Table Of Contents
Release Notes for Cisco VFrame IB Director, Release 3.1.1
Obtaining Platform Information
Cisco Server Fabric Switch Configuration
Cisco Server Fabric Switch Serial Number
Determining the Software Version
VFrame IB Director System Requirements
VFrame IB Director Installation
VFrame IB Scalability Guidelines
Resolved Caveats and Enhancements
Open Caveats in Cisco VFrame IB Director Release 3.1.1 (Build21)
Obtaining Documentation, Obtaining Support, and Security Guidelines
Release Notes for Cisco VFrame IB Director, Release 3.1.1
September 4, 2007These release notes describe the features and known limitations for the Cisco VFrame IB Director, Release 3.1.1, Build21.
Contents
This release note contains the following sections:
•
VFrame IB Director System Requirements
•
•
•
System Requirements
This section describes the system requirements for this software release.
Obtaining Platform Information
To determine platform information and server configuration information, use the information below.
Server Configuration
Type of server, chip set, CPU, amount of RAM, # of nodes
Operating system version (output from "uname -a")
PCI bus devices (output from "lspci -v")
Attached storage devices (output from "/var/log/dmesg")
InfiniBand configuration (output from "/usr/local/topspin/bin/vstat")
InfiniBand HCA info (output from "/usr/local/topspin/sbin/hca_self_test")
Cisco Server Fabric Switch Configuration
Chassis model: Output from the "show running-status all" command
Cisco Server Fabric Switch Serial Number
The chassis serial number and corresponding bar code are provided on the serial number label, as shown below:
Model: 7000, 7008, 3001, 3012
SN: MXxxxxxxxxxx
This information can be found on the bottom of the chassis or the outside of the Cisco chassis box packaging. It can also be found in the output of the show backplane CLI command.
Determining the Software Version
If the VFrame IB Director is running, click on <Help> / <About>. The version and build number will be displayed.
If the VFrame IB Director is not running but is installed on the host, enter the following command at the prompt, to determine the version of the Cisco VFrame IB Director you are running:
$ rpm -qa | grep topspintopspin-vframe-3.1.1-20topspin-ib-mod-rhel3-2.4.21-15.ELsmp-3.2.0-103topspin-ib-rhel3-3.2.0-103Supporting Product Releases
Table 1 identifies the products that work with the VFrame IB Director to form the VFrame solution.
Table 1
Product Platform VFrame IB 3.1.1 Release
See the release notes of the specific products for relevant information on added features, known issues, and workarounds.
VFrame IB Director System Requirements
VFrame IB Director may only be run on a server that meets the following requirements:
•
•
•
•
•
•
VFrame IB Director Installation
See the "Getting Started" chapter in Cisco VFrame InfiniBand Director User Guide for installation instructions.
Feature Overview
This section describes the features of this release. See the Cisco VFrame InfiniBand Director User Guide for a comprehensive list of features.
The VFrame IB Director enables the virtualization of the data center by separating the personality of the server from the physical server. It can flexibly match "bare-metal" servers to various personalities thereby delivering on the value proposition of "utility" computing.
The VFrame IB Directors can be deployed in HA mode with active and standby Directors.The Directors are always in sync, by using the same database stored on the SAN.
The following major features were added to Release 3.1.1:
•
•
•
•
•
•
•
•
•
•
Platform Support
The following platforms are supported with the VFrame IB Release 3.1.1. Additional platforms may also work but have not been tested by Cisco. Contact Cisco at ask-svbu-pm@cisco.com for additional platform requirements.
The following operating system support for compute nodes are supported with Release 3.1.1:
•
(IA32, Opteron 32-bit and 64-bit, EMT64T)
•
(IA32, Opteron 32-bit and 64-bit, EMT64T)
•
(IA32, Opteron 32-bit and 64-bit )
•
(IA32, Opteron 32-bit)
•
(IA32, Opteron 32-bit, 64-bit)
•
Table 2 shows the server support for compute nodes that are supported with Release 3.1.1:
Table 2
Server Support for Compute Nodes
Storage Array Support:
•
•
•
•
Multipathing Software:
•
•
•
•
•
SAN Switch support:
•
•
•
Ethernet Switch support:
•
Lights out Power Management for compute nodes:
•
•
•
•
•
•
•
•
•
•
InfiniBand Host Channel Adapter (HCA) support:
•
•
VFrame IB configurations currently do not support more than 1 HCA on a managed server.
VFrame IB Scalability Guidelines
VFrame IB Release 3.1.1 is recommended for fabric sizes up to 128 physical servers. This is independent of the servers being single or dual connected. A maximum of 32 physical servers per Fibre Channel gateway and a maximum of 64 servers per Ethernet gateway is recommended.
New And Changed Information
This section includes the following topics:
•
•
Resolved Caveats and Enhancements
This section describes the bug fixes and enhancements implemented in Cisco VFrame IB Director Release 3.1.1:
VFrame Operations
•
Support for multiple management subnets. This new feature will allow users to configure one or more management subnets for the Director to use to manage Virtual Servers. Each subnet can be configured for a different IB Partition (PKey) to allow Virtual Servers within a Virtual Server Group to be full-members of the IB Partition while restricting communications of Virtual Servers between Virtual Server Groups. Prior to this enhancement the Director was limited to a single management subnet.
•
Virtual Server and physical server statistic collection and reporting. This new feature adds support to the Director to collect and report statistics for Virtual Server Groups, Virtual Servers, and physical servers.
•
Progress indicator for image creation and deployment jobs. This new feature displays a progress indicator to show the percentage completion of imaging related jobs in the Director.
•
Display Virtual Server Groups and Virtual Servers in alphabetic order. This new feature will sort the display of Virtual Server Groups and Virtual Servers in alphabetic order. Prior to this enhancement these items were displayed in the order in which they were created.
•
Integration with Opsware SAS software for patch management on managed servers. This new feature allows the Director and Opsware SAS applications to co-exist in a datacenter. When the SAS application performs patch management on a managed server SAS will notify the Director to avoid any contention.
•
SOAP based North Bound API (NBAPI) to support third-party management application integration. This new feature will allow users to use a third-party management application to query the Director for all configuration, state and status information. In this solution, the Director acts as a SOAP-Server while the third-party management application acts as a SOAP-Client.
•
•
VSG Summary page allows deletion of VS in bulk create state. If multiple VSs are being created in bulk they should not be allowed to be deleted from the VSG Summary page ("Virtual Server Groups" page) or the VS Summary Page (Virtual Server Groups->VS->Virtual Servers). The user must cancel the bulk VS creation wizard to delete them.
•
Some characters that can be used to define customization properties (e.g., VSG, VS and image) are not being saved properly. These characters are getting encoded incorrectly.
•
When doing a bulk unassign of VSs an incorrect ERROR message may be generated with a wrong physical server count.
The ERROR message generated looks like this:
[ERROR]: Error unassigning physical server from virtual server group RHEL4U3: Cannot unassign physical server from virtual server rhel4u3-vs8. At least 0 physical server(s) are required.
•
Target checkbox in storage pool member selection does not work in Microsoft Internet Explorer browser.
•
The virtual server "ts_customize" script can wipe out similarly named hosts.
•
The Director slows down when an e-mail notification host is misconfigured. If the "Health->e-mail Notification->SMTP Host" is inaccessible, the e-mail logging takes a long time to timeout causing the Director to appear unresponsive.
•
Do not show warnings when storage configuration is asymmetric across SFS. If all ITLs are not available on all SFS switches a VS is configured to use, the "Storage is configured incorrectly" warnings will print during system synchronization.
•
•
User should not be able to select two HCAs while creating a new VSG. Only one HCA is supported at this time.
•
A Director's shared data LUN with device ID 0 is incorrectly displayed as a Boot LUN.
•
The Director UI is incorrectly allowing the selection of the physical server, used by the Director, to be assigned to a VS.
•
A VS may get stuck in the "Exiting Diagnostic" state.
•
A Java Exception is encountered when trying to generate a Diagnostic Report when a VS is in the Diagnostic state.
•
Upgrade the Director to ship with Sun Microsystem's Java Runtime Environment (JRE) version 1.4.2_14 to support the new Daylight Saving Time (DST) standard in the US. http://java.sun.com/developer/technicalArticles/Intl/USDST/
•
The Director's database restore operation does not work when incorrect inputs are entered. The operation fails silently without informing the user a problem has happened.
•
Unable to start the VFrame IB Director after doing a clean install and restoring a backed-up database when more than one web browser is used.
•
Automatically create management interfaces on the VFrame IB Director servers. Prior to this enhancement users had to manually execute commands to create and configure the management interface.
•
Fabric changes are not reflected in the VFrame IB Director if a virtual server is in the diagnostic mode.
•
Failed diagnostic operation results in failed virtual server state.
•
A SNMP policy configured with condition equal to "average value for all virtual servers" or "value for all virtual servers" can result in an ERROR when the policy is triggered.
•
The system synchronization operation is not updating trap receiver information in the SFS switch.
•
When a system synchronization operation is initiated by a SFS switch reboot the operation can hang on rare occasions. The Director will automatically start a system synchronization operation when it receives a SNMP cold-start or warm-start trap from a SFS switch. If the SFS switch reboots again while system synchronization is in-progress the synchronization thread can hang.
•
If a SNMP query for a chassis's version information fails when the Director is performing a system synchronization operation the chassis may be placed in the "Unmanaged" state.
•
Adjust the Director's initialization and failover settings to better provide HA support. Increased the time given to the HTTP Server to start-up. Increased the heartbeat interval between the Primary and Standby Directors.
•
Critical HA properties missing from the "ts_ha.properties" file can result in both Directors going out-of-service.
Storage
•
Support for initializing multiple LUNs at one time. This new feature will allow users to select one or more LUNs to initialize in one job. Previously, only one LUN can be initialized at a time.
•
Display the size of discovered LUNs. This new feature will display LUN sizes in GB/MB/KB resolution.
•
Display discovered LUNs based on the Target View (old format) or LUN View (new format). This new feature will allow users to select how LUNs are displayed in the Director's Storage page.
•
LUN description changes do not persist across chassis reboot. After a SFS chassis reboots the system synchronization operation is not updating the LUN descriptions in the switch.
•
When the Director's shared data LUN (used for VFrame IB Director HA deployments) is configured with four paths the "Multipathing Status" field in the "VFrame Directors->High Availability" page can display an incorrect status.
•
The "Multipathing Status" field in the "VFrame Directors->High Availability" page can display garbage characters when the user enters incorrect information in the form.
•
Automatically run the "fsck -p" command on the Director's shared data LUN before the LUN is mounted to fix potential disk problems. The command is execute each time the Director application starts-up and during a Director failover operation.
General
•
•
•
Open Caveats in Cisco VFrame IB Director Release 3.1.1 (Build21)
This section describes the open caveats with VFrame Release 3.1.1 (Build21). The Cisco defect tracking number, if applicable, is available for tracking.
•
VFrame IB Director GUI might have errors if accessed immediately after VFrame starts. Workaround is to wait 10 seconds after VFrame IB Director has started to access the GUI.
•
Windows may reboot when executing a netrpc shutdown command issued from VFrame. VFrame will power down the server during the system POST.
•
Assigning drive tray to or using removable USB drive with the VFrame IB Director in an IBM Bladecenter may cause imaging operations to hang. The workaround is not to use these devices with the blade running VFrame IB Director or to reboot the Bladecenter after the use of these devices.
•
VFrame IB Director does not restrict multiple users from making changes simultaneously.
•
Failed chassis components may not be visible in the VFrame IB Director "Health" screen. Workaround is to reconcile with Element Manager.
•
VFrame IB Director does not report on database LUN removal if the Target is still accessible. While active director does not have LUN access standby may not detect its failed state. Workaround is manually failover to the standby director.
•
•
VFrame IB Director does not time-out if user defined power control script does not return. Workaround is to restart Director or correct script operation. (
•
Hostnames for windows virtual servers must match Windows conventions for the computer name. For example, the DNS suffix should not be in the Director host-name field.
•
VFrame IB Director gets into a bad state if the application is started in less than 30 seconds after installation. Workaround is to uninstall and re-install VFrame IB Director.
•
Problems copying file onto Standby VFrame IB Director server if the RHEL OS is reinstalled. The problem has to do with a stale SSH key entry for the Standby Director in the "/root/.ssh/known_hosts" file on the Primary Director. Workaround is to manually delete the stale entry on the Primary Director.
•
Initializing many LUNs or deploying images to many Virtual Server LUNs (>20) at the same time can cause the Director to failover and hang in a Director HA configuration. The problem has to do with the Standby Director losing heartbeat with the Primary Director when the Primary Director is busy accessing the LUNs. Workarounds are: (a) select a smaller number of LUNs to initialize or deploy at one time, (b) shut down the standby director during large deploys, (c) use a higher capacity server to host the Directors (e.g., 2 CPUs, 3GB+ of memory).
•
Canceling a Virtual Server creation operation in the VS Creation Wizard can fail. Workaround is to allow the VS creation to complete and then delete the VS afterwards.
•
GUI page for assigning and unassigning a Virtual Server is not auto-refreshing. The user must manually click the "Refresh" button to get updated state information.
•
Explicitly causing a Director fail-over (by clicking the "Fail-over Now" button in the "VFrameDirectors->HighAvailability" page) when the Director is performing an imaging job can cause the Standby Director to not start-up properly. This problem does not happen if a failover is initiated automatically by the Director. Workaround is to not initiate a failover when an imaging job is in-progress.
Related Documentation
You can download the latest Server Networking and Virtualization documentation at http://www.cisco.com/en/US/products/ps6418/tsd_products_support_category_home.html
For more information about Cisco products related to VFrame Director, see the following documents:
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Obtaining Documentation, Obtaining Support, and Security Guidelines
For information on obtaining documentation, obtaining support, providing documentation feedback, security guidelines, and also recommended aliases and general Cisco documents, 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
Any Internet Protocol (IP) addresses used in this document are not intended to be actual addresses. Any examples, command display output, and figures included in the document are shown for illustrative purposes only. Any use of actual IP addresses in illustrative content is unintentional and coincidental.
© 2007 Cisco Systems, Inc. All rights reserved.
