA to a

VSAN that is not in use, and then change it back to the correct VSAN.

Revert to the previous release.

Resolved in

2.2(7b)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 do not allow traffic to pass.

CSCuu16791 The Firmware Auto Sync Server Policy is Auto

Acknowledge by default. This setting attempts to push the default Host Firmware Package firmware in to newly attached blades/rack-servers.

When this issue occurs, run the shut and no

shut commands on the affected ports, for example, vfc or FC uplink.

Change the Auto Sync Server Policy option to "No actions".

2.2(3f)A

Resolved in

2.2(6g)A

2.2(3f)A

CSCuu60867 Unable to discover UCS C240 M3 servers with dual RAID controllers through Cisco UCS

Manager.

When this issue occurs, remove one of the two RAID controllers and re-discover the server.

2.2(3f)A

Resolved in

2.2(5c)A

OL-31127-01


Open Caveats

The following caveats are open in Release 2.2(3d):

Table 51 Open Caveats in Release 2.2(3d)


CSCuu40978 If a large number of syslog messages are generated, the syslog file may fill the Fabric

Interconnect filesystem and prevent it from being written into.

Workaround

If you are unable to write to the filesystem, use the following commands on the Fabric

Interconnect to verify whether the issue is caused by the syslog messages file: connect nxos a | b show system internal flash | grep root

First Bundle

Affected

2.2(3d)A

Resolved in

2.2(3k)A,

2.2(6i)A

Note

In the output, check for close to 100 percent usage show system internal dir

/var/log/external

Note

In the output, note the size of the messages file in bytes

If the file is close to 100 MB:

From Admin > Faults, Events and Audit

Log > Syslog, reduce the size of the local syslog file. You can increase the size of this file later.

CSCuv31912 Cisco UCS Manager iptables are duplicating rules in the FORWARD table.

Longer term workaround:

Ensure that the severity for the local file is set to Critical.

If you still see significant growth of the file even after reducing the size of the local syslog file once, you either will need to repeat the steps or disable the syslog file entirely and use a remote destination.

Reboot the fabric interconnect to clear the entries.

2.2(3d)A

If the duplicate entries are in the thousands, or if there are connectivity issues to Cisco

UCS Manager because of missing rules in the

INPUT tables, contact TAC for help with the workaround.

Resolved in

2.2(3j)A,

2.2(6c)A

7


OL-31127-01

Open Caveats

Table 51 Open Caveats in Release 2.2(3d) (continued)


CSCuv06504 The svc_sam_dme process crashes during steady state of operation when a database(db) corruption is detected primarily on the subordinate fabric interconnect.

Workaround

There is no workaround. The system recovers on its own after svc_sam_dme restarts.

First Bundle

Affected

2.2(3d)A

Resolved in

2.2(3j)A,

2.2(6c)A

When the subordinate fabric interconnect detects a nvram or flash db corruption, it initiates a self-healing mechanism to rectify the sqlite db.

After the sqlite db on the subordinate is rectified, the svc_sam_dme reloads the db and during this reload it sometimes crashes.

CSCus73196 Installation of ESXi 5.5 Update 2 and later updates, or ESXi 6.0 and later updates on the

FlexFlash of a Cisco UCS B200 M4 server fails with the following error:

To avoid this issue, disk scrub the FlexFlash before installing an Operating System.

2.2(3d)B

"partedUtil failed with message: Error:

Can't have a partition outside the disk!

Unable to read partition table for device"

CSCur99740 While upgrading or downgrading the Fabric

Interconnect image on the Cisco 6100 series, you may see the following symptoms:

To recover from upgrade or downgrade failure, reboot using the Cisco UCS Manager

CLI, or power cycle the Cisco 6100 series FI:

2.2(3d)A

Resolved in

2.2(3e)A.

•

The Fabric Interconnect image upgrade fails while upgrading from Cisco UCS Manager

Release 2.2(3d) to Cisco UCS Manager

Release 2.2(3e) and higher releases.

Example:

UCS-A# connect local-mgmt

UCS-A(local-mgmt)# reboot

•

•

•

Auto-install fails during Fabric Interconnect image upgrade from Cisco UCS Manager

Release 2.2(3d) to Cisco UCS Manager

Release 2.2(3e) and higher releases.

To recover from Auto-install failure:

•

Reboot using the Cisco UCS Manager

CLI, or power cycle the subordinate FI.

The Fabric Interconnect image downgrade fails while downgrading from Cisco UCS

Manager Release 2.2(3d) to all lower releases.

Auto-install fails during Fabric Interconnect image downgrade from Cisco UCS Manager

Release 2.2(3d) to all lower releases.

•

•

Directly upgrade or downgrade the primary FI.

Reboot using the Cisco UCS Manager

CLI, or power cycle the primary FI.

CSCus76125 Global service profile will not resolve the LAN ping group even when the ping group is created in Cisco UCS Manager.

This isssue has no workaround. Please contact Cisco TAC for help with this issue.

Resolved in

2.2(3k)A


OL-31127-01

Open Caveats

The following caveats are open in Release 2.2(3c):



CSCus93431 Adding other storage, such as a local disk or a

USB to a SAN boot only policy deletes the SAN boot policy from Cisco UCS Manager.

Workaround

Modify the boot policy description, and save the boot policy again to Cisco UCS Manager.

First Bundle

Affected

2.2(3c)A

Resolved

2.2(3j)A,

2.2(6c)A

When this issue occurs, change the adapter to

Cisco UCS VIC 1240 or VIC 1280.

2.2(3c)B

Resolved in

2.2(4b)B.

CSCus81832 When a Cisco UCS VIC 1340 or VIC 1380 adapter is used with a Sandy Bridge CPU in a

Cisco UCS B420 or UCS B200 M3 server, the server hangs. This causes the KVM console window to become unresponsive and the installation of Red Hat Enterprise Linux 6.5 or

7.0 is not completed.

CSCur70034 Service Profiles with a common VLAN name configured on one of the vNICs fail with the following error: "Incorrect VLAN configuration on one of the VNICs".

CSCut45598 Kernel Panic on Fabric Interconnect.

To resolve this issue, either delete the local-vlan with the same vlan-ID on Fabric

A, or modify the service-profile vNIC to point to the Fabric-A vlan name.

None.

2.2(3c)A

Resolved in

2.2(3e)A.

2.2(3c)A

Resolved in

2.2(3g)A




CSCux85580 Fabric Interconnect cores will be seen on IGMP. This issue has no known workarounds.

CSCux71937 The CPU utilization always displays 100 percent for the kernel in the output of the show system

resources command. This could happen if the system had been up for a very long time (more than 200 days, but this time-frame could vary):

If this issue occurs, reload to reset the memory.

FI(nx-os)# show system resources

Load average: 1 minute: 0.50 5 minutes: 0.71 15 minutes: 1.04

Processes : 563 total, 3 running

CPU states : 0.0% user, 100.0% kernel, 0.0% idle

Memory usage: 3490164K total, 3140304K used, 349860K free

First Bundle

Affected

2.2(3b)A

Resolved in

2.2(7b)A

2.2(3b)A

Resolved in

2.2(3j)A,

2.2(6f)A

7


OL-31127-01

Open Caveats

Table 53 Open Caveats in Release 2.2(3b) (continued)


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.

Workaround

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.

First Bundle

Affected

2.2(3b)A

Resolved in

2.2(3j)A,

2.2(6c)A

If the drive is still accessible it may be possible to repair the file system corruption from the Linux shell as follows: a) unmount all partitions, /dev/sda3 ..

/dev/sda9. You might also have to unmount

/dev/mtdblock3 prior to being able to unmount /dev/sda7.

b) run 'e2fsck -n -f /dev/sdaX' for X = {3, ..,

9} c) 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.

Do one of the following: 2.2(3b) CSCut88909 When trying to map a Virtual Media, KVM becomes unresponsive. Map CD/DVD, Map

Removable Disk, and Map Floppy Options are not displayed.

- Eject the Optical Drive on the computer used to launch UCSM

- Keep a CD/DVD in the Optical Drive.

CSCuu15465 If the version of board controller on a UCS B200

M4 server is higher than the version in the Host

Firmware Pack, Cisco UCS Manager tries to downgrade firmware of board controllers on the servers. This power cycles the servers.

- Use a computer with no Optical Drive.


Resolved in

2.2(6c)

2.2(3b)A

Resolved in

2.2(3h)A

OL-31127-01


Open Caveats



CSCuq74763 On Windows 2012 hypervisors, vmNICs configured with NVGRE network virtualization fail to initiate network traffic when NVGRE NIC task offloads is enabled.

Workaround

To use NVGRE NIC task offload feature, upgrade the hypervisor to Windows 2012 R2.

Alternatively, disable the NVGRE task offload feature using Cisco UCS Manager or in the hypervisor while using Windows 2012.

First Bundle

Affected

2.2(3a)A

Note

Installing patch KB2779768 resolves the issue for ping operation but does not resolve the issue for TCP/UDP traffic.

2.2(3b)A CSCuq88436 XML Parsing Error due to multiple "jbod" diskState attributes is incorrectly specified when a server with two or more local disks is configured with "No RAID" disk config policy.


CSCuq95926 If you unplug the primary management cable when the HA is not complete, the Cisco UCS FI

HA status remains in “switchover in progress”.

CSCup46033 On a FI with a GEM module installed, the total default quantity of licenses does not match the total quantity of licenses available by default.

Restart the primary DME using the pkill

dme command, or reboot the FI.


CSCuq51890 entPhysicalDescr is returning the wrong getnext values during SNMPwalk.

Perform an SNMPwalk of

1.3.6.1.2.1.47.1.1.1.1.2.1 followed by

1.3.6.1.2.1.47.1.1.1.1.2.24 to get the remaining information.


CSCus97608 Faults in Cisco UCS Manager such as “error accessing shared-storage” and timeout/failover warning within the “show cluster extended-state” are displayed when several devices in several chassis are reporting EBUSY within the I2C logs.

CSCut02769

While performing a

dbsync

between Cisco UCS

Fabric Interconnects, the Cisco UCS database copies all the files, and just copies the required files.

This issue has no workaround.

2.2(2c)B

2.1(1f)A

2.2(2c)A

Resolved in

2.2(3g)A

Resolved in

2.2(3g)A

8


OL-31127-01

Open Caveats

The following caveats are open in Release 2.2(3a):



CSCuq57142 In a port channel universe, the following may occur:

•

•

Workaround

If this issue occurs:

1.


Note

Upgrade to a version of Cisco UCS

Manager that fixes the port channel universe

Upgrading Cisco UCS Manager does not fix pre-existing PC ID collisions, but prevents new PC ID collisions

After an upgrade, power loss, FI reboot or failover, the empty port channel universe is incorrectly interpreted as a new installation, and repopulated. This leads to duplicate port channel ID allocation when a server is attached to the FI, or when a server is re-acknowledged.

2.

3.

Identify servers that have port channel

ID collisions

Re-acknowledge servers that have a port channel ID collision

First Bundle

Affected

2.2(3a)A

Resolved in

2.2(3k)A,

2.2(6i)A

CSCuy62783 In a UCS setup with a VIC 13xx adapter on blade servers or rack-mount servers with Microsoft

Windows OS, the server or VIC adapter may become unresponsive after running IO across network file systems. As a result, the vNICs are disconnected from the OS, which leads to connectivity loss until server reboot. In the case of iSCSI booted systems, this may result in a

0x7B INACCESSIBLE_BOOT_DEVICE

BSOD.

If this issue occurs, do the following:

•

•

Reboot the server as an immediate fix

Disable TSO as a long-term workaround until the firmware can be updated

CSCuw44595 DIMMs with correctable ECC errors are marked

Inoperable or Degraded even though correctable errors do not affect normal system operation.

For Cisco UCS Manager Release 2.2(1a) and later versions, clear the error messages with the following CLI commands: scope server x/y reset-all-memory-errors

2.2(3a)A

Resolved in

2.2(6i)A

2.2(3a)A

Resolved in

2.2(3k)A,

2.2(6i)A

For Cisco UCS Manager Release 2.1(x), this example shows how to enter the DIMM mode for server 1/1 to reset-errors on a DIMM module 2 in memory-array 1:

UCS-A # scope server 1/1

UCS-A /chassis/server # scope memory-array 1

UCS-A /chassis/server/memory-array # scope dimm 2

UCS-A

/chassis/server/memory-array/dimm # reset-errors


OL-31127-01

Open Caveats

Table 54 Open Caveats in Release 2.2(3a) (continued)


CSCux68679 When a UCS B460 M4 server is configured with

Fusion IO cards installed in same mezzanine slot of the master and slave blades, actions such as

Cisco UCS Manager upgrade, cluster failover, fabric interconnect reboot trigger server reboot.

Workaround

To avoid this issue, install the Fusion IO card only in the master blade.

First Bundle

Affected

2.2(3a)A

Resolved in

2.2(6g)A

2.2(3a)B CSCut10525 The FlexFlash firmware version is not updating and the

FFCH_ERROR_OLD_FIRMWARE_RUNNIN

G is displaying in the fault summary after you update the UCS B200 M4 server firmware using

Cisco UCS Manager Release 2.2(4b).

CSCuv72975 When upgrading the Cisco UCS Manager infrastructure bundle and using VIC-1340 or

VIC-1380, the backplane port of the IOM and

VIC goes down.

Reset the FlexFlash Controller manually to remove the error.


Resolved

2.2(3j)B

2.2(3a)B

Resolved in

2.2(3j)B

CSCuw78688 When the FI is rebooted or UCS Manager is activated to a new version and the DME database has a corrupted entry, you may see one of the following issues:

•

FI ports will be in unconfigured state


If UCS Manager has lost configuration, you can reimport the configuration using the all-config backup file.

Contact Cisco TAC for more assistance.

2.2(3a)A

Resolved in

2.2(6e)A,

2.2(3j)A

•

No service profiles, templates, policies or pools will display in UCS Manager GUI or

CLI.

CSCus55944 In a Cisco UCS Manager configuration with a

C260 rack-mount server and two adapters, adding a VLAN to a vNIC on an service profile template immediately reboots the rack server that is associated with the inherited service profile.

1.

2.

Run the show detail command for both the adapters.

Note the managing instance of both the adapters and the roles of the fabric interconnects-primary and secondary.

2.2(3a)A

Resolved in

2.2(3j)A,

2.2(4b)A

3.

4.

Shut down the host-facing port from

FI-A, which is connected to the specified server.

Wait 5 minutes.

5.

6.

Repeat Step 1 and ensure that the managing instance of both the adapters is the subordinate FI.

Apply the service profile configuration change. The server reboots.

The activation status of both adapters displays Ready.

Any subsequent service profile configuration change will not result in a reboot unless the configuration itself requires a reboot.


8 OL-31127-01

Open Caveats



CSCur79257 When Cisco UCS Manager is configured as fc-uplink port-channel, and is being upgraded, the FSM fails and the vLANs are not added to FI

B uplinks.

Workaround

When this issue occurs, reboot the fabric interconnect.

First Bundle

Affected

CSCut37134 Cisco UCS B200 M4 servers running ESXi 5.5 will crash intermittently when booting locally.

1. Remove the lsi-mr3 RAID Controller driver and reboot

2. Re-deploy ESXi with the latest ISO available.

CSCup45804 VXLAN offload is enabled. Throughput is reduced because of two-way traffic between

VMs.

CSCuo11544 In some cases, NicAG gets stuck when getting events from adapter when that adapter is busy performing other tasks.


2.2(3a)A

To work around this issue, perform one of the following options:

2.2(3a)B

•

Check for servers that are stuck in discovery, association, or disassociation state and, if found, decommission them.

Then confirm that firmware update on adapters succeeds.

CSCuq09577 In some cases, a system running a 6100 FI setup with a large number of rack servers becomes unresponsive due to a high number of error messages overfilling the Cisco UCS Manager log file when an insufficient number of FEX/IOM uplinks are configured.

•

Instead of performing adapter firmware updates on multiple servers, perform a firmware update on servers in batches.

Make sure a sufficient number of IOM/FEX uplinks are configured so that vNICs configured on each service profile do not exceed the limit.

2.2(2c)A

CSCuq14172 Windows 2012 and Windows 2012R2 systems hang (boot and installation) when TXT is enabled via BIOS setup.

CSCuq28367 Host continuously reboots if scriptable vMedia mount fails and the same device type is also configured in boot policy.


Disable TXT and reboot.

Uncheck to disable the “Retry on mount fail” option in scriptable vMedia policy.

CSCuq33306 SAN boot with Emulex adapter fails with the first vHBA created when creating a Cisco UCS

Manager service profile boot policy to boot from

SAN LUN from the first vHBA.

Use the second vHBA for booting and map boot targets to the second vHBA in the

Emulex adapter.

2.2(2a)B

2.2(3a)A

2.2(3a)A


OL-31127-01

Open Caveats



CSCuq44049 In rare cases, when adding SD cards dynamically

(without powering down server) to a server without SD cards or when dynamically adding one SD card to a server with only one SD card, the server gets stuck in sync mode and disconnects existing drives accessed by host from an SD card.

Workaround

To avoid potential data loss, first disable

HOST VD access (from WEB UI or CLI) if

VD access is enabled. Then, after confirming each SD card is plugged in properly, execute reset of the flex flash controller (from WEB

UI or CLI).

First Bundle

Affected

2.0(3b)A

Place all vHBAs manually on host ports 1/2 or allow system to place all vHBAs.

2.2(3a)A CSCuq63592 In some cases, blade association fails on servers with Cisco UCS VIC 1340 or Cisco UCS VIC

1380 adapters when some vHBAs are placed manually on Host Port 1 or Host Port 2 and some are placed by system when "Desired Host Port" is set to ‘ANY.’

CSCuq64931 Blades with Cisco UCS VIC 1340 or Cisco UCS

VIC 1380 adapters fail remote OS boots when

SRIOV vNICs are placed on first host port and static vNICs (PXE, iSCSI, or vHBA) are placed on second host port.

Place the SRIOV vNICs on the second host port manually via GUI or CLI options to set host port when configuring service profile.

2.2(3a)A

CSCuq81262 If anonymous reporting (AR) is enabled after upgrading Cisco UCS Manager to release 2.2(3a) in an environment specifically configured with

VMs that use VM-FEX on any hypervisor, then

Cisco UCS Manager becomes unresponsive after seven days when VM-related data collection begins.

CSCuq92477 Infra bundle upgrades from Cisco UCS Manager release 2.2(3a) fail if performed from Cisco UCS

Central when upgrade process becomes unresponsive during 'PollActivateOfUCSM' stage if Cisco UCS Manager does not already contain the management image.

CSCut61527 The Cisco B200M4 blade server reboots unexpectedly when BMC returns invalid FRU information.

Disable AR feature in Cisco UCS Manager if running any VMs that use VM-FEX on any hypervisor. (The AR feature is disabled by default.)

If this issue (with AR enabled) is encountered, contact Cisco TAC for solution.

2.2(3a)A

Manually download the Cisco UCS management image or the complete infra bundle to Cisco UCS Manager before triggering the infra upgrade from Cisco UCS

Central.

2.2(3a)A

Resolved in

2.2(3b)A.

This issue does not have a workaround.

Resolved in

2.2(3g)A

CSCur96296

CSCuv53399

FEX goes online and offline in quick succession under scaled setups.

After IOM firmware upgrade, or after an IOM reset while running effected firmware, Cisco

UCS Manager will no longer show Major fault

F0481 indicating the IOM encountered a POST failure.

Avoid IOM/FEX to come back online quickly after it goes offline.

If this issue is encountered, upgrade to a resolved firmware version.

Resolved in

2.2(3g)A

Resolved in

2.2(3k)A

8


OL-31127-01

Open Caveats

The following caveats are open in Release 2.2(2e):



CSCur19358 The firmware upgrade of Nvidia GPUs fails with the error message "FI unreachable". This happens only when Broadcom adapters are used and its firmware is updated along with the firmware of Nvidia GPUs.

Workaround

Upgrade the firmware for Broadcom adapters first. After the firmware update

FSM is complete, another update FSM is triggered with the HFP meant for Nvidia

GPU cards.

First Bundle

Affected

2.2(2e)A




CSCur21496 Even after the firmware update of the K2 Nvidia

GPU adapter is successful, it does not display the version correctly. It displays the older version instead of new version.

CSCuo87059 In rare cases, backplane ports on an IOM that is providing a NULL serial number to Cisco UCS

Manager may be missing after chassis decommission and recommission.

CSCup00833 In rare cases, B440 M1 and M2 servers may fail discovery after CIMC activation during upgrade from release 2.0(2q).

CSCup05598 SNMP Community string shows blank text field even after adding and saving the string.

Workaround

This is a cosmetic issue and does not impact the functionality of the GPU cards.

Perform an additional decommission and recommission cycle.

2.2(2c)B

In Cisco UCS Manager CLI, use slot x/y

B-series server.

reset

where "x/y" designates the


This is a different convention than previously used but running a simple snmpget or snmpwalk against Cisco UCS

Manager VIP with the string just saved should confirm that SNMP is functional.

First Bundle

Affected

2.2(2d)A

Resolved in

2.2(3g)A

2.1(3c)B

2.2(1d)A

Note

Although the field is blank, the display of the 'Set: yes' or 'Set:

No' notifier indicates the value is present.

CSCup20961 Unable to add VIFs to a service profile or make modifications that require a new VIF on an existing interface, such as when adding fabric failover.

Restart svc_sam_dme

process.

2.2(1d)A


OL-31127-01

Open Caveats



CSCup38317 In some rare instances, typically during host power on, a board fails to discover and one of the following errors is displayed:

Workaround

Decommission and reacknowledge the server.

"

Reason: CIMC did not detect storage controller

" or "

CIMC storage subsystem not yet initialized.

"

CSCup40056 In some rare cases, if hypervisor and VM storage traffic are sharing the same fNIC(s) and FI is in switching mode and configured with Cisco VIC

1240 and with NPIV on Windows 2012R2, live migrations may fail.

Assign separate fNICs for hypervisor and

VM storage traffic.

Note

For more information, refer to troubleshooting technote

#117929 .

First Bundle

Affected

2.2(2c)B

2.2(1d)B

CSCup40441 In some cases, attempts to remove a service profile that contains two vNICs and two vHBAs from a

B200 M3 server with a 61xx FI and a M73KR-Q adapter running release 2.1(3b) will fail with the following error:

This issue is caused by an unsupported

WWPN in the service profile. If WWPN has the multicast bit set

(

Ex.20:01:01:01:31:00:00:23

), then

Qlogic flags the WWPN as invalid.

2.1(3b)B

Unconfigure server from service profile pre-boot environment

(FSM-STAGE:sam:dme:ComputePhysicalDisassocia te:PnuOSUnconfig)

Error 4106 - Unable to find FC Device.

To avoid this issue, always use WWPN with multicast bit turned off

(

Ex. 20:01:00:01:31:00:00:23/20:01:

00:00:31:00:00:23

). and Discovery fails with this error:

Discovery Fail Error: Unable to get Fibre

Channel Device Information from the system

(sam:dme:ComputeBladeDiscover:PnuOSInventory

)

Note

This issue can also be avoided by creating WWPN pool from Cisco

UCS Manager WWPN template.

CSCup45930 When upgrading Cisco UCS B230 M2 server from release 2.2(1x) to 2.2(2x) or from 2.2(2x) to

2.2(3x), or when upgrading Cisco UCS C460 M4 server from release 2.2(2x) to 2.2(3x), association fails with “

Unable to find storage controller

” error.

Once the problem occurs, the solution requires an RMA for the card.

Reacknowledge the server.

CSCup88161 In certain race conditions when running 2.2(1d) or later 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.

2.2(1e)A

Resolved in

2.2(2e)A,

2.2(3c)A

2.1(1d)A

Resolved in

2.2(4b)A,

2.2(2a)B CSCuq14172 Windows 2012 and Windows 2012R2 systems hang

(boot and installation) when TXT is enabled via

BIOS setup.


Disable TXT and reboot.


8 OL-31127-01

Open Caveats



CSCuq22707 In some cases, vMotion of multiple VMs at the same time causes stale MAC addresses after VM migration, which results in traffic loss between blades attached to the same FI.

CSCuq46105 In some rare cases, the Cisco UCS 6248 FI will unexpectedly reboot with following FI OBFL log in show tech:

Uptime: 10964, 0 days 3 hour(s) 2 minute(s)

44 second(s)

Reset Reason: Unknown (0)

Reset Reason SW: Unknown (0)

Reset Reason (HW): uC reset code: 0x0100

ADM1066 Power Good Triggered Reset

Card Mode..........................:

Runtime

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:

Cannot Create Certificate Request for default KeyRing

Workaround


The stale MAC addresses will eventually age out.


First Bundle

Affected

2.2(2c)A

2.2(1c)A

Resolved in

2.2(3c)A.

To work around this issue, try the following:

1.

2.

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

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

2.1(3d)A

This is accompanied by the following error messages in svc_sam_dme

log file in Cisco UCS

Manager techsupport log bundle:

=========

[INFO][0xac56abb0][Aug 6

05:20:47.182][exception_handling:rep]

FATAL[5|702]:

../feature/nuova/sam/sam/src/app/sam/dme/imp

/pki/MuPkiEpEndExplicitUpdateCbImp.cc(279):v alidateKeyRingNodes pkiKeyRing[sys/pki-ext/keyring-default] :

Cannot Create Certificate Request for default KeyRing

Note

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

[INFO][0xac56abb0][Aug 6

05:20:47.183][exception_handling:rep]

ERROR[3|702]

../feature/nuova/sam/sam/src/lib/framework/c ore/proc/Doer.cc(874):exceptionCB: exception encountered during processing: "Cannot

Create Certificate Request for default

KeyRing" [702] Cannot Create Certificate

Request for default KeyRing

=========


OL-31127-01

Open Caveats




CSCuw02439 When using Cisco UCS M81KR VIC adapters on a system running Cisco UCS Manager Release 2.2(2c), the adapters crash and generate core files. VIC adapter log files display the following error before the adapter crashes:


ecpumgr.main ERROR: ecpu 1 panic: ASSERT

FAILED (Exception 2 triggered!) @ mips/ecpu_panic.c:138

First Bundle

Affected

2.2(2c)B

Resolved

2.2(3k)A,

2.2(6i)B

CSCux59298 When using UCS B200 M3 servers with VIC 1240 on a system running Cisco UCSM Release 2.2(2c), network and SAN lose connectivity.

When this issue occurs, reboot the server. 2.2(2c)B

Resolved

2.2(3k)A,

2.2(6i)B

CSCus32933 Cisco UCS Manager does not display an error message when a WILL_BOOT_FAULT event is raised because of an incorrect CPLD version.

- Issue the following commands;

# scope server x/x

2.2(2c)B

Resolved

2.2(3j)B

# scope boardcontroller

# activate firmware 13.0 force

# commit-buffer

- Reset CIMC.

You might see the following messages in a var/log/messages file in CIMC show tech.

:will_boot:3438: platforms/castlerock/check.c:89:Procces sor[0]: Type: 1, Version:

17-10291-09_R01

:will_boot:3438: platforms/castlerock/check.c:97:Ivy

Bridge Processor requires update

:will_boot:3438: platforms/castlerock/check.c:89:Procces sor[1]: Type: 1, Version:

17-10291-09_R01

:will_boot:3438: platforms/castlerock/check.c:97:Ivy

Bridge Processor requires update

8


OL-31127-01

Open Caveats

Table 57 Open Caveats in Release 2.2(2c) (continued)


CSCur66094 You may see the following symptoms:

•

•

Multiple components stuck in discovery or upgrade state with no FSM progression.

Control plane service issue triggering faults in

UCSM. However, the OS on the server works normally.

Workaround

Contact Cisco TAC to manually restart the bladeAG process as needed on both

Fabric Interconnects.

First Bundle

Affected

2.2(2c)A

Resolved in

2.2(3g)

•

There may be UCSM process crashes and retries for DME and other processes, but these are not always present.

CSCur05013 Messages.log file on a Cisco UCS B200 M2 blade fails when the log file fills with messages including:

CSCur38408

364:BMC lost control of the bus.[0xd 0x0].

When a PCIe card is present in slot 2 and riser 2 is missing, fabric interconnect HA failover can cause the attached Cisco UCS C460 M4 servers to reboot.


2.2(2c)B

Resolved in

2.2(3d)B.

To work around this issue, try one of the following:

2.2(2c)A

Move the card in slot 2 to a different slot.

Resolved in

2.2(3d)A.

CSCuj09403 When installing RHEL 6.3 in UEFI mode with

M73KR-Q or M73KR-E adapters, the system hangs with a ‘ cannot exit boot services

’ message.

Add PCIe riser 2.

This issue has no known workaround with M73KR-Q or M73KR-E adapters.

2.2(1b)B

CSCuj63232 In some instances, most commonly after doing a

CIMC upgrade, certain long running operation data may indicate that an operation is currently running when it isn't. For example, the consistency check operation may report 0% progress and be stuck in that status.

Use VIC1240 adapter or use legacy boot.


To verify that the data is erroneous, use an LSI tool, such as WebBios or MegaCli, to confirm the operation is still in progress.

CSCul85363 Redhat 6.x installation fails on a B200 M3 that uses a Seagate ST300MM0006 single drive JBOD configuration with the No RAID policy.

To work around this issue:

Install the OS using custom layout by selecting "Create custom layout” when prompted for installation type.

2.2(1b)A

2.2(1b)A

CSCum91845 Server reboots on changing any management policy when Service Profile is derived from update Service

Profile Template that has no vNICs.

CSCun16169 Ethernet uplink port channel member interfaces are permanently down on primary FI after 20 iterations of port channel flap from Cisco UCS Manager.

Create custom layout and swap partition.


Remove and add port channel.

2.2(2c)A

2.2(2c)A


OL-31127-01

Open Caveats



CSCun22459 In some cases when the port is in trunk mode containing more than one isolated-primary VLAN pair, the VM comes up with vNIC configured with two isolated-primary pairs and does not get dynamic

IP address.

Use one of the following two options to resolve this issue:

1.

Do not put the port in trunk mode and configure one isolated-primary

VLAN pair per port.

First Bundle

Affected

2.2(2c)A

CSCun29274 In rare cases and only if blade is already discovered with a network adapter and user replaces it with storage adapter, the association or disassociation of that blade server may fail with error reported as

“

Invalid adapter connectivity, no adapter found

.”

2.

Do not set the native VLAN and, instead, explicitly tag the traffic with one of the VLANs.

Decommission and recommission the affected blade.

CSCun79973 In some conditions, when a Cisco UCS Manager user changes the QoS policy on a vNIC where Netflow session is actively enabled on a vNIC with transmit monitoring, the vNIC will hang.

To workaround this issue:

1.

Disable the Netflow session on the vNIC.

2.

Change the QoS policy.

3.

Reenable the Netflow session.

2.2(2c)A

2.2(2c)A

Resolved in

2.2(2d)

CSCun82594 DIMMS are mapped out during MRC (BIOS POST) when lockstep or mirroring are enabled in configurations with Ivy Bridge or Sandy Bridge CPU and with the following DIMM configurations:

The fix is provided via the Intel Memory

Reference Code update.

2.2(2c)B

•

•

1866 Hynix - Lockstep/Mirroring - 3 DIMMs

Per Channel.

1866 Samsung - Lockstep/Mirroring - 3 DIMMs

Per Channel.

CSCuo12665 In some cases, when user sets SAN boot for boot policy in service profile, the configured boot policy is not reflected on a blade equipped with a M71KR-Q adapter after association so the blade will not boot from SAN.

One of the following workarounds may resolve this issue:

1.

2.

Reset the blade

Upgrade the server firmware version to match with infra bundle version

(FIs / IOMs / Cisco UCS Manager)

2.2(1d)A

2.1(3b)B

9


OL-31127-01

Open Caveats



CSCuo34760 In some rare cases, if primary FI gets rebooted and comes back up as subordinate FI, some vEthernet interfaces will remain in down state on the newly subordinate FI (while previous subordinate FI becomes primary).

Workaround

One of the following workarounds will recover the affected servers:

1.

First Bundle

Affected

2.1(2d)A

(Preferred) CIMC reset of affected servers will force shallow discovery of the server which fixes this issue.

This will not reboot the server.

Resolved in

2.2(2d)

2.

IOM reset (which is connected to the affected FI) will also fix this issue.

However, traffic going through that

IOM for all blade servers on that chassis will be impacted unless there is failover adapter configs supported to redirect traffic through other working IOM/FI.

3.

Reacknowledge affected servers.

However, server gets rebooted.

Reboot subordinate FI to recover.

CSCuo38990 In some rare conditions, servers lose network connectivity during failover.

CSCuo41491 If boot policy has PXE boot listed prior to local disk

/ HBA, a blade my sometimes boot to pnuos instead of following normal boot policy when rebooted.

Re-associate Service Profile.

CSCuo51708 In some cases when attempting downgrade in middle of autoinstall upgrade, Cisco UCS Manager autoinstall downgrade will fail.

Manually activate UCSM version to the higher revision, and then let auto-install downgrade it.

2.1(3b)A

2.2(1b)A

2.2(1d)A

Resolved in

2.2(1g)A.

2.2(2c)B CSCuo55182 IPMITool version 14 or 14rc2 cannot obtain any

IPMI information on the slave blade in a B460-M4 configuration four-socket configuration.

CSCuo65728 In some instances, a MAC move is not learned by the

Cisco UCS FI after a VM is vMotioned between blades when the Cisco UCS FI is enabled with vNIC

Fabric Failover and hypervisor is configured to utilize active/standby for vNICs.

Use ipmitool version 13.

Reset the vEthernet interface on the target vMotion host to force a MAC relearn.

2.2(1c)A

Note

This is not a Cisco UCS best practices configuration.

OL-31127-01


Open Caveats



CSCuo98011 Cisco UCS Manager can crash (dump core) during Infra firmware upgrade

(ucs-k9-bundle-infra.2.2.2c.A.bin) when upgrade is initiated from Cisco UCS Central if Cisco UCS

Manager is registered to Cisco UCS Central using

IPv4 address and image bundles are not present in

Cisco UCS Manager.

Note

The core dump is associated with the

" svc_sam_dcosAG

" process and that process does restart itself after the crash.

Workaround

Avoid this issue by using either of two workarounds:

1.

2.

First Bundle

Affected

2.2(2c)A

Download the Cisco UCS Manager

2.2(2c) Infra bundle

(ucs-k9-bundle-infra.2.2.2c.A.bin) directly on to Cisco UCS Manager before initiating the firmware upgrade from Cisco UCS Central.

Download the firmware image into

Cisco UCS Manager and initiate the upgrade directly from Cisco UCS

Manager after using the following commands to ensure policy control resolution is set to "Local"

Resolved in

2.2(2d)A

CLI in UCSM

>scope system

>scope control-ep policy

>set infra-pack-ctrl source local

GUI in UCSM

Go to Admin > UCS Central > Policy

Resolution Control (Right Pa ne)

CSCuo68579 In some cases after a firmware upgrade, FC traffic stops working on fabric interconnect. However, blades in those chassis have successful flogi into the fabric.

Reboot affected Fabric Interconnect

On the upstream fabric switch fc4 feature types are missing when checking "show fcns database detail" for the affected blades.

2.1(3b)A

CSCuo81734 In some cases, FNIC driver bundled with driver disk does not load when installing on a standalone

C-series server and does not return error to indicate failure.

Perform RPM installation after system is up using the inbox FNIC driver.

2.2(1b)A

CSCup07488 ECC sensors report invalid “Upper

Non-Recoverable” data when there are existing failed PECI transactions on a blade.

Restart the ipmi stack while the host is up and running and

BIOS_POST_COMPLETE

is asserted. However, if PECI failures persist on the blade, this task will need to be repeated.

CSCut03086 When you auto install infrastructure firmware bundle and UCS Manager FI upgrades, in a VM-FEX scale set up with over 100 servers, DME core will happen.

Restart the DME by doing one of the following:

1.

Start and stop the PMON process.

2.

Kill the PMON process.

2.2(2c)B

Resolved in

2.2(2d)B

Resolved in

2.2(1h)A

9


OL-31127-01

Open Caveats




CSCuq63868 When creating a vNIC in a LAN connectivity policy, the show configuration command generates a software error.

Workaround


First

Bundle

Affected

CSCur16493 Customers with the UCSC-GPU-VGXK2 GPU adapter should update to release 2.2(3f) to receive updated firmware.

Upgrade to release 2.2(2e) and verify that the UCSC-GPU-VGXK2 firmware has been updated to version

80.04.F5.00.03_2055.0552.01.08.

2.2(2a)B

Resolved in

2.2(2e)B.

The following caveats are open in Release 2.2(1f):

Table 59 Open Caveats in Release 2.2(1f)


CSCur37260 FI upgrade or downgrade may fail due to lack of disk space in /mnt/pss. The following error may be seen:

PL20-B %SYSMGR-2-NON_VOLATILE_DB_FULL: System non-volatile storage usage is unexpectedly high at 99%

Workaround

Contact Cisco TAC to recover from this issue.

First

Bundle

Affected

2.1(3c)A

Resolved in

2.2(3d)A.

The following caveats are open in Release 2.2(1e):



CSCuo97907 Blades using a BIOS policy that utilizes the core multi- processing option do not honor changes made to the policy.

Workaround

Disable the cores using the operating system.

First

Bundle

Affected

2.1(3b)A


OL-31127-01

Open Caveats




CSCuw36128 Under rare conditions, there may be statsAG mts q leaks, and the following fault may be raised:

Major Fault "F0885: Fabric Interconnect B inventory is not complete card-inventory,eth-pc-inventory,eth-port-inve ntory,fc-pc-inventory,fc-port-inventory,mgmtport-inventory,remote-eth-port-inventory,swit ch-fru" seen.

Workaround

When this issue occurs, restart UCSM processes through the CLI:

UCS-A# connect local-mgmt

UCS-A(local-mgmt)# pmon stop

UCS-A(local-mgmt)# pmon start

First

Bundle

Affected

2.2(1d)A

Resolved in

2.2(3k)A,

2.2(6i)A,

2.2(7b)A

(It could be for either A or B side, or even both)

The error messages shown in the statsAG log file may include the following:

[MAJOR][0x7544eb90][date][app_sam_statsAG:pol lF] Error getting switch FRU inventory; details: SC_Send_tlv(3640) cmd_ucsm_req_send_recv() failed(-1)

[MAJOR][0x7544eb90][ date][app_sam_statsAG:getSw] Error getting

Line-Card inventory; details: Internal Error:

SC_Execute_show_command() returned -1

[MAJOR][0x7544eb90][ date][app_sam_statsAG:getPh] Error getting physical Ethernet port inventory; details:

Error in pm_get_if_index_listing(), errno: 16

[MAJOR][0x7544eb90][ date][app_sam_statsAG:getRe] Error getting remote physical Ethernet port inventory; details: Error in pm_get_if_index_listing(), errno: 16

[MAJOR][0x7544eb90][ date][app_sam_statsAG:getPh] Error getting physical FC port inventory; details: Error in pm_get_if_index_listing(), errno: 16

CSCut28278 After upgrading the infrastructure and the Cisco UCS

Manager image, the subordinate fabric interconnect fails the pre-upgrade check because of insufficient free space in the /var/sysmgr directory.

Contact TAC to verify whether discovery related log files filled the /var/sysmgr directory.

2.2(1d)A

Resolved in

2.2(3j)A

9


OL-31127-01

Open Caveats


Defect ID

CSCup95855

Symptom

CSCur01185 The HA policy of Reset triggers the Cisco UCS

6296UP fabric interconnect to reset.

Cisco UCS C240 M3 server upgrades fail because

FSM tasks are stuck in the throttled state in Cisco

UCS Manager.

Workaround

Check whether NPM is leaking memory by running the show npv internal

mem-stats detail command on the fabric interconnect

First

Bundle

Affected

2.2(1d)A

Resolved in

2.2(3j)A

To determine which adapter is causing the issue, view the output using the show npv

internal errors command.

The error message will list the VFC causing the issue.

-Verify that all CIMCs are accessible from the UCSM CLI (connect cimc X/Y for blades and connect cimc X for racks)

2.2(1d)A

-Resolve all communication issues between UCSM and any inaccessible

CIMC.

-Verify that netstat -anop output does not show receive queues filling up

Resolved in

2.2(3j)A,

2.2(6c)A

CSCuu15250 Toggling the Fabric Interconnect (FI) locator LED in

UCSM fails with the following symptoms:

•

UCSM GUI—The Locator LED will continue spinning and the On/Off options will be grayed out.

-Before running the restart dme command to recover from this scenario, take a core dump of the dme to analyze this situation further. Use the kill -s SIGABRT <dme pid> command to generate the core of the svc_sam_dme_process.

-Copy the generated core and attach it to

SR or the bug for further analysis

When this occurs, contact TAC to toggle the FI locator LED by using the debug plug-in access to NXOS.

2.2(1d)A

Resolved in

2.2(5c)A.

•

UCSM CLI—The following error will appear:

Error: Managed object doesn't exist

CSCus64439 Cisco UCS Manager Mezz logs and VMware vmkernel logs indicate storage latency and numerous

FNIC aborts.

Reboot blade servers where the PGS and

PFC features are not in the right order.

2.2(1d)B resolved in

2.2(3h)B To avoid this issue, 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.


OL-31127-01

Open Caveats



CSCus11782 After rebooting a Cisco UCS 6248UP Fabric

Interconnect (FI) that is operating in the FC end-host mode, some member links of the SAN port channel do not come up. Running the show interface brief command on the FI displays the notConnected status. On the Cisco Nexus switch, the corresponding ports show as invalidCfg.

Workaround


•

First

Bundle

Affected

2.2(1d)A

Run the shut and no shut commands on the non-working interfaces of the

Cisco Nexus switch.

Resolved in

2.2(5b)A,

2.2(3j)A

•

Enable trunking on the Cisco UCS

Fabric Interconnect and Cisco Nexus switch.

There are two possible workarounds: 2.2(1d)A CSCus12019 MAC address table on the FI not updated immediately after a MAC move or change.

•

•

The MAC address is updated automatically after a while (could range from a few seconds to a few minutes).

The MAC address update can be forced by clearing the MAC address table manually by using the clear mac address-table dynamic command.

CSCuo11700 In rare cases, a context switch while applying a new sam.config triggers a race condition that causes the httpd.sh or ucssh process to crash.

The httpd.sh process will recover after the crash.

2.2(1d)A

CSCuo60330 After upgrading the fabric interconnect to release

2.2(2) or later, the IOMs are unreachable, HA is not ready, and previous server interfaces show no configuration.

CSCul44421 You may see the “error accessing shared-storage” fault in Cisco UCS Manager. This message may appear any time but is seen more often during an FI reboot, upgrade, or IOM reset when running release

2.x of Cisco UCS Manager.

To fix the ucssh process crash, re-launch the Cisco UCS Manager CLI.



This fault is auto-cleared and transient so should not persist. The HA framework remains INTACT as long as show cluster extended output reports state as

PRIMARY_OK.

2.2(1d)A

2.1(2d)A

Resolved in

2.2(1e)A and

2.2(2c)A

9


OL-31127-01

Open Caveats



CSCum98771 Upgrading from Cisco UCS Manager 2.1(2a) or higher to release 2.2(1x) with FlexFlash State enabled in local disk policy of service profiles on servers causes the “

FlexFlash controller error, probably not supported, not equipped or inoperable

” config-failure.

This condition prevents updates to Blade and Rack firmware package bundles using auto-install or host firmware pack.

Workaround

First

Bundle

Affected

2.2(1b)A

Avoid this error using either of two workarounds:

1.

Disable FlexFlash in local disk policy of service profile. This will resolve the config-failure and CIMC can be upgraded using auto-install and host firmware pack. Since disabling

FlexFlash requires host reboot, host downtime is required.

Resolved in

2.2(1e)A and

2.2(2c)A

CSCun25187 When using third-party certificates with Cisco UCS

Central, status of Cisco UCS Manager might be displayed in Cisco UCS Central as 'Lost-Visibility' if a keyring with a certificate chain is configured to use https communication. This occurs when certificates are signed by subordinate CA instead of root CA.

CSCun86873 In some cases, during initial discovery or following upgrade-initiated discovery, B230 M2 shows only 8 of 10 cores available.

2.

Modify host firmware pack in service profile to use empty blade and rack package versions. Then directly update and activate CIMC firmware

(Equipment > Chassis > Server >

Inventory > CIMC).

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

Reacknowledge the blade.

2.1(2a)A

Resolved in

2.2(1e)A and

2.2(2c)A

2.2(1c)A

CSCuo12230 When vNIC of a service profile is configured to allow a VLAN defined in appliances cloud in the LAN configuration, the

show run interface vethernet

<x>

and show interface vethernet <x> switchport command outputs report different VLANs allowed on the virtual Ethernet interface.

CSCus74206 After disconnecting then re-connecting a cable between Fabric Interconnect and IOM, all VFC interfaces on the chassis are flapped once.

Check the outputs listed in the symptoms to confirm the problem. If the output shows the problem, then move the VLAN from the appliance cloud to the LAN cloud.

Remove the Pin Group setting in vHBA to use Dynamic Pinning instead of Static

Pinning.

2.2(1c)A

2.2(1d)A

Resolved in

2.2(3g)

OL-31127-01


Open Caveats



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:

Workaround

First

Bundle

Affected

To fix this issue, you can either temporarily lower your Java security settings to add Cisco UCS Manager as an exception or, if you are using Java 7 update 51 or higher, you can add the

Cisco UCS Manager host IP address to the Exception Site list.

2.2(1b)A

Resolved in

2.2(1e)A and

2.2(2c)A

Application Blocked by Security Settings.

Your security settings have blocked an application signed with an expired or not-yet-valid certificate from running.

To perform one of these workarounds, perform the corresponding steps below:

To temporarily lower your security settings:

1.

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

2.

3.

4.

5.

Lower the Security level to Medium.

Start Cisco UCS Manager.

In warning popup, check “I accept the risk and want to run this application” checkbox and click Run.

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

CSCup61601 FCOE_MGR crashes if FI is configured with too many noncontiguous VLANs.

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.

Note

If you use HTTPS to access Cisco

UCS Manager, ensure that you have the correct prefix.

3.

Click OK.

Decrease the number of noncontiguous

VLANs on the FI.

2.2(1d)A

Resolved in

2.2(2d)A

9


OL-31127-01

Open Caveats



CSCup61947 In some cases, MAC learning fails after adding 1000 or more VLANs with a large PV count with 120 or more virtual interfaces each with 1000 VLANs.

First

Bundle

Affected Workaround

To account for this issue, monitor the dleft_lif_vlan_mbr

table to ensure total

VLANs does not exceed 16,000.

2.2(1d)A

Resolved in

2.2(2d)A

If VLAN total approaches or exceeds this limit, reduce VLAN count and then reboot the FI as needed to recover from a

MAC learning failure state.

CSCup82677 A Cisco UCS system with ESXi OS will not exceed

10Gb speed even when QoS is configured with line-rate of 20Gb or 40Gb.

To resolve this issue, set the Rate(Kbps) value in the QoS policy to be the desired value above 10000000. The value must be numeric, such as 20000000 or 40000000.


2.2(1d)A

Resolved in

2.2(2d)A

2.2(1d)B CSCus85186 After activating Cisco Trusted Platform Module

(TPM), the enable and active statuses remain as disabled and deactivated.

Resolved in

2.2(1h)B,

2.2(3j)B,

2.2(4b)B

Resolved in

2.2(1h)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.

CSCuo50049 Cisco UCS Manager will experience HA cluster failover after upgrading from Release 1.4.

Avoid running OS applications that access the host SMBUS.


Resolved in

2.2(1h)A

OL-31127-01


Open Caveats




CSCum82888 After upgrading to Release 2.2(1b), you may see the following errors:

•

•

•

No access to the Cisco UCS Manager GUI.

The virtual IP is not reachable.

•

The virtual IP cannot be accessed by the GUI or

CLI/SSH.

Individual FIs can be accessed using SSH but not with http.

Workaround

To avoid this issue, do not delete the default keyring.

If this issue occurs, stop the bladeAG process.

First

Bundle

Affected

2.2(1b)B

Resolved in

2.2(1d)B

This occurs when the default keyring is deleted before upgrade, or if the default keyring is deleted after the upgrade and the system is rebooted.

CSCum09954 If an FC uplink interface is set to administratively shut, after upgrading to Release 2.2(1b) you may see a “FCoE or FC uplink is down on Vsan 202” error.

This issue is cosmetic and the error message can be safely ignored.

2.2(1b)A

Resolved in

2.2(1d)A

2.1(2c)A CSCum60793 You may see the following fault in Cisco UCS

Manager:

Restore connectivity to Cisco UCS

Central, or register the domain again.

[FSM:STAGE:RETRY:]: Report mount suspend success to operations manager errors

This occurs if Cisco UCS Central is not reachable or the domain has been unregistered.

CSCuo79500 Service dcosag falsely displays information that indicates a vNIC with a usNIC connection policy was automatically changed to dynamic vNIC when a customer adds a second vNIC with a usNIC policy.

This issue does not effect functionality.

The usNIC policies are properly set; the radio button selection falsely displays the selection of dynamic vNIC policy. Saved changes can be seen by clicking on the usNIC radio button.

Reset CIMC of the impacted servers.

CSCut54264 CIMC internal linux iptables deny additional mcserver connections between the CIMC and Fabric

Interconnect.

2.2(1c)A

Resolved in

2.2(2d)A

2.2(1c)A

Resolved in

2.2(3g)B

1


OL-31127-01

Open Caveats




CSCuv20324 The FCoE storage, in rare cases, becomes unavailable or performs very poorly. This can occur after a reset of either host-attached switch in the case of standalone systems, or after the reset of an IOM or FI in

UCS-managed systems when the IOM or FI sends configuration information to the adapter later than 5 seconds after link-up.

Workaround

Reboot the affected host after switch or

IOM reset is completed.

First

Bundle

Affected

2.2(1b)B

Resolved in 2.2(6g)B

CSCum50468 After upgrading the UCS Manager to version 2.2(6c) or higher, false faults for Fabric VSAN Membership

Down are not clearing.


2.2(1b)A

CSCuj63448 When upgrading to a catalog that supports new

DIMMs, some of the DIMM information is not displayed.

CSCui67824 When you disable port security from a network control policy with Hyper-V port profile, security is disabled on the nested port-profile, but security remains enabled on a few interfaces.

Reacknowledge the blade server.

1.4(4k)A


2.2(1b)A

CSCuj26767 Attempting to install UEFI OS second time sets incorrect boot order and stops installation.

CSCui52680 When the blade appears to be in normal operation state,

Cisco UCS Manager SEL message list might display the following error:

EFI Load Image Security Violation

Clean initial installation in the disk and restart installation.

If all blade functions appear to be normal, ignore this error message.

2.2(1b)A

2.2(1b)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

2.0(4a)A CSCug25894 During Cisco 2100 Series IOM boot and chassis reacknowledgment, sysmgr cores are seen.

CSCul74278 In some cases, if the boot policy is initially configured only for SAN devices and the policy is later modified to add local disk or local HDD device, the server fails to boot from the local disk.

The system resumes normal behavior after process restart. This should take approximately three minutes.

Reacknowledge the server.

2.2(1b)A

Resolved in 2.2(2c)A


OL-31127-01

Open Caveats



CSCum02561 When upgrading to 2.2(1b) from 2.1(3a), if you upgrade the infrastructure (A bundle) before the host firmware (B and C bundles), and then modify the service profile you may experience one or more of the following issues:

Workaround

1.

2.

To view missing entries for the configured boot order, use the Cisco

UCS Manager CLI to scope into the server and run

show boot-order

.

1.

2.

The configured boot order in the Cisco UCS

Manager GUI does not display SAN and local disk.

The server reboots every time the service profile is modified.

3.

To avoid server reboots, uncheck the Reboot on Boot Order Change checkbox before upgrading.

Upgrade the B and C bundles.

First

Bundle

Affected

2.2(1b)A

Resolved in 2.2(1d)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.


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

2.1(2a)B

CSCul72408 During upgrade, the following issues occured:

•

The IOM backplane ports show admin down in the

Cisco UCS Manager GUI.

Reboot the IOM in the affected chassis. 2.1(3a)A

•

•

The VIFs show non-participating.

The adapter shows DCE interfaces down.

CSCul95501 IOM FSM failed after downgrading from 2.2(1b).

Connectivity between the IOM and the FI could not be established.

Reboot the FI.

CSCuh89441 On B420 M3, the installed OS fails to boot from SAN

LUN with Gen-3 E.

Enable EDD 3.0 using the Emulex

M73KR-E FCoE option ROM.

2.1(3a)A

2.2(1b)A

CSCuj89557 In some rare conditions, it is observed during Cisco

UCS Manager upgrade, downgrade, or FI reload in a cluster setup that Cisco UCS Manager may fail on the secondary FI due to database corruption.

CSCtl04654 Changing native VLAN on uplink ports will cause the ports state to change and traffic disruption.

If you detect a db corruption in the secondary FI, erase the database from the FI and reconnect it to the cluster. The secondary syncs up with the primary FI and recreates the db.

If native VLAN changes are required on uplink ports, we recommend that they be performed during a maintenance window.

2.2(1b)A

2.1(1a)A

CSCuj78615 The EFI Shell boot option is disabled in the BIOS Setup

Utility > Boot Options page. Shell Boot option is not automatically launched by the BIOS.

Boot UEFI shell by using either one of the following two options:

•

Enter BIOS set up utility and select

Shell boot option from the Exit &

Save page.

•

Hit F6 to access boot override menu.

2.2(1b)A

1


OL-31127-01

Open Caveats



CSCuh13333 When you install ESXi in UEFI mode, the OS fails to boot-up. Even when you seem to have a smooth installation, the server does not boot the ESXi OS.

Workaround

Do the following:

1.

2.

Boot to Shell.

Determine fsxx (xx is where ESX is installed. It will be typically 0 i.e fs0:) Make sure you have the following:

First

Bundle

Affected

2.2(1b)A fsxx:\EFI\Boot\BOOTX64.EFI

3.

Use following command to get current list of EFI Boot Options bcfg boot dump

Note last boot option number to use as LAST_BOOT_No +1

4.

Use the following command to add new Boot Option at position

LAST_BOOT_NO + 1

Last parameter in quotes can be any description for new Boot

Option.

This will be displayed during BIOS

F6 menu.

bcfg boot add LAST_BOOT_NO + fsxx:\EFI\BOOT\BOOTX64.EFI "UEFI

: ESXi "

5.

Make the newBoot Option for ESX as the first by using the following bcfg boot mv LAST_BOOT_NO + 4 1

6.

Issue reset command at the shell and reset the platform. Press F6 when BIOS is booting to get into

BIOS Boot Selection menu. Make sure the new Boot Option is displayed and select this option to

Boot ESXi.

CSCuj31559 Installing Windows EFI PXE using UEFI is slow when compared to the installation using vNIC on MLOM.


2.2(1b)A

CSCuh34516 Installing Linux PXE using UEFI fails with the following message:


2.2(1b)A

Trying to allocate 971 pages for VMLINUZ.

CSCuj34520 TFM information in UCS Manager inventory contains invalid characters.

CSCuj81245 Multiple FNIC aborts in OS kernel and adapter logs when user starts a heavy write operation to SAN disk.


2.1(2a)A

Change IO throttle count from 16 to 256. 2.2(1b)A

Resolved in 2.2(2c)B


OL-31127-01

Open Caveats



First

Bundle

Affected

2.2(1b)B CSCun25692 Under certain timing conditions while allocating memory to critical data structure, M71KR-Q could run out of heap resulting in SAN boot failure."

To restore expected CNA behavior, downgrade to 2.1(3a) release or upgrade to version where this issue has been resolved.

Resolved in 2.2(1e)B and

2.2(2d)B

CSCus42584 The Cisco B200 M4 server unexpectedly reboots after replacing the IOM cables.

CSCut03052 IP QoS core may happen in a VM-FEX scale set up, when you reboot Fabric Interconnects back to back.


Resolved in 2.2(3g)A

Do the following:

1.

Resolved in 2.2(4b)A

Reboot the secondary Fabric

Interconnect and wait for discovery/association to stabilize.

CSCut63966

CSCus83447

Switch will stop at loader prompt upon reboot due to incorrect boot variables or /opt corruption.

A leap second update causes Cisco UCS Fabric

Interconnect to reload or switchover.

2.

Reboot the primary Fabric

Interconnect.

A workaround for this issue is to modify a script file, which is called before any reboots occur. This modification can limit the chances of experiencing this issue during an upgrade to firmware versions in which this issue is resolved.

Contact Cisco TAC for more information regarding this workaround, and to recover from this issue.

Do the following:

Resolved in

2.2(1h)A,

2.2(3g)A

1.

Remove NTP configuration from

Cisco UCS Manager at least 25 hours before the scheduled leap second.

Resolved in

2.2(1h)A,

2.2(3e)A

2.

Restore the previous NTP configuration after the leap second event.

CScur54705 Cisco UCS Manager sends the UCS Manager username and password hashes to the configured SYSLOG server every 12 hours.

CSCur88952 svc_dam_dme core will happen after upgrading or downgrading to Cisco UCS Manager Release 2.1(b).


Resolved in 2.2(1h)A


Resolved in 2.2(1h)A

1


OL-31127-01

Known Limitations




CSCuv55823 Faults like the following may occur when incompatible combinations of CIMC firmware and UCS Manager firmware are installed at the same time:

Controller 1 on server 1/2 is inoperable:

Reason: CIMC did not detect storage controller

Workaround

Use the same bundle version of CIMC firmware and UCS Manager firmware.

First

Bundle

Affected

2.2(1a)A

Resolved in

2.2(6e)A

CSCuv49345 When the server is discovered from only one Fabric

Interconnect (FI), and the DME primary instance is the other FI, disassociation of the associated service profile fails with the following message:

Decommission the server, then recommission it.

No connection to MC endpoint

2.2(1a)

Resolved in

2.2(5c)A

CSCus91342

The show fsm state command under server scope will show that the FSM failure is at

ResetSecureBootConfiguration

The blade server needs to reboot after adding a VLAN to a vNIC template.

CSCuj71400 Cisco UCS Manager displays the ’FCoE or FC uplink is down on VSAN X' fault when the member ports for the VSAN are up.

Reboot the blades for the change to take effect.

If this error occurs, do the following:

1.

2.

Unconfigure the uplink interface.

Reconfigure it as a FC uplink.

Resolved in

2.2(3g)A.

Resolved in

2.2(3k)A,

2.2(6i)A


CSCuo93591 For a fabric interconnect in end-host mode, the MAC address table aging time can get stuck at 300. When this happens, the value cannot be changed through

UCSM GUI or CLI.

Resolved in

2.2(3k)A


•

•

•

Known limitations for Cisco UCS Manager 2.2 releases are described in the following sections:

•

•

Board Controller Firmware in C-Series Servers, page 106

Cisco UCS Manager Discovery with a QLogic QLE8362 Adapter, page 106

•

•

BIOS and CIMC Firmware Downgrade Restrictions, page 106

LAN and SAN Topology Information Limitation, page 106

Default Zoning Not Supported in Cisco UCS, Release 2.1(1a) and Later Releases, page 107

Cisco UCS Manager Downgrade, page 107

Known Limitations and Behaviors, page 108


OL-31127-01


BIOS and BMC Firmware Downgrade

When a new UCS B200 M4 server with the Intel® Xeon® Processor E5-2600 v4 Product Family pre-installed with Cisco UCS

Manager Release 2.2(7) firmware is inserted into an existing Cisco UCSM deployment running the Release 2.2(3) infrastructure version, the firmware sync policy, if enabled, will downgrade the BIOS and BMC versions.

The server will fail discovery after the BIOS and BMC versions are downgraded. To recover from this issue, upgrade the BIOS and BMC versions back to the Cisco UCS Manager Release 2.2(7) firmware.

Board Controller Firmware in C-Series Servers

Upgrading board controller firmware in C-Series servers requires a power cycle for the upgrade to take effect.

Cisco UCS Manager Discovery with a QLogic QLE8362 Adapter

The firmware running version on a QLogic QLE8362 adapter that is managed through Cisco UCS Manager must be version

3.50.14 or later. If the firmware running version is earlier than version 3.50.14, discovery will fail. Firmware upgrade must be done out of a standalone Cisco UCS Manager setup.

BIOS and CIMC Firmware Downgrade Restrictions

The following BIOS and CIMC firmware downgrade restrictions were introduced in Cisco UCS Manager Release 2.2(5a):

•

For UCS C460 M4 servers with Intel

®

Haswell processors installed, Cisco UCS Manager prevents the BIOS and CIMC firmware from downgrading to versions earlier than Release 2.2(5a).

The following BIOS and CIMC firmware downgrade restrictions were introduced in Cisco UCS Manager Release 2.2(3a):

•

For UCS C22 and C24 M3 servers with Intel

®

Ivy Bridge processors installed, Cisco UCS Manager prevents the BIOS and

CIMC firmware from downgrading to versions earlier than Release 2.2(2a).

•

For UCS C220 and C240 M3 servers with Intel

®

Ivy Bridge processors installed:

–

Cisco UCS Manager prevents the BIOS firmware from downgrading to versions earlier than Release 2.1(3a).

–

Cisco UCS Manager prevents the CIMC firmware from downgrading to versions earlier than Release 2.2(2a).

LAN and SAN Topology Information Limitation

LAN and SAN topology information is available only through XML and CLI. This information is not exposed in the Cisco

UCS Manager GUI.

1


OL-31127-01


Default Zoning Not Supported in Cisco UCS, Release 2.1(1a) and Later Releases

Default zoning has been deprecated in Cisco UCS, Release 2.1(1a). Cisco has not supported default zoning in Cisco UCS since

Cisco UCS, Release 1.4 in April 2011. Fibre Channel zoning, a more secure form of zoning, is available in Cisco UCS, Release

2.1(1a) and later releases. For more information about Fibre Channel zoning, see the Cisco UCS Manager configuration guides for the release to which you are planning to upgrade.

Caution

All storage connectivity that relies on default zoning in your current configuration will be lost when you upgrade to Cisco UCS, Release 2.1(1a) or a later release. We recommend that you review the Fibre

Channel zoning configuration documentation carefully to prepare your migration before you upgrade to

Cisco UCS, Release 2.1(1a) or later. If you have questions or need further assistance, contact Cisco TAC.

Cisco UCS Manager Downgrade

Before performing any upgrade or downgrade operation, make sure to perform an all configuration backup of your system. The backup will ensure a seamless downgrade or upgrade, if required.

The following defects are related to downgrade issues:

•

•

•

CSCul55683 - When downgrading the Cisco UCS Manager image from Release 2.2 to any prior version, you must downgrade Cisco UCS Manager before you downgrade the infrastructure firmware on the FI. If you downgrade the FI first, some FI processes will crash.

–


CSCuj87553 - If you downgrade the Cisco UCS Manager image from Release 2.2 to any version prior to Release 2.0(4) before the fabric interconnect images are downgraded, the Cisco UCS Manager GUI may not work.

–

To recover from this issue, downgrade the kickstart and sytem images on the FI to the same version as the downgraded

Cisco UCS Manager image. The Cisco UCS Manager GUI will be able to reconnect successfully after the FI is running the same version as Cisco UCS Manager.

CSCul54029 - When you downgrade a B22 board controller to 2.1(2) version using auto-install, the activate status shows as failed. Also if you initiate individual board activation, the board control firmware downgrade is blocked.

–

To recover from this status, use “force” option to activate the board controller to the same version as the active one.

This clears the fault and brings activate status to “Ready”.

OL-31127-01



Known Limitations and Behaviors

The following known limitations and behaviors are not otherwise documented:

Table 65 Known Limitations in Release 2.2


CSCus53389 A Cisco UCS B200 M3 blade server running a Cisco

UCS Manager version earlier than Release 2.2(3a) with a 32-GB SD memory card (UCS-SD-32G-S) will be able to use only 16 GB memory.

Workaround

To use the entire capacity of the 32 GB

SD memory card, do the following:

1.

Ensure that the CIMC firmware version is Release 2.2(3a) or later

2.

Enable FlexFlash SD card support in a local disk policy

First

Bundle

Affected

2.2(1b)B

3.

4.

5.

6.

7.

8.

Add the local disk policy to a service profile

Associate the service profile with a server

Create a scrub policy

Add the scrub policy to the service profile

Format the FlexFlash SD cards

Reacknowledge the server

CSCuu02813

CSCtq79496

CSCui95113

When the network control policy is configured to include all host VLANs, and the number of MAC addresses exceeds 20,000, server discovery fails.

FCoE traffic may experience momentary disruption when system QoS policy is being modified.

Blade discovery fails with the following error:

Warning F77960 2013-08-29T17:47:46.044

18211988 [FSM:STAGE:REMOTE-ERROR]: Result: end-point-unavailable Code: unspecified

Message: sendSamDmeAdapterInfo: identify failed

(sam:dme:ComputeBladeDiscover:NicPresencePeer)

Warning F16520 2013-08-29T17:47:35.122

18211986 [FSM:STAGE:RETRY:]: detect mezz cards in 6/1(FSM-STAGE:sam:dme:ComputeBladeDiscover:

NicPresencePeer)

CSCun25132 On platforms with OOB storage controller, Cisco UCS

Manager displays usable (coerced) value in disk inventory section, which is different than the raw

'NumberOfBlocks' value displayed in catalog section.

9.

Reacknowledge the server

Change the network control policy to include only native VLANs, or reduce the range of allowed VLANs to keep the number of MAC addresses below 20,000.

This issue has no known work around. To avoid this issue, all QoS policy changes should be made during a maintenance window.

This issue is found when adapters are in

DIAG mode.

This is a non-issue; Cisco UCS Manager is designed to report the coerced, or usable, size as reported by the LSI controller. Both the host and OOB interfaces report this same value.

2.2(4b)A

2.0(1m)A

2.1(2a)B

2.2(1b)T

1


OL-31127-01

Related Documentation

Table 65 Known Limitations in Release 2.2


CSCuo14624 When FlexFlash member of dual SD card setup fails and needs to be replaced, the OS needs to be reinstalled, as well.

CSCup18983 MAC address changes after upgrade from 2.1 to 2.2.

Note

Workaround

This issue has no workaround.

First

Bundle

Affected

2.1(2a)B

This behavior is expected due to fix for

CSCtg93294, which offsets MAC address by 1.

Update MAC security on port with new

MAC address. From local-mgmt for each

FI, use show mgmt-ip-debug

command to find mac address.

2.2(1b)A

Related Documentation

For more information, you can access related documents from the following links:

•

Cisco UCS Documentation Roadmap

•

Release Bundle Contents for Cisco UCS Software, Release 2.2

Cisco UCS C-Series Rack Mount Server Integration with Cisco UCS Manager

For more information, refer to the related documents available at the following links:

•

•

Cisco UCS C-series Rack Server Integration Guides

Cisco UCS C-series Software Release Notes

Obtaining Documentation and Submitting a Service Request

For information on obtaining documentation, using the Cisco Bug Search Tool (BST), submitting a service request, and gathering additional information, see

What’s New in Cisco Product Documentation

.

To receive new and revised Cisco technical content directly to your desktop, you can subscribe to the What’s New in Cisco

Product Documentation RSS feed . The RSS feeds are a free service.

This document is to be used in conjunction with the documents listed in the

“Known Limitations” section.

Cisco and the Cisco logo are trademarks or registered trademarks of Cisco and/or its affiliates in the U.S. and other countries. To view a list of Cisco trademarks, go to this URL: www.cisco.com/go/trademarks . Third-party trademarks mentioned are the property of their respective owners. The use of the word partner does not imply a partnership relationship between Cisco and any other company. (1110R)

Any Internet Protocol (IP) addresses and phone numbers used in this document are not intended to be actual addresses and phone numbers. Any examples, command display output, network topology diagrams, and other figures included in the document are shown for illustrative purposes only. Any use of actual IP addresses or phone numbers in illustrative content is unintentional and coincidental.

© 2013–2015 Cisco Systems, Inc. All rights reserved.


OL-31127-01

Release Notes for Cisco UCS Manager, Release 2.2

Release Notes for Cisco UCS Manager, Release 2.2

Contents

Revision History

Introduction

System Requirements

Updating Cisco UCS Releases

Minimum B/C Bundle Version Requirements for Cisco UCS Manager Features

Hardware and Software Interoperability

Internal Dependencies

Capability Catalog

New Hardware Features in Release 2.2

New Software Features in Release 2.2

Behavior Changes

Security Fixes

The following are security fixes in Cisco UCS Manager Release 2.2:

Resolved Caveats

The following caveats are resolved in Release 2.2 (7b)

The following caveats are resolved in Release 2.2 (6i)

The following caveats are resolved in Release 2.2 (6g)

The following caveats are resolved in Release 2.2 (6f)

The following caveats are resolved in Release 2.2 (6e)

The following caveats are resolved in Release 2.2 (6d)

The following caveats are resolved in Release 2.2 (6c)

The following caveats are resolved in Release 2.2(5d):

The following caveats are resolved in Release 2.2(5c):

The following caveats are resolved in Release 2.2(5b):

The following caveats are resolved in Release 2.2(5a):

The following caveats are resolved in Release 2.2(4c):

The following caveats are resolved in Release 2.2(4b):

The following caveats are resolved in Release 2.2(3k)

The following caveats are resolved in Release 2.2(3j)

The following caveats are resolved in Release 2.2(3h):

The following caveats are resolved in Release 2.2(3g):

The following caveats are resolved in Release 2.2(3f):

The following caveats are resolved in Release 2.2(3e):

The following caveats are resolved in Release 2.2(3d):

The following caveats are resolved in Release 2.2(3c):

The following caveats are resolved in Release 2.2(3b):

The following caveats are resolved in Release 2.2(3a):

The following caveats are resolved in Release 2.2(2e):

The following caveats are resolved in Release 2.2(2d):

The following caveats are resolved in Release 2.2(2c):

The following caveats are resolved in Release 2.2(1h):

The following caveats are resolved in Release 2.2(1g):

The following caveats are resolved in Release 2.2(1f):

The following caveats are resolved in Release 2.2(1e):

The following caveats are resolved in Release 2.2(1d):

The following caveats are resolved in Release 2.2(1c):

The following caveats are resolved in Release 2.2(1b):

Open Caveats

The following caveats are open in Release 2.2(7b)

The following caveats are open in Release 2.2(6e)

The following caveats are open in Release 2.2(6c)

The following caveats are open in Release 2.2(5c)

The following caveats are open in Release 2.2(5b)

The following caveats are open in Release 2.2(5a)

The following caveats are open in Release 2.2(4b):

The following caveats are open in Release 2.2(3j):

The following caveats are open in Release 2.2(3h):

The following caveats are open in Release 2.2(3g):

The following caveats are open in Release 2.2(3f):

The following caveats are open in Release 2.2(3d):

The following caveats are open in Release 2.2(3c):

The following caveats are open in Release 2.2(3b):

The following caveats are open in Release 2.2(3a):

The following caveats are open in Release 2.2(2e):

The following caveats are open in Release 2.2(2d):

The following caveats are open in Release 2.2(2c):

The following caveats are open in Release 2.2(2a):

The following caveats are open in Release 2.2(1f):

The following caveats are open in Release 2.2(1e):

The following caveats are open in Release 2.2(1d):

The following caveats are open in Release 2.2(1c):

The following caveats are open in Release 2.2(1b):

The following caveats are open in Release 2.2(1a):

Known Limitations

BIOS and BMC Firmware Downgrade

Board Controller Firmware in C-Series Servers

Cisco UCS Manager Discovery with a QLogic QLE8362 Adapter