Table Of Contents
Release Notes for Cisco UCS Software, Release 2.0
Hardware and Software Interoperability
Known Limitations and Behaviors
Open Caveats from Prior Releases
New Hardware Features in Release 2.0
New Software Features in Release 2.0
Obtaining Documentation and Submitting a Service Request
Release Notes for Cisco UCS Software, Release 2.0
First Published: September 19, 2011
Updated: May 25, 2012
Part Number: OL-25363-01This document describes the new features, system requirements, and caveats for Cisco UCS Manager Release 2.0(1m), Release 2.0(1q), Release 2.0(1s), Release 2.0(1t), Release 2.0(1w), Release 2.0(1x), 2.0(2m), 2.0(2q), and 2.0(2r) and all related firmware and BIOSes on blade servers and other Unified Computing System components associated with those releases. Use this document with the documents listed in the Documentation Roadmap.
Note
We sometimes update the documentation after original publication. Therefore, you should also review the documentation on Cisco.com for any updates. Documentation updates and errata are also in these release notes. The documentation roadmap for this product is available at: http://www.cisco.com/go/unifiedcomputing/b-series-doc
Separate release notes for the VIC card drivers (which may be released out of synch with other software) are available at: http://www.cisco.com/en/US/products/ps10281/prod_release_notes_list.html
Table 1 shows the online change history for this document.
Table 1 Online History Change
OL-25363-01
A0
September 19, 2011
Created release notes for Release 2.0(1m). 1
B0
October 14, 2011
Updated release notes for Release 2.0(1q).
C0
November 8, 2011
Updated release notes for Release 2.0(1s).
D0
November 29, 2011
Updated release notes for Release 2.0(1t).
E0
December 15, 2011
Updated release notes for Catalog Release 2.0.1o.T.
F0
February 9, 2012
Updated release notes for Release 2.0(1w).
G0
March 16, 2012
Updated release notes for Release 2.0(1x).
H0
March 22, 2012
Updated release notes for Release 2.0(2m).
I0
April 9, 2012
Updated release notes for Release 2.0(2q).
J0
May 25, 2012
Updated release notes for Release 2.0(2r).
1 This release was removed from the download area due to CSCts96949 and CSCts86890. See the software deferral notice.
Contents
This document includes the following sections:
•
•
•
•
•
•
•
Introduction
The Cisco Unified Computing System is a next-generation data center platform that unites compute, network, storage access, and virtualization into a cohesive system designed to reduce total cost of ownership (TCO) and increase business agility. The system integrates a low-latency, lossless 10-GB Ethernet unified network fabric with enterprise-class, x86-architecture servers. The system is an integrated, scalable, multi-chassis platform in which all resources participate in a unified management domain.
System Requirements
To install Cisco UCS Manager your computer must meet or exceed the following minimum system requirements:
•
•
–
–
Adobe Flash Player 10 or higher is required for some features
•
–
–
–
–
Updating Cisco UCS Versions
To update the Cisco UCS software and firmware, see the appropriate Upgrading Cisco UCS document for your installation.
Use the scope firmware and show package filename expand CLI commands to view the contents of a given release package. File names in Table 2 and Table 3 are accurate for their initial release, expect small changes for subsequent patch releases.
Hardware and Software Interoperability
For detailed information about storage switch, operating system, adapter, adapter utility, and storage array interoperability, see the Hardware and Software Interoperability Matrix for this release, located at:
http://www.cisco.com/en/US/products/ps10477/prod_technical_reference_list.html
Note
Internal Dependencies
Table 4 shows interdependencies between UCS hardware and versions of Cisco UCS Manager. Server FRU items like DIMMs are dependent on their server type, and chassis items like fans and power supplies work with all versions of UCS Manager.
Caution
Table 4 Internal Dependencies (continued)
Software
Version
Software
VersionB200 M1
1.0(1)
2.0(2)
B250 M1
1.1(1)
2.0(2)
B440 M1
1.3(1)
2.0(2)
B440 M2
1.4(3)
2.0(2)
B200 M2
1.2(1)
2.0(2)
B200 M3
2.0(2)
2.0(2)
B250 M2
1.2(1)
2.0(2)
B230 M1
1.4(1)
2.0(2)
B230 M2
1.4(3)
2.0(2)
C200 M2
1.4(1)
2.0(2)
C210 M2
1.4(1)
2.0(2)
C250 M2
1.4(1)
2.0(2)
C200 M2 SFF
2.0(2)
2.0(2)
C460 M2
2.0(2)
2.0(2)
C220 M3
2.0(2)
2.0(2)
C240 M3
2.0(2)
2.0(2)
C260 M2
2.0(2)
2.0(2)
UCS 82598KR-CI
UCS M71KR-E
UCS M71KR-Q1.0(1)
2.0(2)
UCS M81KR
1.0(2)
2.0(2)
UCS NIC M51KR-B
UCS CNA M61KR-I 1
UCS CNA M72KR-Q
UCS CNA M72KR-E1.3(1)
2.0(2)
UCS-VIC-M82-8P
2.0(2)
2.0(2)
UCS 6120XP
1.0(1)
2.0(2)
UCS 6140XP
1.1(1)
2.0(2)
UCS 6248UP
2.0(1)
2.0(2)
UCS 6296UP
2.0(2)
2.0(2)
UCS 2104
1.0(1)
2.0(2)
UCS 2208XP
2.0(1)
2.0(2)
UCS 2204XP
2.0(2)
2.0(2)
Cisco Nexus 22482
1.4(1)
2.0(1)
Cisco Nexus 2232
2.0(2)
2.0(2)
Expansion ModulesN10-E0440
N10-E0600
N10-E00801.0(1)
2.0(2)
N10-E0060
1.1(1)
2.0(2)
UCS-FI-E16UP
2.0(1)
2.0(2)
SFP-10G-SR, SFP-10G-LR
SFP-H10GB-CU1M
SFP-H10GB-CU3M
SFP-H10GB-CU5M1.0(1)
2.0(2)
SFP-H10GB-ACU7M
SFP-H10GB-ACU10M1.4(1)
2.0(2)
FET-10G
1.4(1)
2.0(2)
SFP-H10GB-ACU7M=
SFP-H10GB-ACU10M=1.4(2)
2.0(2)
DS-SFP-FC8G-SW
DS-SFP-FC8G-LW1.3(1)
2.0(2)
DS-SFP-FC4G-SW
DS-SFP-FC4G-LW1.0(1)
2.0(2)
GLC-T (V03 or higher)
GLC-SX-MM
GLC-LH-SM1.3(1)
2.0(2)
1 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.
2 The C-series integration using the Cisco Nexus 2248 fabric extender is no longer supported as of release 2.0(2). See the UCS C-Series hardware documentation for details.
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 UCS Manager. Cisco UCS Manager 2.0 releases work with any 2.0 catalog file, but not the 1.0 catalog versions. If a server component is not dependent on a specific BIOS version, using it and having it recognized by UCS Manager is primarily a function of the catalog version. The catalog is released as 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
2.0(1m)
ucs-catalog.2.0.1j.T.bin
2.0(1q) and 2.0(1s)
ucs-catalog.2.0.1m.T.bin
UCS-MR-2X324RX-C on B440 M2
UCS-CPU-X5687 on B200 M2
-
ucs-catalog.2.0.1o.T.bin 1
UCSB-PSU-2500DC48 on the DC UCS 5108 chassis (UCSB-5108-DC)
UCS-HDD900GI2F106 on C200 and C210
UCS-HDD300GI2F105 on C200 and C210
2.0(1t)
ucs-catalog.2.0.1m.T.bin
2.0(1x) and 2.0(1w)
ucs-catalog.2.0.1p.T.bin
2.0(2m) and 2.0(2q)
ucs-catalog.2.0(2f)T.bin
Cisco UCS B200 M3 blade server
UCSB-MLOM-40G-01 on the B200 M3
UCSB-MLOM-PT-01 on the B200 M3
Cisco UCS 6296 fabric interconnect
Cisco 2204 IO module
Cisco VIC 1280 adapter card
Nexus 2232 fabric extender
C200 M2 SFF, C460 M2, C220 M3, C260 M3, and C240 M3 rack mount servers
2.0(2r)
ucs-catalog.2.0.2g.T.bin
1 Available for separate download.
Further details are in the Cisco UCS Manager GUI Configuration Guide.
Known Limitations and Behaviors
The following known limitations found in release 2.0(2) are not otherwise documented:
•
•
•
Pre Upgrade:
Modify any service profiles or service profile templates with iSCSI vNICs to have unique IQNs. Remove any duplicates. If necessary, use the PowerShell script provided in the upgrade notes to find out which iSCSI vNICs reuse the same iSCSI name.
Post Upgrade:
a.
b.
c.
d.
e.
The details of the PowerShell script are provided in the troubleshooting/upgrade guide. (CSCty29247)
•
•
•
The following known limitations found in release 2.0(1) are not otherwise documented:
Cisco UCS Manager
•
•
•
•
•
•
•
•
OS
•
Fabric Interconnect
•
•
BIOS
•
Upgrade and Downgrade Issues
•
•
To avoid this problem, do the following before upgrading from a 1.3.x release to a 1.4.x or a later release:
1.
2.
Resolved Caveats
The following caveats are resolved in the 2.0(2r) release:
•
•
•
The following caveats are resolved in the 2.0(2q) release:
•
•
•
•
The following caveats are resolved in the 2.0(2m) release:
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
UCS Manager
•
•
•
•
•
•
•
•
BMC
•
BIOS
•
Fabric Interconnect
•
•
•
The following caveats are resolved in the 2.0(1x) release:
•
•
•
•
Upgrade
•
The following caveats are resolved in the 2.0(1w) release:
•
•
•
•
•
•
•
•
•
The following caveats are resolved in the 2.0(1t) release:
•
•
•
•
•
The following caveats are resolved in the 2.0(1s) release:
•
•
•
The following caveats are resolved in the 2.0(1q) release:
Cisco UCS Manager
•
•
•
•
•
BIOS
•
Note
•
•
•
•
The following caveats are resolved in the 2.0(1m) release:
Cisco UCS Manager
•
•
•
•
•
•
•
•
•
•
Fabric Interconnect
•
Rack Integration
•
•
CIMC
•
•
Adapter Cards
•
VMware
•
•
RAID Controller
•
BIOS
•
•
Open Caveats
Release 2.0(2)
The following bugs were found in release 2.0(2r):
Symptom Hard drives from one manufacturer are two to three times slower than the hard drives from another manufacturer even though both are sold under the same product ID. This issue is observed with 300 GB SAS 10K RPM SFF drives.
Workaround
Symptom After upgrading from 2.0(2m) to 2.0(2q) release, a few false positive thermal alerts cause the chassis fans to spool up to full speed.
Workaround
Symptom After upgrading the Cisco UCS Manager from 1.4(3m) to 2.0(2m) release, the Cisco UCS Manager alerts a few warning messages to all the associated SPs and the vNICs configured for these SPs. Check if there are issues on the vNICs, the uplinks, and the VLANs. If there are no issues, these messages are not harmful to the network connectivity. (CSCtz36973)
Symptom The Cisco UCS B230 M2 Blade server running the ESX PSOD occasionally displays a processor error message.
Workaround
Symptom The Cisco UCS Manager generates a false VIF Down alarm even though the VIF is active on the Fabric Interconnect.
Workaround
Symptom The FC firmware issues disconnect the SAN and cause a server outage.
Workaround
Symptom The auto core transfer failure fault does not get cleared from the Cisco UCS Manager GUI.
Workaround
Symptom When two Cisco UCS systems push the same VLAN profile, the port profile from one Cisco UCS system disappears.
Workaround
Symptom After upgrading from 1.4(3i) to 2.0(1w) release, a false alert about a conflicting VLAN ID is observed on the VLAN. Deleting the conflicting VSAN does not clear the critical fault on the VLAN.
Workaround
Symptom After sending an inventory message from the Cisco UCS Manager, no registration email is received from the SCH Portal. Devices are registered on the SCH portal but the new inventory messages are not logged and the cases do not get created. The Contact field in the Contact information section contains a < or > character.
Workaround
Symptom While upgrading or discovering the Cisco UCS Manager, when the chassis discovery policy is changed to the set link-aggregation-pref port-channel policy, it creates the port-channel and the Fex gets offline for approximately 40 seconds.
Workaround
Symptom Some VFC interfaces are disabled with an error message after rebooting the Fabric Interconnect.
Workaround
Symptom When a new BIOS policy is created with the C1E state set to disabled from the Cisco UCS Manager, the ESX displays the C1E state as enabled. The BIOS setup menu displays that the C1E is disabled as BIOS policy from the Cisco UCS Manager.
Workaround
Symptom While upgrading CU environment from Release 2.0(1t) to 2.0(2q), one of the internal process crashes as long as the CU configuration is active. As a result, you cannot create service profiles from this template.
Workaround
Symptom Activating the Cisco UCS Manager to version 1.4(2b) results in a blade reboot.
Workaround
The following bugs were found in release 2.0(2q):
Symptom Server discovery fails when Cisco UCS Manager is not able to communicate with an adapter during the identification phase.
Workaround
Symptom If a server is configured to boot from an iSCSI LUN, then disabling the primary and failover NIC from the host OS will result in the host losing its connection to its boot disk which can lead to a host OS panic or BSOD. This occurs when both the primary and failover vNICs are disabled from the host OS.
Workaround
Symptom If a fabric interconnect or IOM is used for migrating VMs hosted on a Red Hat 6.2 KVM server is rebooted during VM migration, the VM migration will not complete until the IOM or fabric interconnect comes back on line. This occurs if the IOM or fabric interconnect reboots while VM migration is in progress.
Workaround
Symptom The Fabric Interconnect to IO Module link does not come up, or it experiences a high degree of packet CRC errors when using gen-2 10 or 7 meter twinax active cable and the 2204XP or 2208XP IO Module. This problem more likely to manifest with uplink number 3 as shown in an nxos cli show interface fex-fabric command.
Workaround
Symptom A server running ESX can only disable C1E when using the default BIOS policy. Once a new BIOS policy is created with C1E disabled from Cisco UCS Manager, ESX does not recognize C1E as disabled while BIOS setup menu and C-state dump from EFI all show C1E is disabled in the BIOS policy from UCS Manager. So as long as the policy is either set to default (not set) or a custom default (platform default), the problem is not seen.
Workaround
Memory DIMMs in a UCS blade may be marked as "Equipped Identity Unestablishable" if they are disabled during the power on self test. They do not have their smbios data filled in with the actual vendor data. Instead the vendor data in the smbios data is shown as "NO DIMM."
Workaround
Symptom During normal operation of the IOM, a kernel panic occurs and the IOM reboots and returns to normal operation.
Workaround
Symptom Blades reboot silently on doing any configuration change on the service profile after upgrading the Board Controller & CIMC firmware images using host firmware package & management firmware package. This is seen on B250 blades (N20-B6620-2, N20-B6625-2) when CIMC & Blade controller are updated at the same time, provided the CIMC running version is between 1.4.0 and 1.4.3s or 2.0 and 2.0.1t. These version ranges don't support board controller firmware upgrade on the above specified models. In the first association, boardControllerUpdate will get skipped if CIMC version doesn't support it. Any further change on the SP will reevaluate the SP and will find out that a board controller-update needs to be done and will trigger it.
Workaround
The following bugs were found in release 2.0(2m):
Symptom Boot order in BIOS setup or F6 menus still show Local HDD even after removing Local Disk option in UCS Manager service profile. This is seen when the boot order is configured by UCS Manager service profile with PXE eth0, PXE eth1, iSCSI iscsi0, iSCSI iscsi1, Local HDD. If you decide to remove the Local HDD option by deleting it from the boot policy service profile, after server rebooting, the boot order still shows Local HDD in BIOS boot order list. This behavior does not effect booting to PXE and iSCSI devices in the order configured.
Workaround
1.
2.
3.
4.
5.
6.
7.
8.
Symptom Under some rare circumstances, issuing NX-OS CLI show fex detail command after an IOM module goes offline and online may cause a Fabric Interconnect to reload.
Workaround
Symptom After an extended period of fabric port flapping configurations on border ports are missing.
Workaround
Symptom When an IOModule goes offline, in some cases the forwarding resource is not freed. When the IOModule comes back online, some VMFEX interfaces may not be able to come up. This requires the chassis connectivity mode be port-channel, and only applies to the case when there are close to 1000 VMFEX interfaces on the same chassis.
Workaround
Symptom In very rare circumstances, an uplink on UCS-2208XP or 2204-XP may experience a rapid link up/down (less than 250 ms interval). This may result in the Fabric Interconnect side of the link being in link down state but the IOModule side of the link being in an up state.
Workaround
Symptom LEDs for ports 1 and 2 on a UCS 6296 behave differently than other ports.
Workaround
Symptom On a 6100 series Fabric Interconnect with 20 chassis, some processes (svc_sam_dme and/or svc_sam_bladeAG) crash with the following message after repeated reack, association, disassociation, decommission, recommission of a chassis:
%KERN-1-SYSTEM_MSG: Proc svc_sam_dme (5082) with Total_VM 706000 KB Resident_Mem 544156 KB Anon_Resident_Mem 501068 KB being killed due to lack of memory - kernelThe processes are restarted automatically.
Workaround
Symptom A Windows Server 2008 baremetal host loses all network configuration and sees a NIC numbering shift. For example if Windows Device Manger showed NIC 1,2,3,4 before the upgrade, after upgrade, NIC 6,7,8,9 will be created. The "new" hardware will have lost all previous configuration such as IP address, DNS server, etc. This occurs after upgrading the UCS blade adapter card firmware to 2.0(2m). This has only been observed on the Microsoft Windows 2008 ("2k8") operating system on B230 (M1 or M2 blades) and B440 with the Cisco VIC M81KR adapter card. On Cisco C-Series servers this issue has been seen on C260 and C460 servers, also running Windows 2008 R2. This has not been observed on any other supported operating systems such as ESXi, or Red Hat Linux.
Workaround
Symptom The Cisco UCS Manager GUI hangs while retrieving data for the performance statistics table.
Workaround
Symptom Configuring a RAID policy on a rack server using an ICH10R controller fails.
Workaround
Symptom If recovery is initiated for a blade with a corrupted BIOS, there is a chance that recovery will hang. This occurs approximately once in 20-30 trials.
Workaround
Symptom When the BIOS setup is controlled by UCS Manager, the option to press F9 to load BIOS defaults in the BIOS menu is still active, which could change BIOS setup such that BIOS setup in the Cisco UCS Manager service profile might conflict with actual BIOS setup.
Workaround
Symptom Changing the dynamic vNIC policy to change the number of vNICs may cause static vNICs to get reordered on PCIe bus. This is seen under the following conditions:
1.
2.
3.
4.
5.
Create static vNICs and then create dynamic vNICs. Either create the dynamic vNIC policy with count less than 50 or greater than 56. (CSCtu34607)
Symptom Under certain conditions, a service profile will generate a configuration failure and a blade that it is associated with is removed from the server pool. This condition can cause service profile re-association to other available blades in the pool.
Workaround
F340-31-9-1-B scope orgF340-31-9-1-B /org # scope service-profile server 1/8F340-31-9-1-B /org/service-profile # associate server 1/8F340-31-9-1-B /org/service-profile* # commit-bufferF340-31-9-1-B /org/service-profile #(CSCtz07798)
Release 2.0(1)
The following bugs were found in release 2.0(1x):
Symptom All other connections are showing up and statistics look normal except a VM itself is still not receiving traffic. This is seen when the VM's vNIC is marked as masked in the VIF list. In this condition, it will not receive traffic.
Workaround
Symptom The Fabric interconnect rebooted with the following error during upgrade of fabric interconnect firmware to version 2.0(1x).
Reason: Kernel PanicSystem version: 5.0(3)N2(2.1w)Service:Workaround
Symptom VFC interfaces remain down and all static veths are stuck in the CR_RE state after a full-state restore. Servers may not work properly after the full-state restore operation if the existing server's configuration does not match the configuration in the backup file used for the restore. This is because UCS Manager does not automatically reconfigure the servers after the restore operation.
Workaround
Symptom Sometimes FEX host facing ports are not created/discovered in UCS Manager at the end of chassis/server discovery. This results in UCS Manager assuming that the adapter has connectivity to only one fabric. So that blade server cannot be used to associate with a service profile which has VNICs that require both fabric or the fabric to which connectivity is not yet discovered. This happens very rarely during chassis and server discovery.
Workaround
Symptom In a server using both a virtualized adapter card and a nonvirtualized card, if there are fewer service profile vNICS than the minimum required physical NIC ports then extraneous NIC ports are generated due to UCS Manager restrictions. (The minimum number of physical ports is 0 for UCS M81KR, 2 for other non-virtualized cards). This happens in the following circumstances:
1.
2.
3.
4.
5.
Workaround
Symptom Under some rare circumstances, a hot removal and insertion of an IO module results in a timeout of backplane port creation.
Workaround
Symptom A blade with a UCS M81KR adapter shows the error "initialize error 1" during iSCSI boot.
Workaround
Symptom Any action on a trunked port-channel (such as link, enable or disable) from the Cisco UCS Manager GUI takes a long time (over 2 minutes) to reflect on CLI. Once ports are up, they show as trunking, however vSANs are stuck in initializing.
Workaround
Symptom Blades power off unexpectedly, and stay off. A shallow discovery has happened which puts the blade into its desired power state. Some examples of actions that can trigger a shallow discovery are:
•
•
•
•
Blades that have been powered on with the following methods will be left in this inconsistent power state:
•
•
•
Workaround
Symptom When you create a vNIC template under the LAN tab a VM-FEX Port-profile is automatically created under the VM tab. This port-profile under the VM tab is created as a convenience for VM-FEX users.
Workaround
The following bugs were found in release 2.0(1r):
Symptom When using a UCS 2208 IO module you see a linkState fault for the UCSM virtual interface corresponding to the CMC management port (port 33 on the IO module).
Workaround
The following bugs were found in release 2.0(1s):
Symptom If you try to download core files from the GUI, the following error messages appear: "Failed to download file... reason: Server return http response code 401........" This is seen when Google analytics is enabled.
Workaround
Symptom The KVM application does not take keyboard inputs in windowed mode. This happens when the UCSM/KVM is run on a Linux Client.
Workaround
Symptom After a VLAN mapping change, a VNIC in ENM source pin will fail.
Workaround
Symptom Virtual machines using VM-FEX (Dynamic vNic) port-profiles lose network connectivity. This occurs on ESX/ESXi 4.1 U2 with UCS B-Series blades running Cisco VIC adapter cards. Some VM's (but not necessarily all) that are connected to the VM-FEX port-profile will lose connectivity. The VM guest OS may or may not show as disconnected.
Workaround
Symptom E2E diagnostic test was not using as much memory as possible depending on blade memory configuration. The test was updated to use as much memory as possible without running out of memory. This test is only available in manufacturing and if a blade is taken out of normal operation to run host diagnostic tests.
Workaround
Symptom If port profiles are configured for VM-FEX, the Fabric Interconnect may crash during upgrade due to a heartbeat failure.
Workaround
Symptom When following the steps in the Cisco upgrade guide and activating the subordinate FI, guest virtual machines will experience a high CPU load, but no other performance problems.
Workaround
1.
2.
3.
Symptom On a Service Profile (SP) configuration change, a Server is rebooted before the maintenance window. This will happen if you make some configuration change on the SP which doesn't require a reboot. Then immediately make another change which requires a blade reboot. The blade will reboot immediately.
Workaround
The following bugs were found in release 2.0(1q):
Symptom On a Service Profile (SP) configuration change, the changes list also shows "Networking" changes to be deployed even if there's no configuration change done in the networking area.
This could happen because of various configuration changes.
Workaround
The following bugs were found in release 2.0(1m):
Symptom In UCS Manager GUI, a power cycle with graceful operating system shutdown does not shutdown the operating system gracefully.
Workaround
Symptom After specifying an attribute setting in either the "General" LDAP setting or under the LDAP Provider setting, LDAPD crashes when testing LDAP. This happens with both the Cisco UCS Manager CLI and GUI.
Workaround
Symptom Currently in UCS Manager there is no auto-creation of IQN identifiers for iSCSI. IQN's must be manually entered for each iSCSI adapter. There is also no validation on the IQN format.
Workaround
Symptom Hard drives sourced from one manufacturer are two to three times slower than hard drives from another manufacturer even though both are sold under the same product id. This is seen with 300 GB SAS 10K RPM SFF drives.
Workaround
Symptom During upgrade from UCS 1.4 to 2.0, an SSLCert error may be written to the log files.
Workaround
Symptom After kernel rebuild or update a server configured for SAN Boot of RHEL 5.6 or RHEL 5.7 may fail to boot. This is a Red Hat issue, and ticket 744330 has been filed. It is a private ticket that may be referenced when calling Red Hat support for more information.
Workaround
Symptom While upgrading to UCS 2.0 with QoS policies defined, critical errors will be displayed for all QoS policies and VIFs with QoS policies defined on them will be down after upgrading the subordinate interconnect but before upgrading the primary interconnect. Expect that during the upgrade there will be a period of downtime between when the primary restarts and when the secondary becomes primary and brings up its VIFs. During this time all blades will lose their connectivity to both LAN and SAN.
Workaround
Symptom A host configured for iSCSI boot will always boot off a LUN exported to the Primary iSCSI VNIC by default, as iBFT is always posted on the primary iSCSI vNIC. The secondary iSCSI vNIC will post iBFT in case the LUN discovery fails on the primary iSCSI vNIC. However, if the Secondary iSCSI vNIC comes up earlier than the Primary iSCSI vnic (due to its overlay vNIC having a lower PCI order than that of the overlay vNIC for the Primary) and LUN discovery fails on the Primary, then there is no iBFT posted and the host fails to boot.
Workaround
Symptom Network connectivity is affected (flapping on uplink ports) on both fabrics during operations such as native VLAN change when the configuration change is done on both interconnects at the same time.
Workaround
Symptom After upgrade to 2.0(1m), blades with UCS M81KR adapters may show the error "initialize error 4" during FC boot.
Workaround
Open Caveats from Prior Releases
Release 1.4(3)
The following bugs were found in release 1.4(3):
Symptom The DME process on one Fabric Interconnect frequently crashes when the peer fabric interconnect is in an inoperable state.
Workaround
Symptom A Fabric Interconnect reboots unexpectedly. Using the show system reset-reason command returns "Reset triggered due to HA policy of Reset. "
Workaround
Symptom PAA for a SPAN session does not works with 8Gb transceivers and Fiber channel expansion modules on the Fabric interconnect.
Workaround
Symptom When a UCS M81KR or UCS M71KR-E/Q (NIV) adaptor is used, and the isolated host is communicating using a MAC configured in the service profile (and registered at the interconnect via a VIC) PVLAN traffic does not flow for an isolated host. Using the show platform fwm info hw-stm command at the NX-OS prompt shows that the isolated host MAC is learned on the isolated VLAN, but not learned on the primary VLAN.
Workaround
Symptom Blades unexpectedly reboot on UCS Manager activation when upgrading from UCS Manager 1.3(1) to 1.4(3s).
Workaround
Symptom When an IOM reboots after a software update on a full width blade, the HIF ports on the second adapter are not brought up by the IOM.
Workaround
Symptom Thermal faults are confusing as they do not have meaningful details.
The causes for thermal conditions are:
•
•
•
•
Workaround
Symptom A BladeAg crash is observed if a request bios_recovery_ctrl message is sent to a blade, but the response came back too late and is ignored by mcclient.
Workaround
Release 1.4(2)
The following bugs were found in release 1.4(2b):
Symptom An Intel Westmere-EP CPU on a B200-M2 or B250-M2 blade may not be able to perform to its full extent when running a subset of 1.4.x BIOS versions. The BIOS may incorrectly initialize a value during boot up which keeps the CPU at P1 even when P0 is requested by an OS.
Workaround
Symptom If the installed DIMMs do not have thermal sensors (the most likely cause as this warning is logged during initial system memory initialization) or the installed DIMMs exceeded the thermal threshold values programmed in either the memory controller or the Memory buffer, then the RankMargintest file in the CIMC shows the following warning code:
MRC - Warning Code:0x9 on Socket#1 Br#0 Ch#00, Ddr#00, Dimm#00, Rank#FF (if applicable)MRC - Warning Code:0x9 on Socket#1 Br#0 Ch#00, Ddr#01, Dimm#00, Rank#FF (if applicable)Workaround
Release 1.4(1)
The following bugs were found in release 1.4(1i):
High Availability
Symptom High Availability does not become ready until all 3 selected HA devices (chassis/rack-unit) have been discovered. The condition that triggers this problem is the fact that a previously functioning and fully discovered device (either chassis or rack mount server) has failed. This may be due to connectivity problems or faulty behavior. In this case the system remains in HA NOT READY state.
Workaround
Symptom When UCS Manager is operated in High Availability mode, SNMP traps stop arriving as expected if the SNMP trap IP header source address field is set to the cluster virtual IP address.
Workaround
BIOS
Symptom Blade and Rack Servers that include unequal sized HDDs or SSDs have encountered various failures intermittently. UCS Manager reports "Error Configuring Local Disk Controller" in most cases during these failures, though other errors are also seen.
Workaround
Symptom The BIOS Setup shows less memory size and some DIMMs disabled on a B230. The SMBIOS table does not report memory info in Type 17 structure for the disabled DIMMs. The UCSM reports less memory size and some DIMMS disabled. This happens when the BIOS disables some DIMMs incorrectly when DIMMs on certain slots do not have the corresponding lockstep pairs installed. For example, configurations that can cause this failure include:
1.
2.
3.
Workaround
CIMC
Symptom When using mirroring mode, if a UCE error happens, there is a Redundancy SEL event and also a UCE SEL event. No other details are available for the Data Parity error.
Workaround
Adapters
Symptom The link from the Rack server adapter to the fabric interconnect port remains down if the SFP type is FET (Fabric extender transceiver). Currently the FET type is supported only between a fabric extender and a fabric interconnect. If the SFP used for the link between the IOM and the rack server adapter is an FET, the link will remain down.
Workaround
Symptom CRC errors reported on an M81KR network interface on SLES 11 SP1. This is seen under High TX and RX traffic on SLES 11 SP1. The FIFO is not cleared as fast as it should because of some delays in the PCI path. An M81KR firmware devcmd storm from the host is also investigated to be one reason for the PCI stalls. These CRC errors are actually caused by FIFO overruns on the M81KR which are in turn caused by PCI stalls. They are not real CRC errors but truncated packets (due to FIFO overrun) flagged as CRC errors.
Workaround
UCS Manager
Symptom LicenseAG crashes during UCS Manager restart after downloading license files.
Workaround
Symptom UCS Manager takes a long time to push configurations containing large number of port-profiles to the VMware Virtual Center (VC). This happens when large number of hosts and virtual machines (VMs) are configured (tested with 500 VMs on 60 hosts managed by same VC) and a large number of port-profiles assigned to multiple DVSes in the VCenter.
Symptom Wait until the operations complete. Configuration FSM could take more than 30 minutes. Then, configure a smaller number of port-profiles (we've measured up to 92 profiles to successfully configure in around 5 minutes in above large number of hosts and VMs situation). (CSCtk97755)
Symptom When 15 or more chassis are configured in release 1.4(1) and the system is downgraded to release 1.3(1), chassis beyond 14 are still there. This may cause some issues as the max chassis support enabled with the 1.3(1) release is 14.
Workaround
Symptom When a rack server has a local disk installed, it does not report real-time disk operability status (disk operability is reported as "N/A").
Workaround
Symptom When the Cisco UCS Manager shell mode is set to s either management or local-management mode, the CLI command terminal monitor is not available.
Workaround
Symptom The minimum power cap that can currently be set is 3400W. The chassis power cap has a lower limit of 3778 W (AC), which is internally converted to 3400 W (DC).
Workaround
Symptom The Cisco UCS Manager shell does not support redirection of show command output to a remote file system.
Workaround
Symptom In UCS Manager there is no option to change the port speed of the SPAN destination port.
Workaround
Symptom UCS Manager reports an unsupported DIMM as missing but does not raise a fault.
Workaround
UCS Manager GUI
Symptom Non-disruptive pending changes may not be shown on a service profile. When a service profile has a maintenance policy that defers the application of disrupting changes to the server, user can see what changes are pending and make further changes. Disruptive pending changes are always visible on the service profile, whereas non-disruptive changes may not be shown. Non-disruptive pending changes are only shown for user convenience. This defect has no functional impact.
Workaround
UCS Manager CLI
Symptom When VIF creation doesn't follow V-motion, NPPM doesn't move the VIF to a new dynamic and the stale VIF doesn't carry the traffic any more. This would cause the SPAN to stop monitoring the traffic from the original VM.
Workaround
Fabric Interconnect
Symptom Fabric interconnect activation during a downgrade from 1.4(1) to 1.3(1) will fail if the setup has an active Nexus 2248 fabric extender.
Workaround
Symptom The server UUID displayed by ipmitool does not match that shown by the Cisco UCS Manager CLI. UCS UUID encoding follows pre SMBIOS 2.6 specified encoding, which is big-endian encoding. Ipmitool does not work well with that encoding. The SMBIOS 2.6 specification mandates mixed encoding (first 3 fields little-endian, last 3 big-endian), which is followed by ipmitool.
For example, The server detail from UCS Manager CLI shows -
Dynamic UUID: 0699a6f3-1b81-45f8-a9f2-c1bbe089324e# ipmitool -H 10.193.142.104 -U gurudev -P password mc guidSystem GUID : f3a69906-811b-f845-a9f2-c1bbe089324eCompared to UCS Manager CLI or GUI output, the first 3 fields f3a69906-811b-f845 show up differently in the output of ipmitool.
Workaround
#ipmitool -H 10.193.142.104 -U gurudev -P password raw 0x06 0x3706 99 a6 f3 1b 81 45 f8 a9 f2 c1 bb e0 89 32 4eThe output matches the value printed by the Cisco UCS Manager CLI. (CSCtk09043)
Symptom The show port-security NXOS CLI command returns a negative value for the Max Addresses. This will occur when a system is configured with more than 8192 Port VLAN instances and. port security is enabled on all interfaces such that more than 8192 MACs are secured.
Workaround
Symptom When an N2XX-ACPCI01 adapter port on a C-series server is connected to an uplink port on a UCS 6100 fabric interconnect, a fault message should appear because this connection is not supported, but there is no such fault message for this situation in this release.
Workaround
CIMC
Symptom The Blade CIMstic management IP address assignment is not included in backups.
Workaround
RAID/Local Disk
Symptom UCS Manager fails to report the Local Disk Failures, Faults, Alarms, Status and Disk Errors/Error Codes from the MegaRAID Controller on the B440 server.
Workaround
Symptom For the B200 and B250 blade servers, the Local Disks 'Operability' field is reported as "N/A". The 'Operability' field of the Local Disks in B200 and B250 is expected to have a correct value and should not be reporting 'N/A'.
Workaround
Symptom For the MegaRAID Controller on the B440 blade server, UCS Manager fails to report BBU Status, Properties and Errors.
Workaround
Symptom If one or more conditions are met, UCS Manager fails to capture certain Local Disk errors. Conditions include: Mixing the SAS and SATA Local Disks in the same server; Disk spin-up or disks present but not reaching 'Ready' state; Missing Disks.
Workaround
Symptom UCS Manager does not include the implementation for the Write Through, Write Back, and Write back with BBU MegaRAID Battery (BBU) Write Policies for the B440 server.
Workaround
Symptom C210 and C250 rack servers are missing RAID 50 and RAID 60 support.
Workaround
Symptom UCS Manager fails to create a single disk striped RAID config in the Storage Controller 1064E environment.
Workaround
Release 1.2(1)
The following caveats were opened in Release 1.2(1):
Red Hat Linux
Symptom Loading multiple driver disks during a RHEL 5.x installation fails.
Workaround
BIOS
Symptom Hubs that only use USB 1.0 may not properly present an attached USB device to the UCS server.
Workaround
Cisco UCS Manager
Symptom The PCIe Address for the Cisco UCS M81KR Virtual Interface Card is not seen in the GUI (or CLI). It causes no functional impact.
Workaround
Symptom Modification of trusted CoS policy in Service Profile does not get immediately applied to the server. If you modify the trusted CoS policy of an adapter profile in a service profile that is currently attached to a physical server, a server reboot is needed. Since it is unsafe to automatically reboot an associated server, UCSM currently does not.
Workaround
Symptom For each Cisco UCS 82598KR-CI 10 Gigabit Ethernet Adapter, 2 interfaces show up in the OS and ethtool reports Link Detected = yes for both of them. This is only seen on Cisco UCS B250 servers.
Workaround
Cisco UCS Manager GUI
Symptom When upgrading from releases prior to 1.1.1, OS-specific default adapter policies will not have the current recommended default values.
Workaround
Eth VMWare->RSS: DisabledEth VMWarePassThru->RSS: EnabledEth default->RSS: EnabledFC (all)->FCP Error Recovery: DisabledFC (all)->Flogi Retries: 8FC (all)->Flogi Timeout: 4000FC (all)->Plogi Timeout: 20000FC (all)->IO Throttle Count: 16FC (all)->Max LUNs Per Target: 256(CSCte58155)
Symptom Logon access is denied for user accounts where the password field was left blank during user account creation.
Workaround
Release 1.1(1)
This section lists the open caveats in release 1.1(1j).
BIOS
Symptom With the B-250 blade server, the displayed ESX and Linux OS HDD Boot Device Order is the reverse of the BIOS HDD Boot Order.
Workaround
Symptom FSM gets stuck in an Error Configuring the Local Disk Controller state due to various underlying conditions. Those can include but are not limited to the following:
•
•
Workaround
Symptom The Disk Fault/Error Codes, Disk Status, Alarms and the failures forwarded by the SAS Controller are not received by Cisco UCS Manager.
Workaround
Symptom After resetting the CMOS the system date needs to be reset to current.
Workaround
Red Hat Linux
Symptom When a vNIC is not in failover mode and a link down event occurs, the network traffic on the blades is disrupted with a system running RHEL 5.3.
Workaround
1.
service network restartor
ifdown ethxifup ethx2.
•
•
•
Note that network performance may be affected since the system is running in legacy mode. (CSCte44548)
Cisco UCS Manager
Symptom After the removal or insertion of one or more local disks, their full discovery fails.
Workaround
Symptom For a given port profile with existing VIFs, if the "Max-Ports" setting is reduced from the currently configured value to a value less than the "Used-Ports" value reported for that port profile by VMware vCenter, this is a mis-configuration. The new value for "Max-Ports" for that port profile will only be updated in Cisco UCS Manager and its update in VMware Center will fail, causing a inconsistency between Cisco UCS Manager and VMware Center Server.
Workaround
Cisco UCS Manager GUI
Symptom When a cluster configuration is set up such that I/O module 1 goes to fabric interconnect B and I/O module 2 goes to fabric interconnect A, then the Ethernet devices are given ports 1 and 0. However if the setup is straight, with I/O Module 1 connected to fabric interconnect A and I/O Module 2 to fabric interconnect B, then the devices are assigned ports 0 and 1.
Workaround
UCS Manager CLI
Symptom The UUID of the VM changes in VMware vCenter. After a VM restarts, the virtual machine node on the VM tab shows multiple instances of the same VM with one online and one offline.
Workaround
Release 1.0(2)
The following caveats were opened in UCS software release 1.0(2) and are still unresolved.
BIOS
Symptom With various Local Disk Configurations, the LSI SAS Configuration Utility fails to launch while in BIOS.
Workaround
Symptom When the memory mirroring configuration is destroyed by removing a DIMM, the BIOS will switch to the Performance mode, and will not log a message that mirroring was disabled.
Workaround
Symptom When plugging or removing USB devices at BIOS Setup -> Advanced -> USB, the Setup Utility may hang.
Workaround
Fabric Interconnect
HTTP
Symptom HTTPD process crashed, with the following event log:
Process crashed. Core file 1253640662_SAM_ucs-6120-1-A_httpd_log.3114.tar.gz (SAM/Switch Core Dump) detected on fabric interconnect A
Workaround
UCS Manager GUI
Symptom When waking up from sleep, the Cisco UCS Manager GUI will detect an event sequencing error and display the error: "Event Sequencing is skewed" because the JRE doesn't have a sleep detection mechanism.
Workaround
Symptom Downloads may be slow if TFTP is used.
Workaround
Release 1.0(1)
The following caveats were opened in UCS software release 1.0(1e) and are still present.
AAA
Symptom Local user passwords can not contain "$" character.
Workaround
Adapters
Symptom When a service profile containing two vNICs and having failover enabled is applied to QLogic or Emulex CNAs, the failback timeout specified in the adapter policy for the second vNIC has no effect. The failback timeout specified in the adapter policy and applied to the first vNIC is applied to the whole adapter and is effective for both vNICs.
Workaround
BIOS
Symptom Installing EFI Native SLES 11 is currently not supported.
Workaround
Symptom One vNIC defined in the Cisco UCS Manager service profile boot order results in two BIOS vNICs.
Workaround
Fabric Interconnect
Symptom When a fabric interconnect boots, the "The startup-config won't be used until the next reboot" message appears on the console. Fabric interconnect configuration is controlled by the UCS Manager, so this message has no meaning on the fabric interconnect configuration and has no functional impact.
Workaround
Symptom Console logon user names on the fabric interconnect are not case sensitive. For example, there is no differentiation between admin and ADMIN.
Workaround
Symptom When the system is under high stress, with repeated port flapping (ports rapidly going up and down) and default (native) VLAN change, the FWM process may core and cause the fabric interconnect to reload.
Workaround
Symptom The show cdp neighbor CLI command does not display information for CDP neighbors seen from the management interface, nor does it display the fabric interconnect CDP information corresponding to the management interface.
Workaround
Faults and Alerts
Symptom In rare cases the Cisco UCS Manager reports the link absence fault between the fabric interconnect server port and the fabric extender during the internal inventory collection. The following is an example of such a fault:
***************************Severity: ClearedCode: F0367Last Transition Time: 2009-07-15T11:47:49ID: 646445Status: NoneDescription: No link between fabric extender port 2/1/1 and switch A:1/9Affected Object: sys/chassis-2/slot-1/fabric/port-1Name: Ether Switch Intfio Satellite Connection Absent Cause: Satellite Connection AbsentType: ConnectivityAcknowledged: NoOccurences: 1Creation Time: 2009-07-15T11:46:49Original Severity: MajorPrevious Severity: MajorHighest Severity: Major*******************************Workaround
Inventory
Symptom Hardware revision numbers for fabric interconnect components are not populated in the Cisco UCS Manager.
Workaround
1.
2.
Server
Symptom The disk scrub policy needs enhancements to meet DOD compliance.
Workaround
SNMP
Symptom SNMP shows the fabric interconnect name rather than system name.
Workaround
Symptom An SNMP user name can not be the same as a local user name.
Workaround
SMASH
Symptom Any SMASH command entered with wrong option should give "INVALID OPTION" error message.
Workaround
UCS Manager CLI
Symptom Statistics counters cannot be cleared using the Cisco UCS Manager CLI.
Workaround
UCS Manager GUI
Symptom When several KVM Consoles are launched, the SUN JRE sometimes reports an error and the KVM Console fails to launch.
Workaround
Symptom In the Cisco UCS Manager GUI, if the Reboot on boot Order Change checkbox is checked for a boot policy, and if CD-ROM or Floppy is the last device in the boot order, then deleting or adding the device does not directly affect the boot order and the server does not reboot.
Workaround
Symptom Fibre Channel port and server port events do not appear on the Fibre Channel port and server port Events tabs.
Workaround
New Hardware Features in Release 2.0
Release 2.0(2m) adds support for:
•
•
•
•
•
•
•
•
Release 2.0(1w) adds support for:
•
Release 2.0(1s) adds support for:
•
Release 2.0(1m) adds support for:
•
•
•
New Software Features in Release 2.0
Release 2.0(1m) adds support for:
•
•
•
•
•
•
•
•
•
•
Related Documentation
For related documentation, see this document:
•
Obtaining Documentation and Submitting a Service Request
For information on obtaining documentation, submitting a service request, and gathering additional information, see the monthly What's New in Cisco Product Documentation, which also lists all new and revised Cisco technical documentation, at:
http://www.cisco.com/en/US/docs/general/whatsnew/whatsnew.html
Subscribe to the What's New in Cisco Product Documentation as an RSS feed and set content to be delivered directly to your desktop using a reader application. The RSS feeds are a free service. Cisco currently supports RSS Version 2.0.
This document is to be used in conjunction with the documents listed in the Cisco UCS Documentation Roadmap.
CCVP, the Cisco logo, and Welcome to the Human Network are trademarks of Cisco Systems, Inc.; Changing the Way We Work, Live, Play, and Learn is a service mark of Cisco Systems, Inc.; and Access Registrar, Aironet, Catalyst, CCDA, CCDP, CCIE, CCIP, CCNA, CCNP, CCSP, Cisco, the Cisco Certified Internetwork Expert logo, Cisco IOS, Cisco Press, Cisco Systems, Cisco Systems Capital, the Cisco Systems logo, Cisco Unity, Enterprise/Solver, EtherChannel, EtherFast, EtherSwitch, Fast Step, Follow Me Browsing, FormShare, GigaDrive, HomeLink, Internet Quotient, IOS, iPhone, IP/TV, iQ Expertise, the iQ logo, iQ Net Readiness Scorecard, iQuick Study, LightStream, Linksys, MeetingPlace, MGX, Networkers, Networking Academy, Network Registrar, PIX, ProConnect, ScriptShare, SMARTnet, StackWise, The Fastest Way to Increase Your Internet Quotient, and TransPath are registered trademarks of Cisco Systems, Inc. and/or its affiliates in the United States and certain other countries.
All other trademarks mentioned in this document or Website 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.
© 2011-2012 Cisco Systems, Inc. All rights reserved.
