brocade fabric os v5.2 - fujitsu.com pki certificates ... the section also includes the updated...

Brocade Fabric OS v5.2.1 Release Notes v1.0 of 110

Brocade Fabric OS v5.2.1 Release Notes v1.0

February 9, 2007

Document History

Document Title Summary of Changes Publication Date

Brocade Fabric OS v5.2.1 Release Notes v1.0 First Release February 9, 2007


Copyright © 2001 - 2007 Brocade Communications Systems, Inc. All Rights Reserved.

Brocade, the Brocade B weave logo, Fabric OS, File Lifecycle Manager, MyView, Secure Fabric OS, SilkWorm, and StorageX are registered trademarks and the Brocade B wing logo and Tapestry are trademarks of Brocade Communications Systems, Inc., in the United States and/or in other countries. FICON is a registered trademark of IBM Corporation in the U.S. and other countries. All other brands, products, or service names are or may be trademarks or service marks of, and are used to identify, products or services of their respective owners.

Notice: The information in this document is provided “AS IS,” without warranty of any kind, including, without limitation, any implied warranty of merchantability, noninfringement or fitness for a particular purpose. Disclosure of information in this material in no way grants a recipient any rights under Brocade's patents, copyrights, trade secrets or other intellectual property rights. Brocade reserves the right to make changes to this document at any time, without notice, and assumes no responsibility for its use.

The authors and Brocade Communications Systems, Inc. shall have no liability or responsibility to any person or entity with respect to any loss, cost, liability, or damages arising from the information contained in this book or the computer programs that accompany it.

Notice: The product described by this document may contain “open source” software covered by the GNU General Public License or other open source license agreements. To find-out which open source software is included in Brocade products, view the licensing terms applicable to the open source software, and obtain a copy of the programming source code, please visit http://www.brocade.com/support/oscd.

Export of technical data contained in this document may require an export license from the United States Government.


Contents Document History.................................................................................................................................................................1

QuickLook for Fabric OS v5.2.1...........................................................................................................................5 Fabric OS v5.2.1 Overview ...................................................................................................................................5

Hardware ..............................................................................................................................................................................5 The Brocade 5000 Switch................................................................................................................................................ 5

Software................................................................................................................................................................................6 New Features in Fabric OS v5.2.1........................................................................................................................................6

Access Gateway............................................................................................................................................................... 6 Dynamic Ports on Demand.............................................................................................................................................. 6

New Features in Fabric OS v5.2.0........................................................................................................................................7 RAS (Reliability, Availability, Serviceability) ................................................................................................................ 7 Management .................................................................................................................................................................... 7 Security............................................................................................................................................................................ 7 Other ................................................................................................................................................................................ 8

Optionally Licensed Software ..............................................................................................................................................9 Licensed Software as Standard.............................................................................................................................................9 Supported Switches ..............................................................................................................................................................9 Standards Compliance ..........................................................................................................................................................9 Technical Support...............................................................................................................................................................10

Important Notes...................................................................................................................................................11 Fabric OS Compatibility.....................................................................................................................................................11 Firmware Upgrades and Downgrades.................................................................................................................................12 Fabric Scalability................................................................................................................................................................13 FICON Support...................................................................................................................................................................14 Brocade PKI Certificates ....................................................................................................................................................14 Fabric OS............................................................................................................................................................................14

Diagnostics backport test ............................................................................................................................................... 14 Diagnostics spinsilk Test ............................................................................................................................................... 14 Others ............................................................................................................................................................................ 15

RFEs Implemented in Fabric OS v5.2.1 ............................................................................................................21 RFEs Implemented in Fabric OS v5.2.0 ............................................................................................................22 Fabric OS v5.2.1 Documentation.......................................................................................................................22

New Hardware Documentation ..........................................................................................................................................22 Brocade 5000 Hardware Reference Manual (Publication number: 53-1000424-01)..................................................... 22 Brocade 5000 QuickStart Guide (Publication number: 53-1000425-01)....................................................................... 22 Brocade 5000 Power Supply/Fan Assembly Replacement Procedure (Publication number: 53-1000426-01) ............. 22 Brocade 5000 Rack Mounting Ears Installation Procedure (Publication number: 53-1000451-01).............................. 23


Updated Software Documentation......................................................................................................................................23 Brocade Fabric OS V5.2.X Software Addendum (Publication number: 53-1000429-01)............................................. 23

New Software Documentation............................................................................................................................................23 Brocade Fabric OS V5.2.1 Access Gateway Administrator’s Guide (Publication number: 53-1000430-01) ............... 23

Documentation Updates.....................................................................................................................................23 Open Defects In Fabric OS v5.2.1......................................................................................................................25

Deferred Defects Found in a Release Other than Fabric OS v5.2.1....................................................................................25 Deferred Defects Found in Fabric OS v5.2.1 .....................................................................................................................62

Resolved Defects in Fabric OS v5.2.1...............................................................................................................72


QuickLook for Fabric OS v5.2.1 If you are already using the most recent version of the Fabric OS v5.2.0b Release Notes, here are the changes in this version of the Release Notes.

• The Fabric OS v5.2.1 Overview lists the new features in Fabric OS v5.2.1, and the New Features in Fabric OS v5.2.1 section describes those features.

• The Important Notes section contains a new update regarding the diagclearerror command on certain platforms. The section also includes the updated Fabric OS Compatibility Matrix.

• The RFEs Implemented in Fabric OS v5.2.1 section lists the Requests for Enhancement included in Fabric OS v5.2.1.

• The Documentation section lists and describes the new Fabric OS v5.2.1 documentation.

• The Documentation Updates section includes recent additions or corrections to the Fabric OS v5.2.0 documentation.

• After the Fabric OS v5.2.1 Overview, the rest of the front matter of this Fabric OS v5.2.1 release note will consist of the front matter from the previous release note, Fabric OS v5.2.0b, with the following update:

Two new Important Notes regarding the Access Gateway feature.

The Evmd (event monitor daemon) entry was removed from the bulleted list of restart/monitoring daemons in the RAS (Reliability, Availability, Serviceability) section.

Fabric OS v5.2.1 Overview

Hardware Fabric OS v5.2.1 supports the merge of the Fabric OS 5.0.x software features with those of the Fabric OS v5.2.x software features. This means that in addition to the switches already supported by the Fabric OS v5.2.x software, the following embedded switches are supported in Fabric OS v5.2.1:

o SilkWorm 3014

o SilkWorm 3016

o SilkWorm 4012

o SilkWorm 4016

o SilkWorm 4018

o SilkWorm 4020

o SilkWorm 4024

In addition, Fabric OS v5.2.1 supports the new Brocade 5000 switch.

The Brocade 5000 Switch The Brocade 5000 switch is a 32-port autosensing 4/2/1 Gbit/sec Fibre Channel switch that includes advanced interoperability capability for McDATA fabrics, enhanced operating efficiency and optimized system design. The Brocade 5000 is targeted for use as a standalone switch in small SANs, or as an edge switch in larger SAN environments.


Software

Fabric OS v5.2.1 supports a new feature, Access Gateway only on the following embedded switches: o SilkWorm 4012

o SilkWorm 4016

o SilkWorm 4020

o SilkWorm 4024

NPIV functionality has been available on the Fabric OS v5.1.x and 5.2.x platforms and is also supported in Fabric OS v5.2.1 on embedded blade server SAN switches including the SilkWorm 3014, 3016, 4012, 4016, 4018, 4020, and 4024. No NPIV license is required.

Web Tools in Fabric OS v5.2.1 supports the enabling and disabling of Access Gateway, as well as firmware download and various monitoring functions. There is also SNMP support for Access Gateway. Fabric Manager 5.2.0 does not support the Access Gateway feature. Both Web Tools and Fabric Manager 5.2.0 support the new Brocade 5000 switch.

Fabric OS v5.2.1 supports Dynamic Ports On Demand (DPOD) on the following embedded switches: o SilkWorm 4016

o SilkWorm 4018

o SilkWorm 4020

o SilkWorm 4024

In addition, Fabric OS v5.2.1 includes fixes for various FOS defects and various Requests for Enhancement (RFEs).

New Features in Fabric OS v5.2.1

Access Gateway

Access Gateway allows a switch to operate in a special ‘agmode’ that allows simplified connectivity between large numbers of servers and the SAN. Access Gateway leverages NPIV (N_port ID virtualization) to hide the complexity of the servers (both physical and virtual) attached to it while allowing easy SAN connectivity. The edge fabric switch provides all the fabric services while Access Gateway connects to the edge switch by what appears as an HBA connection. This architecture allows the deployment of many additional servers without requiring a domain and the associated fabric rebuild traffic that is prevalent in dynamic blade server environments. On Fabric OS v5.2.1 Access Gateway is available on the SilkWorm 4012, 4016, 4020, and 4024.

Dynamic Ports on Demand Dynamic Ports On Demand (DPOD) is an optional feature on selected embedded switches. DPOD takes the expansion capability of fixed Ports On Demand (POD) and adds the flexibility of connecting to any available port as long as a valid license is available. Previously, POD allowed only specific fixed ports to be utilized. With DPOD, any physically available port can be made active as long as a valid license is available. This allows customers the flexibility of automatically changing port assignments where previously the port assignments were fixed and inflexible. DPOD is tailored for dynamic environments such as blade server deployments and is available on the SilkWorm 4016, 4018, 4020, and 4024.


New Features in Fabric OS v5.2.0 Brocade Fabric OS v5.2.0 supports two new hardware blades for the SilkWorm 48000 director: SilkWorm FC4-48 Fibre Channel port blade and SilkWorm FC4-16IP iSCSI blade.

• The FC4-48 port blade offers 48 1-, 2-, and 4-Gbit/sec Fibre Channel ports for the SilkWorm 48000 director. Brocade continues to provide its customers with state-of-the-art scalability and a SAN enterprise solution with the industry’s lowest power consumption.

• The FC4-16IP iSCSI blade enables the SilkWorm 48000 director to provide iSCSI initiators to FC target connectivity. It features eight auto-sensing 1-, 2-, and 4-Gbit/sec Fibre Channel and eight 1 Gbit/sec Ethernet (1000Base-T) RJ-45 ports.

Fabric OS v5.2.1 supercedes Fabric OS v5.2.0. All users are strongly encouraged to upgrade to v5.2.1 as soon as they have access to it.

NOTE: Install Fabric OS v5.2.1 software before you install the new blade types (FC4-16IP or FC4-48).

New features in the Fabric OS v5.2.0 release are summarized in the following sections.

RAS (Reliability, Availability, Serviceability) • Audit logging provides logs per user-generated events, such as security violation, zoning, firmware

download, and configuration changes.

• Configuration management enhancements improve switch availability by allowing Fabric Watch and SNMP parameter changes to be non-disruptive.

• Firmware upgrade enhancements provide clearer error messages and remove the need to enter “release.plist” in the command line.

• Daemon restart/monitoring restarts management daemons automatically when they fail without switch reboot:

o Snmpd - simple network management protocol daemon

o Webd - web server daemon

o Cald - common access layer daemon

o Rpcd - remote procedure call daemon

o Arrd - asynchronous response router daemon (send management data to hosts when the switch is accessed via FA API or SMI-S).

o Trackd - track changes daemon

• Port Mirroring captures traffic between two devices for non-disruptive traffic analysis (available on the SilkWorm 4100, 4900, 48000, and 7500).

Management • Role-Based Access Control (RBAC) adds support for the new RBAC roles: Operator, Zone Manager, Fabric

Administrator, and Basic Switch Administrator.

• Virtual fabrics through administrative domains (Admin Domains or AD) provides data, management, and fault isolation through administrative domains.

• DHCP support for standalone switches.

Security • Device Connection Control (DCC), Switch Connection Control (SCC), and the ability to manually

distribute passwords among participating switches in the base Fabric OS.

• Internet Protocol Security (IPSec) ensures private, secure communications over Internet Protocol (IP) networks to prevent network-based attacks, which could potentially result in denial of service, data


corruption, data theft, user credential theft, and so on. IPSec will be available as a standard license for the SilkWorm 7500 and FR4-18i blade in the SilkWorm 48000 director.

Other • FCR enhancements for the SilkWorm 7500 and FR4-18i blade in the SilkWorm 48000 director:

o Front domain consolidation providing one front domain per chassis projected to edge fabrics regardless of the number of EX_ports connected from the SilkWorm 48000 or FR4-18i blade in the SilkWorm 48000 to that edge fabric.

o McDATA interoperability in both McDATA Fabric and Open Fabric modes

o EX_port trunking providing high bandwidth across the router

o Router port cost providing users flexibility to determine the preferred route between two destinations across a metaSAN

• FCIP enhancements for the SilkWorm 7500 and FR4-18i blade in the SilkWorm 48000 director:

o Internet Protocol Security (IPSec) ensures private, secure communications over Internet Protocol (IP) networks to prevent network-based attacks, which could potentially result in denial of service, data corruption, data theft, user credential theft, and so on.

o Fastwrite reduces the number of round-trips required to complete a SCSI Write IO, which both reduces IO completion latency and increases FCIP ISL bandwidth utilization.

o Tape Pipelining accelerates SCSI Write IOs between geographically remote initiators and tape devices on Fibre Channel SANs linked via FCIP ISLs.

o WAN tool, the ipperf option has been added to the portCmd command to characterize end-to-end IP path performance factors, such as bandwidth, loss rate, roundtrip time, and path MTU (Maximum Transmission Unit) between a pair of Brocade FCIP ports.

• The tstimezone command provides an interactive interface to select Daylight Savings Time based on the country and region.

• The number of user accounts is increased from 15 to 256.

• Zoning database size increased from 256 KB to 1 MB.

• Long distance mode simplification:

o LD is a dynamic distance mode that automatically discovers lengths and assigns the correct amount of buffer credits with an Extended Fabrics license.

o LS is a static distance mode that allows you to specify the number of buffer credits required with an Extended Fabrics license.

o LE supports up to 10 kilometers at any speed and does not require an Extended Fabrics license.


Optionally Licensed Software This Fabric OS release includes all basic switch and fabric support software, as well as the following optionally licensed software, which is enabled via license keys:

• Brocade Extended Fabrics—Up to 500 km of switched fabric connectivity at full bandwidth over long distances

• Brocade ISL Trunking Over Extended Fabrics—Enhanced to enable trunking over long-distance links of up to 250 km

• Brocade Fabric Manager—Administration, configuration, and maintenance of fabric switches and SANs with host-based software

• Brocade Advanced Performance Monitoring—Performance monitoring of networked storage resources

• Brocade Fabric Watch—Monitoring of mission-critical switch operations

• FC-IP—Fibre Channel over IP extension includes FC-IP trunking, multi-tunnel support, and compression

Licensed Software as Standard The following licensed software is available with the hardware and no additional purchase is necessary:

• Brocade Web Tools—Administration, configuration, and maintenance of fabric switches and SANs

• Brocade Advanced Zoning—Division of a fabric into virtual private SANs

• IPSec—IP Security (for the SilkWorm 7500 and FR4-18i blade in the SilkWorm 48000)

Supported Switches Fabric OS v5.2.0 adds support for the FC4-48 and FC4-16IP blades for the SilkWorm 48000 director. It also supports the SilkWorm 200E, 3250, 3850, 3900, 4100, 4900, and the SilkWorm 7500, 24000, and 48000, and the Brocade 5000, new with Fabric OS v5.2.1.

IMPORTANT: The SilkWorm 12000 is not supported in this release; defect fixes for this platform will be delivered on the Fabric OS v5.0.x releases.

Standards Compliance This software conforms to the Fibre Channel Standards in a manner consistent with accepted engineering practices and procedures. In certain cases, Brocade might add proprietary supplemental functions to those specified in the standards. For a list of standards conformance, visit the following Brocade Web site: http://www.brocade.com/sanstandards


Technical Support Contact your switch supplier for hardware, firmware, and software support, including product repairs and part ordering. To expedite your call, have the following information immediately available:

1. General Information

• Technical Support contract number, if applicable • Switch model • Switch operating system version • Error numbers and messages received • supportSave command output • Detailed description of the problem and specific questions • Description of any troubleshooting steps already performed and results

2. Switch Serial Number

The switch serial number and corresponding bar code are provided on the serial number label, as shown here.

*FT00X0054E9 FT00X0054E9

The serial number label is located as follows:

• SilkWorm 3014—Top of the chassis, under the insertion arm

• SilkWorm 3016, 4012, and 4024 —Bottom of the switch module

• SilkWorm 4016 and 4018— Top of the switch module

• SilkWorm 4020—Side of the switch module

• SilkWorm 200E, 3200, 3250, and 3850—Bottom of the chassis

• SilkWorm 3800 and 3900—Nonport side of the chassis

• Brocade 5000, SilkWorm 4100, 4900, and 7500—On the switch ID pull-out tab located inside the chassis on the port side on the left

• SilkWorm 12000, 24000, and 48000—Inside the chassis next to the power supply bays

• SilkWorm Multiprotocol Router Model AP7420—Bottom of the chassis and on the back of the chassis.

3. World Wide Name (WWN)

• Brocade 5000, SilkWorm 200E, 3014, 3016, 3250, 3850, 3900, 4012, 4016, 4018, 4020, 4024, 4100, 4900, and 7500 switches and SilkWorm 12000, 24000, and 48000 directors—Provide the license ID. Use the licenseIDShow command to display the license ID.

• SilkWorm Multiprotocol Router Model AP7420—Provide the switch WWN. Use the switchShow command to display the switch WWN.

• All other SilkWorm switches—Provide the switch WWN. Use the wwn command to display the switch WWN.


Important Notes This section lists information you should consider before you use this firmware release.

Fabric OS Compatibility The following table lists the earliest versions of Brocade software supported in this release, that is, the earliest supported software versions that interoperate. Brocade recommends using the latest software versions to get the greatest benefit from the SAN.

For a list of the effective end-of-life dates for all versions of Fabric OS, visit the following Brocade Web site:http://www.brocade.com/support/end_of_life.jsp Fabric OS Interoperability with Brocade Switches and Firmware Switches: SilkWorm 2000 Series and 6400 Fabric OS v2.6.2 1 Switches: SilkWorm 3000, 3200, 3600, and 3800 Fabric OS v3.2.X Embedded Switch: SilkWorm 4012 Fabric OS v5.0.0 Switches: SilkWorm 200E, 325x, 385x, 3900, and 4100 Embedded Switches: SilkWorm 3014, 3016, an 4012 Directors: SilkWorm 12000, 24000, and 48000 (without FR4-18i blade)

Fabric OS v5.0.1

Embedded Switch: SilkWorm 4020 Fabric OS v5.0.2 Switches: SilkWorm 200E, 325x, 385x, 3900, and 4100 Embedded Switches: SilkWorm 3014, 3016, 4012, and 4020 Directors: SilkWorm 12000, 24000, 48000 (without FR4-18i blade)

Fabric OS v5.0.3

Switches: SilkWorm 200E, 325x, 385x, 3900, and 4100 Embedded Switches: SilkWorm 3014, 3016, 4012, 4016, and 4020 Directors: SilkWorm 12000, 24000, and 48000 (without FR4-18i blade)

Fabric OS v5.0.4

Switches: SilkWorm 200E, 325x, 385x, 3900, and 4100 Embedded Switches: SilkWorm 3014, 3016, 4012, 4016, 4018, 4020, and 4024 Directors: SilkWorm 12000, 24000, and 48000 (without FR4-18i blade)

Fabric OS v5.0.5

Switches: SilkWorm 200E, 325x, 385x, 3900, 4100, 4900, and 7500 Directors: SilkWorm 24000 and 48000 (with or without FR4-18i blade) Router: SilkWorm 7500

Fabric OS v5.1.0

Switches: SilkWorm 200E, 325x, 385x, 3900, 4100, 4900, and 7500 Directors: SilkWorm 24000 and SilkWorm 48000 (any combination of

FC4-16, FC4-32, FC4-48, FC4-16IP, and FR4-18i blades) Router: SilkWorm 7500

Fabric OS v5.2.0 1

Switches: SilkWorm 200E, 3014, 3016, 325x, 385x, 3900, 4012, 4016, 4018, 4020, 4024, 4100, 4900, Brocade 5000 and SilkWorm7500 Directors: SilkWorm 24000 and SilkWorm 48000 (any combination of

FC4-16, FC4-32, FC4-48, FC4-16IP, and FR4-18i blades) Router: SilkWorm 7500

Fabric OS v5.2.1

Router: SilkWorm AP7420 XPath 7.4.x OS 2


Fabric OS Interoperability with McDATA Switches and Firmware 2

Intrepid 6140 and 6064 EOS v7.x, v8.x 2 Sphereon 3232, 4300, 4500 and 3216 EOS v7.x, v8.x 2

(1) Fabric OS v2.6.2 can interoperate with Fabric OS v5.2.0 through the FC routing capability of the SilkWorm AP7420, SilkWorm 7500, or FR4-18i blade in the SilkWorm 48000 director. Customers who wish to have Fabric OS v2.6.2 and v5.2.0 mixed in the same fabric should consult their equipment provider for a detailed list of limitations. New fabric-wide features introduced in Fabric OS v5.2.0, such as Virtual Fabrics, Access Control security policy, new hardware, etc., will not be compatible with Fabric OS v2.6.2.

(2) Fabric OS and McDATA E/OS v4.x, v5.x, 6.x can interoperate through the FC routing capability of the SilkWorm AP7420 only. Fabric OS and McDATA E/OS v7.x, 8.x can interoperate through the FC routing capability of the SilkWorm AP7420, SilkWorm 7500, or FR4-18i blade in the SilkWorm 48000.

Firmware Upgrades and Downgrades Brocade does not support upgrading from more than two previous releases. For example, upgrading from Fabric OS v5.0.x to v5.2.x is supported, but upgrading from Fabric OS v4.4.0 or a previous release directly to v5.2.x is not.

Upgrading a switch from Fabric OS v4.4.0 or a previous release to v5.2.0 requires a two-step process: first upgrade to v5.0.x or v5.1.x and then upgrade to v5.2.0.

In addition, the following conditions must be met before upgrading to v5.2.0:

• Device-based routing must not be in use, otherwise the upgrade will fail. You can use the aptPolicy command to verify the routing policy.

• Chassis configuration options 3 and 4 are no longer supported for the SilkWorm 48000; see the “SilkWorm 48000 Chassis Configuration Options” table for details.

Install the new blade types (FC4-16IP or FC4-48) only after you have installed the Fabric OS v5.2.0 software.

Brocade supports downgrading up to two previous releases, for example, if you upgrade to Fabric OS v5.2.0 from v5.0.x, you can revert back to v5.0.x. However, you cannot downgrade from Fabric OS v5.2.0 to v4.4.0 or to a previous release.

NOTE: If the SilkWorm 48000 has FC4-48 or FC4-16IP blades installed or any new software features in Fabric OS v5.2.0, such as FCR trunking and administrative domains for virtual fabrics, you cannot downgrade below Fabric OS v5.2.0. If you wish to do so, you must a) remove these features, b) physically remove the blade, and then c) downgrade firmware.

For any other new software features or increased scalability limits supported by Fabric OS v5.2.0, downgrade will be disruptive and requires cold reboot.

A v5.2.0 configuration file cannot be used on the same switch after the switch has been downgraded to firmware version v5.0.x or v5.1.x.

When you downgrade to Fabric OS v5.0.x, you will also need to remove additional v5.1.x features (and any installed FR4-18i blades). The firmwareDownload command will guide you to remove any features and blades that need to be removed.


Fabric Scalability Fabric OS v5.2.0 supports the same fabric scalability as Fabric OS v5.0.x and v5.1.x, that is, 2,560 ports with 50 domains.

For FC Routing environments, the following scalability numbers apply:

Fibre Channel Routing Scalability

Max # edge fabrics per metaSAN 32

Max # edge fabrics per chassis 16

Max # local switches per edge fabric 26

Max # front domains per edge fabric 10

Max # translate domains per edge fabric 33

Max # total domains per edge fabric 69

Max # local switches per backbone fabric 5

Max # translate domains per backbone fabric 33

Max # total domains per backbone fabric 69

Max # FCR switches per metaSAN 10

Max # local WWNs per edge fabric 1200

Max # local WWNs per backbone fabric 512

Max # imported devices per fabric 1000

Max # local & remote WWNs per fabric 1300

Max # device database entries per metaSAN 10000

Max # LSAN zones per metaSAN 2500 (with a v5.2.0 only FCR backbone)

Max # entries per LSAN zone 64

Max # hops between edge switches 12

Max # EX_Ports to an edge fabric from FCR 8(4G)

EX_ports per FCR 32


FICON Support With this release, the Switch Connection Control high integrity requirement for cascading FICON is available in the standard base Fabric OS. End users can now deploy new cascade FICON directors without purchasing a separate Secure Fabric OS license.

To add a new FICON director into existing cascaded configurations that are already running Secured Fabric OS, it is recommended that users continue to deploy Secure Fabric OS on the new FICON director instead of migrating to FOS ACL configuration.

NOTE: The FC4-48 Fibre Channel port blade is not supported to connect to System z environments via FICON channels or via FCP zLinux on System z. To attach the SilkWorm 48000 director to the System z environment, use an FC4-16 or FC4-32 Fibre Channel port blade.

Brocade PKI Certificates As of May 15, 2005, Brocade no longer includes a PKI Certificate as part of the installed Secure Fabric OS. If you wish to activate Secure Fabric OS on a supported director or switch, you must contact Brocade to obtain a PKI certificate.

Refer to the Secure Fabric OS Administrator's Guide, Chapter 2, ”Adding Secure Fabric OS to the Fabric,” for a description of how to obtain certificates from the Brocade Certificate Authority.

Fabric OS

Diagnostics backport test

The backport test passes only in a) a pure SilkWorm 24000 director or b) a SilkWorm 24000 system with no FC4-16 blades and under Option 5.

Do not run backport tests in any configuration other than the two listed above; use the minicycle test instead.

Diagnostics spinsilk Test

The following configurations will pass the spinsilk test:

• Pure SilkWorm 24000 director (only CP2 and FC-16 blades)

• Pure SilkWorm 48000 director, option 5

• Pure SilkWorm 48000 director, option 5 (with FC4-16 blades)

The following configurations will fail the spinsilk test; use the minicycle test instead:

• Mixed SilkWorm 24000 director (with either CP4 or FC4-16 blades)

• Pure SilkWorm 48000, option 1

“Pure SilkWorm 48000” refers to a director with CP4 and FC4-16 blades only.


Others

The following are known issues in this release of Fabric OS.

Area Description

Diagnostic Command on the SilkWorm 4016, 4018, and 4024

When run as a separate command from the CLI, diagclearerror might hang or cause the system to crash. This affects only the SilkWorm 4016, SilkWorm 4018, and SilkWorm 4024 platforms. To run diagclearerror from the CLI, first switch directories as shown here: cd /fabos/sbin diagclearerror

SilkWorm 7500 Silkworm 7500 fans operate at the correct speed, that is, at maximum on bootup. However, this initial speed may trigger an error message that indicates that the speed is too high (“above threshold”). You can disregard this message; the fan speed is adjusted to a nominal speed shortly after bootup. This message is benevolent. The fan speed will be adjusted to a nominal speed shortly after bootup.

SilkWorm 48000 • Customers upgrading SilkWorm 24000 switches from Fabric OS v5.0.5 to v5.2.0 while in chassisconfig option 3 or 4 will not get sufficient notification in the session error message or firmwaredownloadstatus command about how to correct the problem.

• Before moving the slider UP on a Control Processor blade that is being activated, observe that the amber LED is not ON for the active CP for at least 5 seconds and all LEDs are off on new inserted CP.

• In a core-edge design, when a fully populated 384-port SilkWorm 48000 (populated with 8 FC4-48 blades) is an edge switch in a large SAN, it can experience high CPU utilization and may panic if it becomes a principal switch. SAN design best practice recommends deploying a high port-count switch as both core and principal switch to reduce fabric stress and provide ease of management.


Area Description

FC4-48 port blade for the SilkWorm 48000

• configure command only gives a maximum login per port setting. The command allows over 127, where ports for the FC4-48 blade will honor that value as long as its share areas values are 127 or less.

• Before replacing an FC4-32 blade with an FC4-48 blade, restore ports 16 – 31 of the FC4-32 blade if these ports are used for port swapping. Failure to do so will fault the FC4-48 blade. The only way to restore back to original settings is to add the FC4-32 blade back into the slot and port swap the ports back to the ports’ default setting.

• FC4-48 ports should not belong to the zone or in an administrative domain in which FICON devices are present.

• FC4-48 blade does not support loop. Private L_Ports will be shown on these ports in switchShow, but will not participate in the fabric.

• The porttest and spinfab commands on any platform will not work on E_Ports connected to an FC4-48 port.

• The FC4-48 Fibre Channel port blade is not supported to connect to the System z environments via FICON channels or via FCP zLinux on System z. To use the SilkWorm 48000 director to attach to the System z environment, please use the FC4-16 or FC4-32 Fibre Channel port blades

• In a zoning configuration with members D and P, where "P" is greater than or equal to 256, remove these configurations before downgrading to a lower firmware version (5.1.x or 5.0.x). Otherwise, the downgrade will not be HA compatible with earlier versions.

• Do not insert the blade until the system is running Fabric OS 5.2.0.

FC4-16IP iSCSI blade for the SilkWorm 48000

• iSCSI virtual target creation involves adding LUNs to the virtual targets. The user discovers the LUNs by executing the fclunquery command. Testing has revealed that some devices do not respond properly to the LUN query. The user will need to use tools from the array vendor to determine LUN information required for iSCSI target creation.

• Any upper case letters used for the CHAP user name will be transformed to lower case.

• Users may install up to four FC4-16IP iSCSI blades per SilkWorm 48000 chassis, or any combination of up to four FC4-16IP and FR4-18i blades, not to exceed two FR4-18i blades per chassis. Some valid combinations are:

o three FC4-16IP blades + one FR4-18i blade

o two FC4-16IP blades + two FR4-18i blades

o one FC4-16IP blades + two FR4-18i blades

o two FC4-16IP blades + one FR4-18i blades

• Do not insert the blade until the system is running Fabric OS 5.2.0.


Area Description Firmware upgrade/downgrade

• When upgrading from Fabric OS v5.1.0x to v5.2.0x, if there are 2 or more inter-fabric links (IFLs) connected to an edge fabric, one IFL will stay online and the other IFLs will go online and offline. This will cause a temporary traffic disruption going from multiple IFLs to 1 IFL and then back to multiple IFLs. This is due to the new front domain consolidation feature in Fabric OS v5.2.0 where the IFLs connected to the same edge share the same front domain.

• When downgrading from Fabric OS v5.2.0 to v5.1.0x, FC traffic will be disruptive if there is front domain consolidation prior to the downgrade, even in the case of a single IFL.

• Upon firmware download the FC4-16IP blade does not preserve disabled GE_Ports in a disabled state. If you wish to retain GE_Ports in a disabled state across afirmware download, you must configure them as persistently disabled.

• In a large fabric with a large zoning database (e.g., 2560 ports with a 1MB zoning database), a non-disruptive firmwaredownload on a SilkWorm 3850 or 3900 can result in an E_port offline transition. The E_Port offline transition causes a fabric reconfiguration and can cause momentary frame loss. 4 gb/s switches do not experience this issue.

• FCIP traffic is disrupted in an upgrade from Fabric OS v5.2.0a to Fabric OS v5.2.0b.

Fabric OS – CLI commands

• This release does not support underscore (_) as part of the name for dd and ddset in the iscsicfg command.

• The slotOff and slowOn commands are now obsolete; use slotPowerOff and slotPowerOn instead. The portLogPortShow command is also now obsolete.

• The QuickLoop feature and related commands (listed below) are no longer supported on Fabric OS versions v5.1.0 and higher.

o qloopAdd

o qloopCreate

o qloopDelete

o qloopRemove

o qloopShow

Distance mode • Distance setting is not persistent. After a configuration uploads and downloads, distance settings will be lost and the desired distance will be shown as 0.

FC Routing • If a SilkWorm AP7420 is present in the backbone fabric, the command fcrDisable may take up to 8 minutes to complete. If the AP7420 is replaced by an FR4-18i or a SilkWorm 7500, the command completes immediately.

• EX_Port trunking is not enabled by default.

• Fabric OS v5.2.0 introduces the EX_Port trunking feature. This feature should only be enabled if the entire configuration is running Fabric OS v5.2.0 or later. Enabling the EX_port trunking feature on a switch running Fabric OS v5.2.0 or later in a configuration containing a Fabric OS v5.1.0 switch will cause the Fabric OS v5.1.0 switch to panic.

• When an unstable edge fabric that has multiple EX_Port connections is in a transitional state, on rare occasions one of the EX_Ports may detect an FID conflict and be disabled. If this occurs, manually re-enable the port.


Area Description Security Remove any password enforced expiration of admin or root accounts before

downgrading firmware to Fabric OS v5.0.1 or lower versions.

Diagnostics • All offline diagnostics commands should be used only when the switch is disabled.

• POST can fail if new SFPs are added during POST. SFPs should only be added while the switch is “online” or if the switch is powered off.

• When you use the diagnostic commands systemVerification and diagSetBurnin, the switch or blade will fault when the burn-in error log is full. Clear the burn-in log before running systemVerification or diagSetBurnin.

• If there are ISLs present on the switch that are not used for routing because they have higher link costs, disable the links before running spinfab.

HA If there is an already segmented port and backbone devices are exported to an edge fabric, a build fabric/fabric reconfiguration can occur after running haFailover. Ensure that there are no segmented ports before upgrading firmware.


Area Description IPSec for FR4-18i blade

• IPSec implementation details:

o Pre-shared key

o Main mode (IKE negotiation protocol)

o Tunnel mode in ESP (Encapsulating Security Payload)

• IPSec specific statistics not provided

• No NAT or IPV6 support

• FastWrite and Tape Pipelining will not be supported in conjunction with secure tunnels.

• Jumbo frames will not be supported on secure tunnels.

• ICMP redirect is not supported for IPSec-enabled tunnels.

• Only a single secure tunnel will be allowed on a port. Non-secure tunnels will not be allowed on the same port as secure tunnels.

• Modify operations are not allowed on secure tunnels. To change the configuration of a secure tunnel, you must first delete the tunnel and then recreate it with the desired options.

• Only a single route is supported on an interface with a secure tunnel.

• An IPSec tunnel cannot be created using the same local IP address if ipperf is active and using the same local IP address (source IP address).

• Unidirectional supported throughput is ~104Mbytes/sec and bidirectional supported throughput is ~90Mbytes/sec.

• An IPSec tunnel takes longer to come online than a non-IPSec tunnel.

• User is not informed with the IPSec mismatch RAS event when configuring a tunnel with IPSec mismatch on either end.

Fabric Merge Do not try to merge fabrics with conflicting domain IDs over a VE_Port. Before merging two fabrics over FC-IP with VE_Ports at each end, it is recommended that all domain ID and zoning conflicts are resolved.

Scalability • Support for Default Zoning policies has been added to Fabric OS v5.1.0. Typically, when you issue the cfgDisable command in a large fabric with thousands of devices, the name server indicates to all hosts that they can communicate with each other. To ensure that all devices in a fabric do not see each other during a cfgDisable operation, you can activate a Default Zone with policy set to “no access”. If Default zoning policies are enabled, all cfgEnable/Disable commands and zoning changes must be run from a switch in the fabric running Fabric OS v5.1.0/v5.2.0.

• In large fabrics with more than 1,000 ports, it is recommended that the MS Platform Database is disabled. It is also required that the Platform DB be disabled before downgrading to previous versions of Fabric OS. This can be done using the msPLMgmtDeactivate command.

FRU insertion The FW_FRU_INSERTED message is displayed twice when a power supply FRU is inserted and powered on. There is no functional impact.

System boot Not all Fabric OS services are available when the prompt becomes available during boot up. Wait for all the services to come up before using the switch or performing zoning actions.


Area Description Performance Monitoring

If the user tries to save more than 512 monitors using the perfCfgSave command, some of the monitors may be lost.

Management – Proxy switches

If you are using a Fabric OS v4.x switch as an API or SMI-S proxy to manage a v5.1.0 switch, you must be running Fabric OS v4.4.0d or higher.

FCIP • Frame drops observed on FCIP slow links:

o The frame drops occur when the FCIP tunnel bandwidth is set to 10 Base-T (10Mbps), E1 (1.048Mbps), or T1 (1.544Mbps).

o With E1 or T1, frames are dropped even without an impaired link.

o With 10 Base-T, frame drops may be observed when a low impairment is put to the link.

• portperfshow indicated incorrect (smaller) bidirectional throughput on the FCIP tunnel when Fastwrite/Tape Pipelining is enabled.

• Fastwrite/Tape Pipelining did not inform the user when it failed due to multiple equal paths configured on 2 GbE ports.

• Backup jobs initiated from the Symantec BackupExec application slowed noticeably after adding significant IO traffic from regular hosts and targets to the FCIP tunnel. A port-based routing policy must be used for Tape devices.

Access Gateway vs. Standard Switch Mode

• When using the Brocade blade server SAN switch in Access Gateway mode, most switch features are no longer applicable. These features include Admin Domains, Advanced Performance Monitoring, direct connection to SAN target devices, Fibre Channel Arbitrated Loop support, Fabric Manager, FICON, IP over FC, ISL Trunking, Extended Fabrics, Management Services, Name Services (SNS), port mirroring, Secure FOS, SMI-S, and Zoning. These switch features are available in the default switch mode of operation.

Access Gateway Mode Port State

• When a disabled port on a switch in Access Gateway mode is connected to a configured loop HBA, the port state alternates between Nosync and Insync. The switchShow command displays the state of the remote HBA port that is continuously attempting to reconnect to the disabled port.

• Brocade Access Gateway only supports FCP initiator connections on the F_Ports. Note that cascading Access Gateway devices or connecting FCP targets, loop devices, or FICON channel/control units on the F_Ports is not supported.


RFEs Implemented in Fabric OS v5.2.1

RFE Number Description

3868

Disable console magic key break sequence to avoid unintentional switch reboot/hung/panic due to user input or serial line settings.

3863 Add syslog IP addresses to Configuration upload file to allow upload/download.

3814 Log a syslog message when syslogd destination address is removed by syslogdipremove command

3797 When port speed is auto negotiated, Web Tool’s ports overview shows port speed as (AN-2), (AN-4), etc,. rather than just 1,2,4.

3615 Change HELP to be case insensitive.


RFEs Implemented in Fabric OS v5.2.0 RFE Number Description

3791 Set backspace key (^H) as erase key for firmwareDownload, configUpload and configDownload commands.

2953 Provide consistency in IP administration and configuration across different features and platforms.

3142 RLIRs are sent only to listeners within the same Brocade zone.

2487 Add domain ID for each ISL in the output for clarity (islShow and trunkShow.) Previously, the port numbers were shown. Adding the domain ID helps identify the destination switch.

2537 Allow administrator to clear port counters when necessary.

3082 Add information to supportShow help file that supportShow is a diagnostic tool.

3099 Add bsn (Brocade Serial Number) in supportShow to identify switch while trouble-shooting.

3114 Add date field to logging field in the portLogDump/Show.

3152 Add "top" (CPU util output) to supportShow.

3273 Successful login message in event log should show IP address of station logging in.

3532 supportShow now includes sfpShow –all.

Fabric OS v5.2.1 Documentation This section provides information on the documentation for Fabric OS v5.2.1.

New Hardware Documentation The following new manuals support the Brocade 5000:

Brocade 5000 Hardware Reference Manual (Publication number: 53-1000424-01) The Hardware Reference Manual is written for network administrators to provide a complete set of Brocade 5000 switch installation procedures and an overview of the switch hardware. This document is specific to the Brocade 5000 switch running Fabric OS v5.2.1.

Brocade 5000 QuickStart Guide (Publication number: 53-1000425-01) The QuickStart guide is intended as an overview to help experienced installers unpack, install, and configure a Brocade 5000 switch quickly. For detailed installation and configuration instructions, refer to the Brocade 5000 Hardware Reference Manual.

Brocade 5000 Power Supply/Fan Assembly Replacement Procedure (Publication number: 53-1000426-01) This document provides instructions to replace a power supply/fan assembly unit in the Brocade 5000 switch.


Brocade 5000 Rack Mounting Ears Installation Procedure (Publication number: 53-1000451-01) This document provides instructions to install mounting ears to the switch and install the switch in a rack with the mounting ears.

Updated Software Documentation The Brocade Fabric OS V5.2.x manual contains important last minute updates to the Fabric OS v5.2.0 Family Documentation set as well as instructions on using the DPOD feature. Use the Software Addendum and the Fabric OS v5.2.0 documentation set for instructions on administering a Fabric OS SAN.

The most recent Fabric OS v5.2.0 documentation manuals are available on the Brocade Partner Network: http://partner.brocade.com/.

Brocade Fabric OS V5.2.X Software Addendum (Publication number: 53-1000429-01) The Software Addendum is written for SAN administrators to provide a complete description of the DPOD feature and important last minute changes to the Fabric OS v5.2.0 documentation manuals. This document is specific to switches running Fabric OS v5.2.1.

New Software Documentation The following manual supports Brocade Access Gateway only. For detailed Fabric OS administration instructions, refer to the Fabric OS V5.2.X Software Addendum and the Fabric OS V5.2.0 Family Documentation set.

Brocade Fabric OS V5.2.1 Access Gateway Administrator’s Guide (Publication number: 53-1000430-01) The Access Gateway Administrator’s Guide is written for SAN administrators to provide a complete description of operating and managing a switch in Access Gateway mode.

Documentation Updates This section provides information on last-minute additions and corrections to the documentation. The most recent Fabric OS v5.2.0/v5.2.1 documentation manuals are available on the Brocade Partner Network: http://partner.brocade.com/

Brocade 5000 Hardware Reference Manual (Publication Number 53-1000424-01) On page 12, under the heading “Installation and Safety Considerations,” replace the following bullet:

“To install and operate the switch successfully, ensure that the following requirements are met:

• The primary AC input is 90-264 VAC (switch autosenses input voltage), 47-63 Hz.”

With:

“To install and operate the switch successfully, ensure that the following requirements are met:

• The primary AC input is 100-240 VAC (switch autosenses input voltage), 47-63 Hz.”


On page 30, under the heading “Facility Requirements,” replace the following bullet:

“Electrical:

- Primary AC input 90-264 VAC (switch autosenses input voltage), 47-63 Hz.” With:

“Electrical:

- Primary AC input 100-240 VAC (switch autosenses input voltage), 47-63 Hz.”

On page 31, in Table 2 “Power Supply Specifications,” replace the “Input voltage value” with the following:

“100 - 244 VAC, Universal”

On page 11, under the heading “Items included with the Brocade 5000,” replace the following bullet:

“• Power plug current/voltage rating: 15A/125V”

With:

“• Power plug current/voltage rating: 1.4A/125V”

Brocade 5000 QuickStart Guide

(Publication Number 53-1000425-01)

On page 5, under the heading “Items included with the Brocade 5000,” replace the following bullet:

“• Power plug current/voltage rating: 15A/125V”

With:

“• Power plug current/voltage rating: 1.4A/125V”

On page , under the heading “Site Planning and Safety Guides,” replace the following bullet:

“The primary AC input is 90-264 VAC (switch auto-senses input voltage), 47-440 Hz.”

With:

“The primary AC input is 100-240 VAC (switch auto-senses input voltage), 47-440 Hz.”


Open Defects In Fabric OS v5.2.1 Including defect disposition as of February 06, 2007, the following table lists defects that, while still formally “open,” are unlikely to impede Brocade customers in their deployment of this version of Fabric OS.

Note that when a workaround to an issue is available, it is provided; otherwise, no recommended workaround is available at this time.

Deferred Defects Found in a Release Other than Fabric OS v5.2.1 This table lists defects that were found in a release other than Fabric OS v5.2.1 and are being deferred to a future Fabric OS release.

Deferred Defects Found in a Release Other than Fabric OS v5.2.1

Defect ID Severity Description DEFECT000065145 High Summary: Kernel Oops error when the counter is overrun during stress-to-fail

test. Symptom: Switch failed with kernel trace. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000069537 High Summary: During stress testing of an unsupported configuration, chassis containing more than the supported number of FC4-18i and FC4-16IP blades, on one occasion the switch panicked and rebooted. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000070043 High Summary: When performing port mirroring, a SilkWorm 48000 printed an exception error and the kernel rebooted. Symptom: Error message: "Exception in kernel mode: sig=4, WARNING, SilkWoOops: Exception in kernel mode, sig: 4". This is a rare occurrence with a heavily loaded system. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000071195 High Summary: configdownload with new timezone is not being validated nor

applied completely; date uses old timezone, tstimezone shows new time zone Symptom: tstimezone command does not show the correct time and date if configdownload interface is used to update timezone. Workaround: After doing configdownload, run 'tstimezone' CLI to update Time Zone. The CLI validates the timezone as well as applies it. A reboot is required to make timezone completely effective in either case (config download or using CLI). Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000072085 High Summary: Switch panicked when the switch was being configured to Interop mode and domain id was changed at the same time. Symptom: While setting up the fabric to interopmode, when domain id was also changed, one of the switches in the fabric panicked. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.0.5

DEFECT000074395 High Summary: After executing trunk disable/enable on all edge switches in a large scalability fabric, one core SilkWorm 24000 E_port gets stuck in "Loopback" mode. Symptom: Cfgtrunkdisable/Cfgtrunkenable will cause the ports to be permanently in IN_SYNC. Workaround: Issue a portdisable/portenable on the problem port to recover it back to E_port online state. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000074978 High Summary: AD filter: portmirror allowed the addition of a mirror port in AD3 in

which the SID and DID were not port members. Symptom: The AD Enforcement on SID/DID is not performed. The traffic can be monitored between any SID/DID even though the ports are not under your AD Solution: Added SID and DID check for the AD Enforcement. The SID and DID both need to be part of AD apart from the mirror port to form a portmirror connection Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000075088 High Summary: A FR4-18i blade became faulty after running bladedisable/bladeenable then a continuous hafailover. Symptom: Error message: 2006/08/21-00:18:59, [EM-1034], 7846,, ERROR, Silkworm48000, Slot 10 set to faulty, rc=2001 Workaround: Perform slotpoweroff/slotpoweron for the affected FR4-18i blade. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.0

DEFECT000075139 High Summary: Issue hafailover command and the new active CP got Oops and reboot during initialization. Symptom: New active CP got Oops and reboots with error "kernel BUG at cachemap.c:91! Oops: Exception in kernel mode, sig: 4; >>NIP; c0010bb0 <consistent_alloc+0xa0/0x1c8 [kernel]> <=====" Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000075170 High Summary: GE0 port on an FR4-18i blade was stuck in No_Sync state,

preventing a VE tunnel from coming up. Symptom: After reboot of a CP, the VE tunnel does not come up. Workaround: Refrain from using copper SFPs. Remove and reinsert the cable on the GE ports Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000076053 High Summary: When there are more than the supported number of EX-ports attached to the same edge fabric, the switch might panic. Symptom: Continuous rebooting (due to panics) on FCR switches. Solution: Fixed code to gracefully handle an excessive number of EX_ports. Workaround: Use only the supported number of EX_ports. Customer Impact: Minimal, since there is a workaround. Probability: High Reported in Release: V5.1.0

DEFECT000076244 High Summary: Unexpected termination of zoned, switch crashes after upgrade Symptom: Switch crashes after upgrade between internal builds. Workaround: Perform upgrade while fabric is stable. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076284 High Summary: With more than the supported number of AP blades (five) enabled,

powering on the 4th and 5th AP blades at the same time immediately results in a failover. Symptom: A cold boot of the system with more than the allowable number of AP blades may result in all AP blades enabled. Solution: Changes to EM recovery and HSM code. Workaround: Poweroff one of the enabled AP blades (poweron and subsequent failovers should work). Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Medium Reported in Release: V5.2.0

DEFECT000076350 High Summary: During Scalability testing, after firmwardownload to Fabric OS v5.2.0 build, FR4-18i blade faulted. Symptom: Due to a rare occurrence an FR4-18i blade might become faulty after firmware download and slotpoweron followed by slotpoweroff. Customer Impact: This problem has been observed only once and several attempts to reproduce it failed. Probability: Low Reported in Release: V5.2.0

DEFECT000076525 High Summary: Tape Pipelining needs to properly handle target recovery in a congested IO condition Symptom: Tape backup job failures were occasionally observed in congested IO conditions Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076541 High Summary: After changing FID and enabling FCR switch in backbone, all

connected ports remain disabled due to a timing window of the LE-port offline and switch offline, leading to a computed stale link-state. Symptom: After enabling switch, all connected/active ports are disabled. Workaround: Perform a switchdisable/switchenable. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000076578 High Summary: After swapping an FC4-16IP blade and an FC4-48 blade, followed by failover, the FC4-16IP blade is faulted due to BM-BC heartbeat died during execution of POST2 iscsitest. Symptom: The FC4-16IP blade gets faulted and slotshow shows this blade is stuck at "DIAG RUNNING POST2" state. The following message will be seen on the console: 2006/09/22-17:38:39, [BL-1003], 228255, FFDC, CRITICAL, saturn64, Faulting blade in slot 9 Workaround: Remove the affected FC4-16IP blade, then re-insert it in chassis. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000076598 High Summary: Slotpoweron on an FC4-16IP port blade and hafailover after its FC initialization completes causes all its iSCSI ports to be stuck at "No_Sync" while the FC ports come online. Symptom: If the active CP fails over after an FC4-16IP port blade has been slotpowered on, all the iSCSI ports of the FC4-16IP can get stuck at "No_Sync", while the FC ports come online. Workaround: Issue portdisable and portenable for all affected iSCSI ports. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076640 High Summary: After an all switches disable on a SilkWorm 48000 with FC4-48

port blades, the standby CP asserted and rebooted Symptom: Error message: "Failed expression: snode != NULL". Solution: Standby CP automatically rebooted. Workaround: No workaround is needed. After the reboot the standby CP comes up online and HA is synchronized Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.0

DEFECT000076657 High Summary: HAfailover during FCIP IO (through an IPSec enabled tunnel) sometimes causes VE port to go down/up. Symptom: With IPSec tunnel configured, the VE might go down and then come back up during HAfailover. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000076677 High Summary: ISCSI stress testing: targets lost, PDUs for 1 session sent to another session Symptom: Hosts are losing targets and event logs on hosts: indicate incorrect ITT. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076686 High Summary: ISCSI: isnsccfg --set command should not accept an FR4-18i port.

Symptom: User will be able to configure the isns client to communicate with an isns server through a port on the FR4-18i blade. This will not work. Workaround: The user should either a) Configure the isns client to communicate with the isns server with a management port, or b) Run slotshow first, confirm the slot number to be that of a FC4-16IP and then use this slot/port combination. Customer Impact: Minimal, since there is a workaround. Probability: High Reported in Release: V5.2.0

DEFECT000057804 Medium Summary: EZM: Display device does not show legends when Restore Fixed Zone is done. Symptom: Usability issue which will give incorrect information to the user. Solution: Commented the code to throw Exception when hostConnsNum > maxHostConnsNum. Also uncommented the code to set the initiators and targets in WTConfig. Workaround: Close and reopen the EZManager window after restoring fixed zoning. Customer Impact: Users will not see proper legends on improper connection of devices. Probability: Medium Reported in Release: V5.0.1

DEFECT000059663 Medium Summary: When disabling external ports, the port diagnostic LEDs blinking are not synchronized Symptom: The physical LEDs for disabled ports flash independently of each other. Workaround: None Customer Impact: No customer impact on switch operation. The amber flashing may appear disorganized to some observers, however, the switch is functioning as expected. Probability: Medium Reported in Release: V5.0.2



Defect ID Severity Description DEFECT000063046 Medium Summary: External port with faulty SFP is disabled then enabled with good

SFP and still shows as faulty. Symptom: Port will seem faulty, even though it is actually healthy. Workaround: After enabling the port with the good SFP, remove and re-insert the good SFP. The port will show as healthy. Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.0.4

DEFECT000063566 Medium Summary: With Web Tools, the error message is not clear when the fmsmode setting fails. Symptom: It is unknown what the exact error is that causes setting of fmsmode to fail when using Web Tools. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.1.0

DEFECT000068760 Medium Summary: Usability: AD0 allowed user to disable or enable VE ports but prevented user from creating or deleting VE tunnels Symptom: Gigabit Ethernet (GE) ports can be configured only from AD0 or AD255. Configuring AD1-AD254 is not allowed. VE ports can be created with a GE port. The created VE ports can be controlled from any ADs. Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000068888 Medium Summary: Web Tools is missing "Invalid CRCS" under Fabric Watch end to end class pull down menu. Symptom: Missing "CRC Error" option in Web Tools under End-to-End class. In CLI "Invalid CRCS" is displayed. Customer Impact: Usability issue, with no impact to functionality. Probability: High Reported in Release: V5.0.5



Defect ID Severity Description DEFECT000069724 Medium Summary: SilkWorm 4018: Internal ports can be configured as L_Ports under

special circumstances for SilkWorm 4018 (portcfgdefault). Symptom: With the "portcfgdefault" command, the port information window for internal ports on Web Tools will display the Port type and Allow port type as "U-Port" and if the end user executes "portcfglport" it prevents the F-port server from logging in and could result in unexpected behavior of the internal ports. Workaround: config upload/download may be used to reset the configuration of a switch Customer Impact: Minimal, since there is a workaround. Probability: High Reported in Release: V5.0.5

DEFECT000069966 Medium Summary: Usability: firmwaredownload downgrades from v5.2 do not offer automatic feature removal to the customer. Symptom: Firmwaredownload does not offer automatic removal of some software features when they block customer downgrade capability. Customer Impact: Usability issue, with no impact to functionality. Probability: High Reported in Release: V5.2.0

DEFECT000070707 Medium Summary: In AS4 and Sol10, All the Applets like Zone admin and name server windows are not refreshed properly after resizing. Symptom: GUI components size will be a bit squeezed. Solution: new gui alignment solved the problem Workaround: Resizing the panel, Scrollbar comes. This happens only for the first time. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000071060 Medium Summary: Web Tools display problem aligning traffic information.

Symptom: The customer will find inconsistency in the representation of the traffic corresponding to the [slot] ports in the graph . Customer Impact: Cosmetic issue, with no impact to functionality. Probability: High Reported in Release: V5.0.5

DEFECT000071471 Medium Summary: Need errlog messages for DHCP configuration/operations on a switch Symptom: errlog (raslog) messages will not be available for diagnosing DHCP problems. Workaround: none Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000071738 Medium Summary: Web Tools->Fabric Watch->Environment Class displays incorrect "Last Value" for "Temperature" area. Also Web Tools "Current value" erroneously takes CLI "fwShow" "Last Value". Symptom: The 'Last Value' in Web Tools does not match the 'Last Value' in CLI displayed by fwshow. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.0.5

DEFECT000072169 Medium Summary: Web Tools allows (switch,port) members to be added as a member of a zone when the switch is in interop mode. Symptom: The user might be led to believe that a zone with a member as (switch,type) is actually created successfully. Customer Impact: Usability issue, with no impact to functionality. Probability: High Reported in Release: V5.0.5



Defect ID Severity Description DEFECT000072269 Medium Summary: Some switches are segmented with reason alias ID overlap when

selected switches in fabric are configured to use Alias Server Symptom: Some switches are segmented with reason alias ID overlap when selected switches in fabric are configured to use Alias Server This defect will only show up when there is any switch in the fabric with active alias ID tokens defined. In this case, the FabricOS v5.2 switch will segment out of the fabric with reason "domain overlap" in switchshow. But the console log will display a message with "alias id overlap". Although these messages are misleading the situation can be determined by fabricshow as it will display some alias tokens after the domain list. Workaround: Fabric OS v5.2.0 does not support Alias Server. Refrain from using the Alias Server feature in fabrics that include Fabric OS v5.2.0 switches. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000072479 Medium Summary: Solaris- Events list in Call Home e-mail is not getting updated with

exact reason for alert because the switch is already in marginal state and the marginal event does not impact the current marginal status. Symptom: Customer will not get an event message for call home when switch status has an event that doesn't change the current switch status. For example, a switch in marginal state has another event that is marginal. Solution: Fabric Watch sends message when ever there is change in switch status i.e from Healthy to Marginal/Marginal to Down /Down to Healthy/Down to Marginal.And it gives a message describing the contributing factor for the status change. When switch status is changed to Marginal or down state,the first contributing factor which has made the status change is reported.But if the switch remains in Marginal or down state and if someother factors(Power supply,Fan etc) go down yet not changing the switch status such instances are not reported.Due to these we do not get the exact picture of all the factors that contributed for the switch status.This problem is resolved and messages are displayed reporting the status change with the contributing factor when there is switch status change also when the switch is in marginal/down state are reported as as well. Workaround: user has to check switchstatusshow CLI for contributing factors. Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000072551 Medium Summary: Extra RSCNs seen during HAfailover Symptom: Offline/Online events may be observed during hafailover Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000072617 Medium Summary: FCIP:Error message is not proper while editing\adding the subnet

mask of an IP address conflicting from the existing IP Addresses. Symptom: In the Port Administration Services window, "IP Interfaces" tab, if the user edits/adds a subnet mask of an IP address that conflicts the with the exisiting IP Addresses the following behavior occurs: In the GigE Port Configuration Wizard, an error dialog is shown with the message " Failed to edit IP Interface;Reason:com.brocade.dm.Dao Exception...". Solution: corrected error message displayed. Workaround: Input the correct subnet mask to correct the problem. Another workaround is to use the CLI command "portcfg". If the user makes the same mistake, a proper error message is generated. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000072655 Medium Summary: Setting portcfgilsmode causes E-port to become G-port on 5.2 switch and E-port to unknown on 5.0.4 switch when connected together Symptom: The E-port of the 5.2 switch becomes a G-port and the E-port of the 5.0.4 switch becomes unknown when portcfgislmode is set for the 5.2 switch Workaround: Set portcfgislmode port# 1 on both ports. This way the ports will come up correctly on both switches. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.0.4

DEFECT000072778 Medium Summary: Device Connection Step in EZManager not differentiating ports as Storage and Hosts Symptom: User will not be able to differentiate between Storage and HBA ports if no devices/hosts connections are made. Solution: To show storage and host images for ports in device connection view. Workaround: Use EZSwitchSetup screen shot to see storage and HBA ports classification. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000072942 Medium Summary: send class 2 SOFi error frames from SANtester to attack switch with

600+ devices, 48K switch reboots with wdtd on, extremely slow if wdtd is turned off. Symptom: Generation of a very big core file causes the system to remain busy for a long time. This results in a watchdog timer timeout. This is a rare case as nearly all daemons and applications do not generate a very big core file if they crash. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.0

DEFECT000073512 Medium Summary: Some invalid zone objects are accepted by zonecreate command. Symptom: Under normal circumstances, there is no impact to user. If user is performing an unsupported activity like zonecreate "zone_test", "10:0:0:0:000:00:0:00", the switch will not provide an error about the WWN. Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000073600 Medium Summary: Web Tools should display "logged in" Radius account information Symptom: Web Tools is not displaying "logged in" Radius account information. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000073608 Medium Summary: SFP tab: SFP tab not showing the Units for temperature, and Rx power and TX power. Symptom: In Web Tools under SFP tab, the units for temperature, Rx and Tx power are not shown. Customer Impact: Usability issue, with no impact to functionality. Probability: High Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000073903 Medium Summary: Fabric Topology Window does not reflect the units for Bandwidth

Demand and Total Bandwidth on Web Tools. Symptom: Customer will not be able to see the units of Total Bandwidth and Bandwidth Demand from Web Tools. Workaround: none Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000074119 Medium Summary: fwshow continues to show old temperature sensor data after a blade is removed from chassis Symptom: FabricWatch temperature sensor data is skewed up when a blade is removed from a chassis. Also the "last value" is not updated though sensor data changes. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000074418 Medium Summary: SilkWorm 7500 GE port was stuck in "No_Sync" state Symptom: On rare occurrences, if a cable is first partially inserted into the SFP of a switch connected to a Brocade GE port, then fully inserted, the GE port can get stuck in the No_Sync state. Workaround: 1. Issue portdisable followed by portenable on the GE that is stuck in No_Sync state. 2. Pull out the cable and plug it back in (without pausing during the insertion of the cable). Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000074463 Medium Summary: Web Tools should provide real fail reason when config upload fails

due to invalid file path. Symptom: Customer will see an error message of " Directory change failed" instead of " File path provided is not valid" Customer Impact: Cosmetic issue, with no impact to functionality. Probability: High Reported in Release: V5.2.0

DEFECT000074563 Medium Summary: Modifying port speed and trunking on FC4-48's ports from Web Tools followed by hafailover and then trying to re-launch, "Port Admin Window" on the same ports fails. Symptom: Changing the port speed on a few ports, disabling trunking on some other ports from Web Tools on an FC4-48 port blade, hafailover and then trying to re-launch, "Port Admin Window" on the same ports fails to reload the Web Tools window. Workaround: Close the Web Tools window and non-responsive Ports Admin window. Relaunch after a few minutes and then access the Ports Admin window. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000074926 Medium Summary: EZSwitchSetup: Device display is overlapped in Display

Connection window. Symptom: If the window size is not maximized, the images of hosts and ports don't overlap. Even if they overlap, the two images do not merge completely and are easily distinguishable. Solution: Fixed the x-cordinate of switch image as zero so that on resizing the image, the x-cordinates of host/storage will not change and the images will not overlap. Reduced the y-cordinate of host/staorage image by 20 pixels to avoid overlapping Workaround: Reduce the browser window size. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000074981 Medium Summary: Some storage device port comes up as L-port instead of F-port after reboot and the hosts no longer see that storage device. Symptom: F_port may come up as L_port after the storage device is rebooted. Solution: If problem happens, lock gport by issuing portcfggport Workaround: If problem happens, lock gport by issuing portcfggport Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000075058 Medium Summary: portstatsclear help page description of behavior on 2GB and 4GB platforms is ambiguous Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Service Request# RQST00000051350 Reported in Release: V5.0.4



Defect ID Severity Description DEFECT000075112 Medium Summary: FR4-18i tracedump files contain many BP error messages and one

empty file. Symptom: No customer symptom except possible incomplete support files. Workaround: None Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000075198 Medium Summary: PerfMon:Switch level graphs display inconsistencies in Traffic patterns for Logical FC ports in GiGE0. Symptom: Basic Performance Monitor Graphs in WebTools such as Switch Throughput Utilization and Switch Percent Utilization for GigE0 , is inconsistent with the PortPerfshow command. Logical Port 16 ,with no tunnel configured on it is displayed with traffic pattern instead of Logical Port 17, where the actual traffic is flowing through the tunnel. Workaround: Configure the tunnels in some other Logical Port in GigE0 and observe the traffic information. Another workaround is to use the PortPerfShow command (CLI) to observe the traffic statistics. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000075226 Medium Summary: Usability: Ctrl-C in legacy configdownload code is not blocked during downloads Symptom: Due to considerably longer configdownload times in v5.2, users may get nervous about a configdownload in progress and use the ctrl-c key to escape from the utility. Depending on when this is done, it could result in an ill-defined switch state. Workaround: Most configdownloads are fast and will not be interrupted by a user ctrl-c sequence. The window is also small in which this can create problems. The workaround is to avoid use of ctrl-c to interrupt this command. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000075261 Medium Summary: Scal_2560: After a couple of iterations all core switch fastboot,

some of the E-Ports on core switch stuck in "Mod_Val" Symptom: When rebooting all core switches in a loop, E-Ports may get stuck in "Mod_Val" mode. Solution: Do a portdisable/portenable on the problem port and it will come online. Workaround: Do a portdisable and portenable on the problem ports and it will come online. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.0

DEFECT000075341 Medium Summary: Incorrect WARNING messages displayed when SFPs are removed from ports on SW48K(FC4-48 port blade), moved to other ports and hafailover is issued. Symptom: Error messages: "SW48Kchassis, Error reading SFP ID. Media may not be plugged in properly (Code = -21, slot = 1 port = 0)., i2c_func.c, line: 1302, comp:module-99-th, ltime:1970/01/01-00:00:00" Solution: Looks like its in the switch polling routine that want to read the gbic information, but it's not there. Solution is to not read it if SFP isn't there. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000075389 Medium Summary: commitTransaction does not return 'specific error code' when zone fails to propagate in the fabric because the zoneDB reached the maximum size. Symptom: FOS does not return the appropriate error code to the SMIAgent when the commit of a zoning transaction fails due to reaching the maximum zone database size. Instead FOS returns a generic error code that tells the user 'Zoning transaction commit failed' but doesn't indicate the specific reason for why it failed. Workaround: None Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000075467 Medium Summary: Error messages are generated when applying iSCSI easycreate

operation from Web Tools on unreachable FC Targets. Symptom: No error messages if LUN information is not retrievable from an FC Target. Errors are only displayed when applying the operation. Workaround: Do not use physical FC targets where LUNS cannot be retreived(invalid luns). Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000075502 Medium Summary: Web Tools Switch Admin is sometimes unable to turn Dynamic Load Sharing (DLS) and In-Order Delivery (IOD) on/off. Symptom: If Dynamic Load Sharing or In-Order Delivery is disabled, it cannot be enabled from Web Tools. Workaround: There are no workarounds from Web Tools interface. Use CLI to set/reset DLS. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000075565 Medium Summary: Web Tools event filtering uses a time range of 0-11 as opposed to 1-12. Symptom: In Web Tools Fabric Events and Events window, on invoking filter button and filtering based on from time or to time, on providing a time greater than 12 (e.g. 12:30 PM) will result in an error stating that time should be in range of 0-11 Solution: -->Webtools can go with 24hrs notation for event filtering instead of 12hrs.(or) -->Webtools can go with "(1-12)" hour in am/pm instead of existing "(0-11)" hour in am/pm. Workaround: Not required. Existing "(0-11)"hour time format is working fine. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000075615 Medium Summary: Additional Port Info column in FC Ports tab shows "None (FMS

Mode)" instead of "Disabled (FMS Mode)" Symptom: Additional Port Info column in FC Ports tab shows "None (FMS Disabled)", although CLI shows it as "Disabled (FMS Mode)". Solution: An API similar to switch_port_is_active_fms_port() should be provided to web linker. Workaround: The user can refer to the jPortStatus column in FC ports tab to know the port status. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000075760 Medium Summary: During firmwaredownload upgrade process, seeing "Blade rebooted during firmware commit The operation will be restarted." messages for FR4-18i blades after the commit has started on these blades. Symptom: firmwaredownloadstatus output shows that FR4-18i blades both rebooted during a firmware commit process that had already started. The process recovers. Solution: Changed the BM code. Workaround: none required Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000075911 Medium Summary: EZSwitchSetup: Fixed zone cannot be configured with all ports designated as storage or hosts. Symptom: The correct constraint is being enforced, but the message displayed does not describe how to correct the misconfiguration. Solution: Modified the message to be displayed to be meaningful when user selects all ports either as storage/host. Workaround: Change port assignments so that there is at least one host port and one storage port. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076025 Medium Summary: The current password is taken as an invalid password for the first

time to change password after an account is expired. Symptom: Sometimes, after current password is provided correctly in this scenario, Web Tools responds as an invalid password; this can be recovered by trying the same again. Workaround: Retry the same correct password. Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000076026 Medium Summary: Error messages observed during IPsec FCIP IO. Symptom: On initiating IPsec FCIP IO, error mesage displays on console: "[CDR-5206], 0,, ERROR, sprint_73, UPSM [OID 0x4302880a] (1) (UP10): Invalid event [0x1030035, 0x0, 0x0], OID:0x4302880a, proto_upsm.c, line: 1797, comp:snmpd, ltime:1970/01/01-00:00:00". Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000076037 Medium Summary: LinkRNIDDeviceRegistration trap provides wrong nodeRNIDIncidentPortWWN and nodeRNIDConnectedPortWWN Symptom: Wrong information during request node identification data (RNID) registration. Solution: Fill the incident port wwn and connected port wwn fields before calling ms_ficon_notification() Probability: Medium Reported in Release: V5.2.0

DEFECT000076060 Medium Summary: Port type gets swapped in the SAM Report when the port indices of 2 ports are swapped. Symptom: Customer will see wrong port type after port swap on issuing fwsamshow from CLI. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076112 Medium Summary: FCIP:In the absence of IPSEC license,"Show IP Security policies"

doesn't give the policy details where as secured tunnel gives the IPSEC policies associated with it. Symptom: The customer will be unable to view the policy details in the absence of IPSEC license from "Show IP Security policies". Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000076155 Medium Summary: Observed frames being discarded from frame filter on secondary port when disabling the primary port that shared the same area on a FC4-48. Symptom: When both the shared ports are receiving traffic and the primary port goes offline, all the frames that are out for delivery for the primary port will be dropped (as usual) but the counters will show them as dropped on the secondary port. Workaround: Customer can clear the counters on the secondary port after primary port goes offline. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: High Reported in Release: V5.2.0

DEFECT000076195 Medium Summary: After changing an expired password and selecting another AD during the same session, the password expiration message and password prompt are displayed again. Symptom: User has to change the password on every ad select in the same session. Workaround: Exit session and re-login after changing an expired password. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000076255 Medium Summary: Error messages from ACL configuration through SMI are too generic and do not contain sufficient information to debug the problem. Customer Impact: Usability issue, with no impact to functionality. Probability: High Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076260 Medium Summary: Portstats: Selecting the "Show delta" radio button and clicking on

the clear counter button does not clear the counters; instead it changes the values to negative. Symptom: Clicking on the clear counter button does not clear the counters; instead it sets the values to negative. Solution: On selecting show delta button, stats will be first shown as zero, then after first refresh cycle, it will show delta values. Even on clicking 'Clear Counters' button, it will first show all zero values, then after first refresh cycle, it will show new values. Workaround: You may need to click the clear counters more than once to set to zero. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000076265 Medium Summary: FFDC files are generated when ipaddress is manually set to a switch which is in DHCP network Symptom: customer may see ffdc warning messages when manually changing switch ip address Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000076269 Medium Summary: No warning message is displayed when using Web Tools to configure/Edit FCIP tunnels with the fastwrite -tape pipelining feature Symptom: When configuring fast write using Web Tools, the customer will be not receive a warning indicating that the fastwrite and tape pipelining features should not be used with multiple equal cost paths. Solution: Fix some FCIP defects related to message, dialog exception throw out. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076272 Medium Summary: Able to establish API session to AD 255 using "user" login which

has AD membership only to AD 0 which is not allowed from CLI. Symptom: End users of SMI and Web Tools will not observe this error because all operations are performed via an Admin role user. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000076308 Medium Summary: Web Tools should make the feature consistent with CLI and allow user to configure long distance for ports if current AD owns these ports but not switchmembership Symptom: Port long distance settings are read-only for owned ports on non-owned switches in WT Extended Fabric tab, but editable for owned ports in WT port configuration wizard. Workaround: Use the Web Tools port configuration wizard Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000076372 Medium Summary: Web Tools does not show the reason for configdownload failure while CLI shows the reason. Symptom: Customer will not know why a particular configdownload operation failed. He/she will have to use configdownload again in CLI to know the reason. Workaround: Use configdownload in CLI Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076400 Medium Summary: Firmware downgrade to Fabric OS v5.1 or prior is blocked even

after ports 256-383 are removed from a DCC policy created using [*] option. Symptom: Every time ports are removed, VERIFY is displayed on console. Deleting the DCC policy will resolve the problem. Workaround: Delete that DCC policy and continue Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000076401 Medium Summary: Admin having permissions to just AD255 is able to create ADs through configdownload. Symptom: Admin user account with permissions to only AD255 will be able to create Admin Domains. Expected behavior is that only Admin user accounts with permissions for all ADs should be able to create/modify ADs. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000076430 Medium Summary: The confirmation dialogue is not appearing when we close the name server table by clicking the 'X' button. Symptom: User will experience an unexpected close by clicking the 'X' button accidentally. Solution: Conformation dialog is removed from the Close button. Workaround: Customer can click on the Yes / No instead of the "X" button Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076433 Medium Summary: Detailed view of any device does not refresh properly when Auto

refresh option is selected. Symptom: Customer will see empty values for Name Server entry in the detail view, upon attempting to resize/scroll the window. Workaround: Close and open the detail view dialog after the refresh cycle. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000076462 Medium Summary: Role name(and sometimes user name also) missing from error message when a user exits from a telnet session Symptom: SEC-3022 misses the role name or user name. Solution: The solution is getting the ENV variables from the pam and then displaying them in the AUDIT message. Role name was not appearing when a user exits out of the telnet session. Explicitly setting the role name in pam_sm_close_session will make the role name appeared in the SEC-3022 Explicitly setting the role name using gset_role for getting appeared in the AUDIT messages Workaround: Look at SEC-3020 audit message (logged when a user logins) for the particular user and find his role. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076490 Medium Summary: SilkWorm 3850 with port configured for longDistance and R_DY

transmits ARB(vc). Symptom: SilkWorm 3850 running Fabric OS v5.2 with the following port configuration: - portcfglongdistance <port> LS 1 50 - portcfgislmode <port> 1 produces ARB(vc) fill words instead of idles. This might cause failures with certain long distance devices Workaround: fabric.ops.mode.longDistance must be set on all switches in fabric. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000076492 Medium Summary: When a zoning transaction is opened and a switchdisable/enable or ISL remove/reconnect is performed, the transaction might be aborted. Symptom: The expected behavior is that the transaction should remain open Workaround: If the local transaction drops due to zonemerge, the user needs to create a new transaction. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000076498 Medium Summary: Web Tools is not correctly reflecting the physical switch state of a persistently disabled switch when the switch is fastbooted. Symptom: Web Tools shows port LEDs as dark (not lit) instead of slowly blinking yellow when switch does a fastboot after it is persistently disabled. Solution: update the port speed and led even when switch is persistently disabled or investigate further and see if this is expected behavior. Workaround: none Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076509 Medium Summary: No event is generated when the IP address of a switch is changed.

Symptom: Customer cannot determine that the ip address of the switch was changed by issuing the command errdump or errshow because an ip address change event was not registered. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000076514 Medium Summary: Several issues observed in ZoneAdmin and in AD when the switch is configured in PID format 2 Symptom: Customer will not able to see the ports 16-32 when configuring a switch in PID 2 format Solution: Enhanced the area ID mapping when PID format change to 2. Workaround: Perform zoning operations from the CLI when the switch is configured in PID 2 format. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000076522 Medium Summary: During slot commands test, a SilkWorm 48000 asserts and tracedumps with error message. Symptom: Executing a script that issues a series of bladedisable/enables. slotpoweron/off is required to produce this issue. Following a slot power on, assert and tracedump is encountered with "ASSERT - Failed expression: (data != NULL) && (iuRx != NULL), file = fc/fcop.c, line = 209, user mode" Solution: Fix where the data field in the exchange manager goes to NULL. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076528 Medium Summary: cfgtransabort and ad --transabort commands are shown as run when

only one of them are executed by user; note that cfgtransabort is invalid operation in AD255 and similarly ad --transabort is invalid in AD0-254 Symptom: Aborting Admin Domain transaction generates redundant message about cfg transaction being aborted. Similarly, aborting cfg transaction generates redundant message about ad transaction being aborted. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Medium Reported in Release: V5.2.0

DEFECT000076535 Medium Summary: Traffic stops for a moment before taking redundant path after rebooting one of the FCR in a dual BB fabric. Symptom: Customer may see significant delay in traffic coming back in case of one FCR failure in a fully redundant SAN configuration. The problem does not show up with SAN Tester device. Could be specific to a host or a setup since removing the other redundant path does not cause traffic disruption. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Medium Reported in Release: V5.2.0

DEFECT000076538 Medium Summary: Fabric is busy after running a group of commands repeatedly and not able to execute any commands. Symptom: Switch may report fabric busy following a long string of operations Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076545 Medium Summary: When using interactive mode to execute pathinfo with an invalid

source port value, there is no errors message. If execute the pathinfo with an invalid source port value, it will tell user there is no response. Symptom: User may think that the invalid source port value is being taken because there is no errors or warning messages. Workaround: Use pathinfo cli command in batch mode. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000076550 Medium Summary: iSCSI: Inconsistent results when a login request is rejected by FC4-16IP Symptom: After receiving a login reject, intermittently under some conditions the subsequent login attempts may fail as well. Solution: Move the session back to init state after rejecting a login. Workaround: None Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Medium Reported in Release: V5.2.0

DEFECT000076585 Medium Summary: iSCSI: Switch runs out of memory when runnning slotpoweroff/on continually on two separate FCR-16IP blades. Symptom: Switch may run out of memory if continually running slotpoweroff/on on multiple iSCSI gateway blades for a long period (> 9 hours). Workaround: None Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076587 Medium Summary: After changing zone configuration in the edge fabric, one of the

FCRs in the backbone fails to display an xlate domain in fabricshow Symptom: Customer will notice a difference in the fabricshow outputs between two FCRs in the same backbone Workaround: If EX port is disabled and enabled the xlate domain will appear correctly in both FCRs. This should be minimal impact to the user considering the entire zoning configuration was changed and a certain level of interruption was already expected. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000076591 Medium Summary: Need explicit way to clear iSNS server IP address. Symptom: No CLI option to clear the IP address of iSNS server. Workaround: Either use fosconfig --disable isnsc to disable isns which clears the Server IP address or run isnsccfg --set slot/port -s 0.0.0.0 to reset the server IP address. Customer Impact: Minimal, since there is a workaround. Probability: High Reported in Release: V5.2.0

DEFECT000076596 Medium Summary: No warning message is given when a lsan zone with a zone name larger than 58 characters is created in AD1 - AD254. Symptom: LSAN zones in AD1- AD254 with names having more than 58 characters are not propagated and no warning is displayed. Workaround: Use LSAN zone names shorter than 58 characters in ADs 1 to 254. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076597 Medium Summary: In Secure FOS, Web Tools gets null pointer from java console when

trying to modify default admin and user accounts Symptom: In SFOS, when the attempt is made to modify default admin account (login as root) and user account (login as root or admin), Web Tools should bring up the modify dialog for user to modify description and/or status. A null pointer exception appears in java console. Solution: Enhanced AD capable check in Secure FOS Add protection to null object comparasion. Workaround: Use CLI to complete modifying the default user account in secured FOS. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000076605 Medium Summary: Supportsave on a SilkWorm 48000 displays error message. Symptom: While collecting the supportsave info on a SilkWorm 48000, a user might see the following message on the active CP's console log: "SW48Kchassis, supportSave's ftp operation aborted error:0x0 cmd:module:BURNINSTAT dump cli failure, Error:0 127" Workaround: none Customer Impact: Usability issue, with no impact to functionality. Probability: High Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076681 Medium Summary: iSCSI: SendTargets=All sends portal information of disabled

ge_ports to 3rd-party host during login. Symptom: 3rd-party host takes a long time to log in due to wrong portal information being sent. Solution: If the portal is configured with ipaddress, the port is considered for reachability by the host at any point in time. To interop with 3rd-party OS, user should remove the ipaddress of the ports that are disabled or not reachable by the 3rd-party port. Workaround: Remove the IP address of the disabled port. portcfg ipif slot/ge# delete <ipaddress> <netmask> Customer Impact: Usability issue, with no impact to functionality. Probability: High Reported in Release: V5.2.0

DEFECT000076687 Medium Summary: ISCSI: iSNS console log reports incorrect name for management port. Symptom: The console log for iSNS reports the management port as slot/port 0/0. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: High Reported in Release: V5.2.0

DEFECT000076693 Medium Summary: iSCSI: AIX host is not able to see disks after successful login (with and without CHAP) when immediate data is set to enable Symptom: The user may not see the targets after successful iSCSI login using AIX if immediate data is enabled. Solution: Fix text parameter processing code to return the negotiated value instead of configured value. Workaround: Disable immediate data Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076710 Medium Summary: ISCSI: with multiple switches in the fabric and ISNSC enabled,

error is logged on only one switch. Symptom: iSNS error message is logged on only one switch in the fabric: it should be logged on all switches. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000076732 Medium Summary: After changing Backbone FID, switchshow not displaying some LE Port information. Symptom: Switchshow output is incorrect; LE port information not displayed for some EX or VEX ports. Workaround: LE Ports are functioning correctly, only a display issue. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000076736 Medium Summary: LE ports are not up after trunkdisable/enable. Symptom: EX-Port does not come online Solution: unknown - need to recreate the problem to get the fcrlog right after the erro occurs. Workaround: The work around is switchdisable/enable of the FCR switch. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0



Defect ID Severity Description DEFECT000076740 Medium Summary: EX PORT Disabled (Setting VC Credits failed) after enabling

switch Symptom: EX-Port fails to come online after switch is enabled. Workaround: Disable and then enable the EX-Port Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000076760 Medium Summary: execquery(Brocade_SwitchFcPortRateStats) is not returning correct Instances, When AD's are created on the Switch Side without stopping the SMI-A, it always returns Instances with respect to AD0. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000076786 Medium Summary: Sometimes when trying to add/modify AD list for multiple user accounts simultaneously, AD list for user accounts changes. Symptom: When this happens, the AD list for the first add/modify user account will be added to the second or following add/modify user accounts AD list. Solution: Enhance the string buffer handling for user AD list Workaround: Apply the change for adding or modifying user account AD list one user at a time. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000076815 Medium Summary: Port Detail Report displays the iSCSI Ports. Symptom: Customer sees the iSCSI ports in the Port Detail Report, also shows incorrect values. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0


Deferred Defects Found in Fabric OS v5.2.1 This table lists defects that were found in Fabric OS v5.2.1 and are being deferred to a future Fabric OS release.

Deferred Defects Found in Fabric OS v5.2.1

Defect ID Severity Description DEFECT000077966 High Summary: On an already fully utilized trunking group, a newly added slave

port may not be ready to pass the first few frames right away. Symptom: Testing tools that do not retry may detect a few frames lost on the path through the new slave port. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1

DEFECT000078694 High Summary: After continuous slotpoweroff/on, a panic happened on a Silkworm 48000 with AP blade Symptom: Error message: "kSWD: Detected unexpected termination of: ''[6]fabricd . . . " Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.1

DEFECT000081361 High Summary: IOF: luns continously disappear from iscsi initiator host Symptom: When header digest error is detected by iSCSI port, customer may see iSCSI sessions go away that will lead to devices go away. Solution: On header digest error handling, the response buffer is free twice. One in TCP close callback and one at the caller of the TCP close API. This lead to iSCSI port to hang and heartbeat failure detected on BP. BP then reboot iSCSI port. Hence all iSCSI sessions are cleanup. The fix is to remove the call to free the buffer in the caller of the TCP close API since it will be done in the close callback function. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1



Defect ID Severity Description DEFECT000074876 Medium Summary: F_Port Route cleanup does not happen correctly when NPIV is used

Workaround: Disable NPIV on all F_Ports on Access Gateway using "portcfgnpivport <port> 0". Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1

DEFECT000076864 Medium Summary: SW4020: The "restart module and run standard diagnostics" operation takes longer than the specified 60 second limit. Symptom: No visible symptoms, such as error messages in the management module logs. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.1

DEFECT000077034 Medium Summary: SilkWorm 4020: Unwanted message displays when we try to download a non-Access-Gateway configuration file in Access Gateway mode. Symptom: Message: "Warning:non-Access Gateway configuration file cannot be downloaded on a Access Gateway platform . . . configDownload not permitted." Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.1

DEFECT000077037 Medium Summary: The userrename command may be used to rename a user created user with "User" role to Brocade default user "admin". Symptom: Web Tools would mistakenly identify "admin" as an Admin role user and incorrectly grant it Admin role privileges.Only occurs on an embedded blade Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1



Defect ID Severity Description DEFECT000077592 Medium Summary: In Web Tools, if the incorrect username is entered at login, the

prompt displays an unhelpful error message. Symptom: Message: "invalid encryption data." Behavior of the return value to display the proper error message is inconsistent. Solution: A common error message can be returned even when the geps_auth_type() function fails or and incorrect username or password results in a login failure. Workaround: A common error message can be returned even when geps_auth_type() functions fails or username incorrect or password incorrect ,error being Login Failure. Customer Impact: None if recommended service procedures are followed carefully. Probability: High Reported in Release: V5.2.1

DEFECT000077823 Medium Summary: Web Tools should display a list of devices connecting to the NPIV port. Symptom: Customer will not be able to see the list of connected devices from Web Tools in Fabric OS v5.2.1. Workaround: Customer can use CLI to see the required information. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.1

DEFECT000077972 Medium Summary: Cannot enable Access Gateway mode without a Zoning License. Symptom: The agmode enable command fails with error message "Failed to clear Zoning/Admin Domain configuration". Workaround: Make sure zoning license is installed before doing ag --modeenable Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1



Defect ID Severity Description DEFECT000078109 Medium Summary: configdownload exits with "segmentation fault" when invalid value

passed to parameter "Do you want to continue [y/n]:" Symptom: configdownload terminates with "segmentation fault" when invalid value passed to parameter "Do you want to continue [Y/N]:". Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1

DEFECT000078237 Medium Summary: Executing fwfrucfg on a switch in access gateway mode displayed an incorrect error message. Symptom: Message: "There are no FRUs present on this platform." Message should be "Command not supported on this platform." Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1

DEFECT000078313 Medium Summary: SilkWorm 4016 displays incorrect raslog message for a failover situation when the Access Gateway switch is rebooted. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.1

DEFECT000078355 Medium Summary: Virtual port login (FDISC) from NPIV-capable HBA on an F_Port is not handled correctly. Workaround: Disable NPIV on all F_Ports of Access Gateway switch using "portcfgnpivport <port> 0" Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1



Defect ID Severity Description DEFECT000078463 Medium Summary: SilkWorm 4020 and 3016: restore factory default from MM will not

restore the default switchname Symptom: SilkWorm 4020 and 3016: restore factory default from MM will not restore the default switchname Workaround: After restore factory default, the user may run the following command to restore the default switchname, as a workaround: switchname "brocade4Gb" Customer Impact: Cosmetic issue, with no impact to functionality. Probability: High Reported in Release: V5.2.1

DEFECT000078592 Medium Summary: Ports come up as "No_Sync Disabled (Switch not ready for F or L port)" after rebooting switch Symptom: After switch reboot, besides E-Ports come up, all other ports are disabled as " In_Sync Disabled (Switch not ready for F or L ports)" in switchshow Customer Impact: This problem has been observed only once and several attempts to reproduce it failed. Probability: Low Reported in Release: V5.2.1

DEFECT000078852 Medium Summary: SilkWorm 4900: console message "Unknown bridge resource 2: assuming transparent" seen when rebooting. Symptom: Message: "Unknown bridge resource: assuming transparant" This is a Linux message. Workaround: No workaround is needed. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.1



Defect ID Severity Description DEFECT000079176 Medium Summary: Console error "Can't locate module char-major-4" when rebooting

switch Symptom: Message: "Can't locate module driver char-major-4" This is a generic Linux message. Workaround: No workaround is needed. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.1

DEFECT000079304 Medium Summary: Access Gateway should handle an IPv6 address in a GMAL payload. Symptom: IP Address in "ag --show" truncated to 16 characters (which is the size of IPv4 address) depending on the size of IP address string for edge switch to which N_Port is attached. Fabric OS v5.2.1 supports IPv4, not IPv6. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1

DEFECT000079455 Medium Summary: SilkWorm 4020: The "Restore Factory Default" operation leaves ports disabled if the Access Gateway mode is enabled. Symptom: For a switch that leaves the factory configured as a non-AG mode system, if subsequently AG mode is enabled, the "Restore Factory Defaults" operation does not completely restore the configuration. In particular the ports will be disabled. For a switch that leaves the factory with Access Gateway enabled, the "Restore Factory Defaults" operation will leave all ports disabled and the command "ag --show" will show no mapping. Workaround: If the factory default configuration is with the Access Gateway mode enabled, after performing the Restore Factory Defaults operation, run the "ag --modedisable" command. After the reboot, run "ag --modenable". If the factory default configuration is with the Access Gateway mode not enabled, do the opposite, i.e., "ag --modeenable" followed by "ag --modedisable". Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Medium Reported in Release: V5.2.1



Defect ID Severity Description DEFECT000079551 Medium Summary: IOF: Extra translated domain is created when there is no LSAN zone

configured between 2 fabrics after switchdisable/enable on all switches in the fabric. Symptom: Extra translate domain shows up when all the switches in all the edge fabrics are disabled and enabled one after another during overnight stress test. There is no other functional impact observed on the fabric. Workaround: Switchdisable, wait a minute, then switchenable on routers in the backbone fabric. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.1

DEFECT000079553 Medium Summary: switchdisable and switchenable on switch that directly connects to storage will cause the storage to see GLIEN reject error message. Symptom: GLIEN reject error message. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Medium Reported in Release: V5.2.1

DEFECT000079569 Medium Summary: Port GE0 went down and recovered after 5 seconds: no loss of data and no affect on iSCSI traffic. Symptom: Customer sees approximately a 5-second delay when the iSCSI port comes Online. Customer sees raslog entries indicating iSCSI port goes down and subsequently initializes. iSCSI hosts do not see any loss of operability (traffic is unaffected). Workaround: No workaround required. Service is established without operator intervention. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.1



Defect ID Severity Description DEFECT000079747 Medium Summary: SilkWorm 4100: F_Port count is incorrect in agshow --name <AG>

Symptom: User will observe that F_Port count is incorrect in agshow --name <AG>. Workaround: Make sure N_Ports of Access Gateway are connected to switches in a single FC fabric Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1

DEFECT000079968 Medium Summary: CLI: iscsicfg --show/delete ddset -n xxx complains "Requested object doesn't exist" even though that object does exist. Symptom: User could not use iscsicfg --show ddset -n or iscsicfg --show dd -d or iscsicfg --delete ddset -n, ... on those DD/DDSETs received from Microsoft iSNS server. Solution: iSNS client to convert DD and DDSet name to lowercase. Workaround: Use iscsicfg --clear dd for clear ddset and dd. Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.1

DEFECT000079971 Medium Summary: IOF: a few extra xlate domains with a missing ipaddress appear in the fabricshow results for an edge fabric after a disable/enable of all switches in one backbone. Symptom: When all the switches in a backbone fabric of a dual backbone FCR are disabled and enabled, the fabricshow will still show the missing translate domains. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.1



Defect ID Severity Description DEFECT000080024 Medium Summary: If the failover policy is disabled, failover should not take place

during either physical switch reset or reboot. Symptom: Failover will be triggered even when the N_Port failover policy for that N_Port is disabled if it fails to come Online after reboot. Workaround: Always make sure that the N_Ports to which F_Ports on Access Gateway are mapped are Online, i.e., by ensuring they have a POD license, NPIV configured on the fabric side, etc. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.1

DEFECT000080033 Medium Summary: Disabling F-ports due to N-ports not being available keeps changing the port state from "No_Sync" to "In_Sync". Symptom: Status of F_Port keeps changing from No_Sync to In_Sync or In_Sync to No_Sync in output of "switchshow" CLI when N_Port to which F_Port is mapped is Offline. Workaround: Ensure that the N_Port to which the F_Port is mapped is Online. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.1

DEFECT000080439 Medium Summary: iSCSI: During 1024 session testing, one of the ports lost multiple sessions after approximately 4 hours of traffic run Symptom: switchshow -iscsi shows less than 64 sessions on the port, but the iSCSI initiator will not be able to log in to the target (out of resources error). Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Medium Reported in Release: V5.2.1



Defect ID Severity Description DEFECT000081855 Medium Summary: Run from the CLI, diagclearerror causes the console to hang or the

system to crash. This affects only the SilkWorm 4016, SilkWorm 4018, and SilkWorm 4024 platforms. Symptom: When run as a separate command from the CLI, diagclearerror might hang or cause the system to crash. Workaround: To run diagclearerror from the CLI, first switch directories as shown here: cd /fabos/sbin diagclearerror Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: High Reported in Release: V5.2.1


Resolved Defects in Fabric OS v5.2.1 This section lists defects that have been closed since the last Fabric OS GA release, v5.2.0.

Defects Resolved In Fabric OS v5.2.1

Defect ID Severity Description

DEFECT000080217 Critical Summary: SilkWorm 48000 running FICON with fsmode enabled panics after an upgrade from Fabric OS v5.0.x to v5.1.x or v5.2.0. Symptom: Problem occurred on a SilkWorm 48000 with FICON fsmode enabled, dynamic load sharing (DLS) on, and with a port based route policy. After the upgrade and a disable/enable on a port to trigger a reroute, the switch panics with the following message displayed on the console: "ASSERT - Failed expression: !old_path, file = rte_path.c, line = 1366;" Solution: Fabric OS v5.1 introduced a new flag during route calculation. After an upgrade from Fabric OS v5.0 to v5.1, a new active CP running Fabric OS v5.1 uses the old way of calculating the route, rather than using the new flag. This causes a route miscalculation and the switch panics. Workaround: Avoid the combination of "Port based routing (external = 1, internal = 2) + DLS ON" in a FICON fsmode enabled setup. DLS off is recommended for a FICON environment. Service Request# RQST00000056030 Reported in Release: V5.1.0

DEFECT000052600 High Summary: System status LEDs become unstable with SilkWorm 24000 - blade status LEDs blinking from previous diag failure not cleared even after switch is enabled. Symptom: The system status LEDs may blink, indicating a failure when executing some diag commands, even though the slotshow command shows all slots enabled and functioning properly. This can occur when a previous diag returns a failure and the log is not cleared. Workaround: Do not abort diag test or Run "diagclearerror -all" to clear the condition. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Service Request# RQST00000035033 Reported in Release: V4.4.0




DEFECT000065369 High Summary: SilkWorm 4016 port speed changed when port is empty Symptom: Port speed is changed from N4 to AN when a port is empty. This only applies to the SilkWorm 4016. Solution: Filters chassis noise in the port speed display. Probability: Medium Reported in Release: V5.0.4

DEFECT000069240 High Summary: FR4-18i becomes "FAULTY" after FCR related stress tests and slotpower off/on. Symptom: POST failed. 2006/05/12-18:40:08, [DGD-5003], 8023/0,, ERROR, SilkWorm48000, Slot 8 has failed the POST tests. FRU is being faulted., main.c, line: 1824 Workaround: Slotpoweroff/on. Probability: Low Reported in Release: V5.1.0

DEFECT000070545 High Summary: hafailover after a few diagnostic tests have been run results in error messages. Symptom: This warning message "VERIFY - Failed expression: 0, file = fab_ha.c, line = 4610, user mode args = 0, 128, 1, 0, 0, 0, 0, 16417, 1552, 4294967295" comes on when lot of diagnostic tests are run and the consistancy check between the fabric and kernel fails. This happens during the warm recovery on standby. Solution: This happens due to a consistency check and a possible solution is to eliminate the check. or enter a small code change to make sure the check is done at the appropriate place. Customer Impact: This problem has been observed only once and several attempts to reproduce it failed. Probability: Medium Reported in Release: V5.2.0




DEFECT000071142 High Summary: In rare cases, the zoning daemon (zoned) asserts and causes the switch to panic during HAfailover. Symptom: In these rare cases, if the switch panics during HAfailover, the console log displays "do_assert (expression=0x10062ae8 "ns_zt[port].ns_zonetype != 0", file_name=0x0, line_num=970) at dbg_assert.c:268." Solution: Properly handles the error condition during the zone type recover. Probability: Low Service Request# RQST00000049244 Reported in Release: V5.0.3

DEFECT000071486 High Summary: Observed condor-module panic on new active CP after hafailover with running I/O through AD, Mirrored port, FC4-16IP and FC4-48. Symptom: Observed panic on the switch after hafailover Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000072181 High Summary: An hafailover on an SilkWorm 48000 causes kernel panic and error message. Symptom: An hafailover on a SilkWorm 48000 causes Software Fault:Kernel Panic with the message of "kernel BUG at cachemap.c:91!Exception in kernel mode: sig=4" Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Medium Reported in Release: V5.2.0




DEFECT000072370 High Summary: Switch panic during power on while running diagnostic test. Symptom: The following error messages displayed after power on of a SilkWorm 3014. The code fix applies to all platforms. "?[PDTR-1001], 921,, INFO, ?,pdcheck: info: found new pd: mtd_ts=1151588734 DIE Thu Jun 29 13:45:34 2006, cf_ts=0" and "?[HAM-1004], 922,, INFO, SilkWorm3014, Processor rebooted - Software Fault:Kernel Panic" Solution: Enhanced CAM diagnostic test code to allow CPU sharing so other processes get a chance to run. Service Request# RQST00000049809 Reported in Release: V5.0.1

DEFECT000072552 High Summary: Fastboot on remote switch followed by hafailover on SilkWorm 48000 causing multiple kernel panics. Symptom: Error message: "kernel BUG at cachemap.c:91!, kernel BUG at mtracer.c:521!" Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000073418 High Summary: Too many Fabric Watch email alert messages caused unexpected termination of other processes as switch runs out of memory. Symptom: When using the continuous mode in Fabric Watch with a timing interval of 1 second, a large number of Fabric Watch alert INFO emails hold chunks of memory and cause other processes to panic. Error messages similar to: "memorywt_pool_get: pthread_create: Cannot allocate memory on console," In addition, "sendmail" processes flood the switch. This typically happens when using Fabric Watch to monitor port performance. Solution: Fixes code to reduce the number of Fabric Watch email messages queued in the system. Workaround: Set up Fabric Watch email address properly and disable email alerts in Fabric Watch class on console monitor. If you do not use the Fabric Watch send mail alert feature, disable all alerts with fwalarmsfilterset. Customer Impact: Minimal, since there is a workaround. Probability: Low Service Request# RQST00000050466 Reported in Release: V5.0.3




DEFECT000073777 High Summary: Sometimes the SNMP configuration set via the agtcfgset command is lost when upgrading from one firmware version to another. Symptom: Some of the SNMP configurations (such as Trap recipient) are initialized (back to default) when upgrading the firmware from one version to another. Customer Impact: This problem has been observed only once and several attempts to reproduce it failed. Probability: Low Service Request# RQST00000049349 Reported in Release: V4.4.0

DEFECT000073835 High Summary: Out of memory messages seen on serial console during multiple scripted logins (command testing). Symptom: A script that makes heavy use of login changes to all supportable login roles stops functioning after several minutes and various error messages are seen in the telnet session. Solution: duplicate of 76913, see that defect. Workaround: Refrain from issuing new logins from inside an existing session. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Medium Reported in Release: V5.2.0

DEFECT000074048 High Summary: ISCSI: Target lost when using persistent reservation on two connections Symptom: The host cannot access its target when using persistent reservation in the redundant two connection configuration. Workaround: 1. Have both tcp connections on the same iSCSI portal. 2. Avoid having reservations from different hosts sharing the same iSCSI portal and the same iSCSI virtual target Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0




DEFECT000074848 High Summary: After a non-disruptive firmwaredownload on a SilkWorm 3900 in a large fabric (2560 ports), the core switch experienced an E_port offline transition and caused a fabric reconfiguration. Symptom: The firmware download operation on that switch will be disruptive. Workaround: Use a zone cfgsize smaller than 256kb on SilkWorm 3900, 3250, or3850 Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.0

DEFECT000075091 High Summary: Running portshow command in a loop causes telnet to hang. Symptom: Running portshow continuously will hang the telnet session. Solution: Drop any update from the active cp to the standby cp if the fspf neighbor database is in any transaction state(not init yet. value will be -1). Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Service Request# RQST00000047507 Reported in Release: V5.0.1

DEFECT000075126 High Summary: During slotpoweroff/slotpoweron CLI test, an error message is displayed, then an FC4-16IP blade slot is faulted. Symptom: Error message: "boot failure faulting blade.." and an FC4-16IP blade generates a faulty (21). Solution: The change applies to the FC4-16IP blade, to set the boot_status register to BS_LOAD_OS (0xB0) only after successfully loading and validating the kernel image from CF. Probability: Low Reported in Release: V5.2.0




DEFECT000075449 High Summary: Switch panic or CP failover when a kernel process is spawned by a non-root user. Symptom: "free_uid()" is on the console log backtrace. This problem applies to all platforms. Solution: Applies a Linux kernel patch to allow correct tracking of the root process. Probability: Low Service Request# RQST00000051415 Reported in Release: V5.0.3

DEFECT000075526 High Summary: All backup jobs sent to the 2Gb 3rd party tape library over a 200ms delayed link failed. Symptom: All backup jobs initiated from Symantec BackupExec to a 2Gb 3rd party tape library over a 200ms delayed link and under high loss conditions (> 1%) failed. Workaround: The issue will show up with higher latency (200ms RTT) along with very high loss environments (more than 1%). The normal operating latency will be max of 200ms RTT and loss will be < 1%. The workaround is to enable Tape Pipelining on lower latency (less than 200ms RTT) and low loss environment (< 1%). Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000075874 High Summary: In interop mode, no registered state change notification (RSCN) sent to hosts after zoning changes using cfgEnable. Symptom: In interop mode, following cfgEnable, switch does not generate RSCN to host(s). As a result, the host(s) do not PLOGI, nor do they discover target(s). Solution: Updates the nameserver code in interop mode to send out the RSCN after the zone change. Workaround: Perform portdisable/portenable on host port after cfgEnable (only if in interop mode). Customer Impact: Minimal, since there is a workaround. Probability: High Service Request# RQST00000051734 Reported in Release: V5.1.0




DEFECT000076137 High Summary: I/O Ops is detecting Unit Checks if PDCM CUP Port x'FE' self prohibit bit is set Sense code =x'29CA' Symptom: Unit checks running I/O Ops. Solution: This defect is already fixed in the 5.2.0 code base. The fix was not verified and closed until after OEM qual. Probability: Medium Reported in Release: V5.1.0

DEFECT000076245 High Summary: FC4-16IP blade came up in Faulty(51) state after firmwaredownload and reboot Symptom: After upgrading, one out of four FC4-16IP blades in the chassis was in Faulty(51) state Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Reported in Release: V5.2.0

DEFECT000076352 High Summary: During AD domain ID change test, CAM entries were not set up on F-Port. Symptom: After manually changing the switch domain ID, then changing it back, the CAM entries for a (D,P) zoned device in AD might not be properly installed, halting traffic. Workaround: None Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0




DEFECT000076440 High Summary: The Web Tools GUI hangs upon invoking fabric wide polling when there is no IP connection between the Web Tools proxy switch and the other switches in the fabric. Symptom: On a large fabric (e.g., 30+ embedded switches), the GUI hangs upon launch and no access via Web Tools is possible. This impacts all platforms if the proxy has no IP connection to a large number of switches in the fabric. It may be more apparent after an upgrade from Fabric OS v4.x to v5.0.x if there are switches in the fabric previously unrecognized by Fabric OS v4.x, and then after the upgrade is newly recognized by Fabric OS 5.0.x. These switches are on a different subnet from the proxy and without an IP connection to the proxy. Solution: Support the enable/disable to check the value line fabric license and the fabric event, which use the switch proxy mode to get all info from the switches in fabric. Service Request# RQST00000049863 Reported in Release: V5.0.3

DEFECT000076551 High Summary: Without displaying an error message, a hardware ASIC failure causes performance degradation on a SilkWorm 48000, or a switch fault on other 4Gb switches. Symptom: Porterrshow indicates no errors on F_ports, though the ISLs on the remote switch show a large number of frame errors. Performance drops dramatically without traffic failover to the redundant data path. Solution: Added external critical RASLOG CDR-1003 message to identify the failed slot and ASIC chip. For Fabric OS v5.2.1and later, also provided option to allow customer to shut down the problem port blade or CP blade on a Silkworm 48000, through a persistent configurable parameter under system/bladeFaultOnHwErrlevel. Note: the CP shutdown will only mean that the dual core path will be reduced to a single core path for all slot cards, and HA will remain unchanged. Workaround: Manually shut down the blade when the HW ASIC error occurs, or configure it to allow autoshutdown. Default configuration is for manual shutdown. Service Request# RQST00000052235 Reported in Release: V5.0.3




DEFECT000076560 High Summary: Error message displayed after upgrading the firmware on a local SilkWorm 48000 and the FR4-18i blade in the SilkWorm 48000 on the remote side of an FCIP tunnel is faulted. Symptom: Error message: "CP-4350 heartbeat dead" and FR4-18i blade generates faulty (21) message. Solution: Increased delay from 2 to 4 seconds between the blade daemon and an IPSEC daemon. Probability: Medium Reported in Release: V5.2.0

DEFECT000076571 High Summary: After two hrs of routing stress test, an FR4-18i port blade in one SilkWorm 48000 is faulted and an FC4-32 port blade is faulted another SilkWorm 48000. Symptom: An FR4-18i port blade in one SilkWorm 48000 is faulted with error code (FAULTY 21) and in another SilkWorm 48000 an FC4-32 port blade is faulted with error code (FAULTY 3) after running a stress-to-fail routing test. Workaround: slotpoweroff and slotpoweron the faulty port blades Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.0

DEFECT000076589 High Summary: Running a script that is repeatedly adding and removing lsan zones failed when an lsan zone is added to an edge fabric and the zoning config does not propagate to both FCRs. Symptom: LSAN zones not equal across multiple FCRs (SilkWorm 7500 or SilkWorm 48000 with an FR4-18i blade) in backbone. Devices not imported as expected. Solution: Set appropriate flag when an LSAN zone is created so the zone will be propagated across the backbone fabric. Workaround: Workaround 1 - Perform a zoneadd or zoneremove in conjuction with a zonecreate. Workaround 2 - Disable/enable ports connecting into the backbone fabric or disable/enable ports connecting into the edge fabric. Probability: Medium Reported in Release: V5.2.0




DEFECT000076599 High Summary: FR4-18i blade went faulty (62) after firmware download Symptom: One of two FR4-18i blades in a SilkWorm 48000 switch went faulty (62) after firmware download using firmwaredownload -sf and reboot -f on both CPs simultaneously. Workaround: Slotpoweroff/slotpoweron Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000076603 High Summary: FR4-18i displays continuous Device Off messages when devices are intact. Symptom: If all FCR events messages are turned on, one may see FCR-1031 event Device Off message continuously on console. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000076639 High Summary: ISCSI: FC4-16IP blade running 1024 sessions stops responding after running for three days Symptom: An FC4-16IP blade stopped responding to commands after running stress-test traffic for three days. Solution: Fixes enforcement API to free memory for each target in the target list returned by DB when DDSET is enabled. Probability: Medium Reported in Release: V5.2.0




DEFECT000076674 High Summary: An FR4-18i blade reported as faulty (21) when portdisable and portenable were performed on a target port on an adjacent SilkWorm 7500. Symptom: Portdisabled and portenabled a 2Gb JBOD port on a SilkWorm 7500, which is connected to an FR4-18i blade over a VE-VE port, and observed the blade went faulty (21). Workaround: Slotpoweroff/on. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000076688 High Summary: ISCSI: unable to disable iSNS after peering with iSNS server Symptom: The error "An outstanding transaction on switch" is reported when disabling the iSNS client with the command fosconfig --disable isnsc. Solution: When any DD/DDSet operation fails, it may leave an outstanding transaction that was started by the iSNS client. The fix is to abort any outstanding transaction when any DD/DDSet operation fails. Workaround: Step One: display the outstanding transaction: iscsicfg --show transaction Step Two: abort transaction that is displayed in Step 1 above: iscsicfg --abort transaction -x <transaction id> Probability: Medium Reported in Release: V5.2.0

DEFECT000076747 High Summary: ISCSI: host loses target with any IO when port is set to Error Recovery Level 2 (ERL2), digests enabled. Symptom: When port configuration is set to ERL2 with digests enabled, the host loses its targets. Solution: Write data was sent to the target before the digest calculation ended, creating buffer pointer mismatch. The fix disables digest when immediate data is enabled. Workaround: From initiators, reconnect to targets. Probability: Low Reported in Release: V5.2.0




DEFECT000077099 High Summary: Using the tape pipelining feature, accessing data spanning on a 1Gb tape library causes the backup application to remain stuck in running state. Symptom: Using the tape pipelining feature, backing up a set of data whose size is larger than the available space on a tape may cause the backup application to get stuck in the running state. This was observed with a 1Gb tape library; however, the faster 2Gb tape library successfully spans the data across two tapes. Solution: The fix checks for an early end of medium status and pauses the tape pipelining operation until it encounters the write file marker. Probability: Medium Reported in Release: V5.2.0

DEFECT000077202 High Summary: Blade fault on bloom based platforms such as SilkWorm 24000, 12000, 3900, and 38x0. Symptom: Blade faults with data dumped in /core_file/fdet directory and with raslogs: [BLL-1000], 1114,, CRITICAL, ? , ASIC driver detected Slot 1 port 8 as faulty (reason: 13) and [BLL-1000], 1962/1743,, CRITICAL, ?, ASIC driver detected Slot 0 port 21 as faulty (reason: 9) Solution: Additional enhancement added to prevent two types of Fdet: TXQ# error and Mini-buffer state checking error. Service Request# RQST00000052999 Reported in Release: V5.0.1

DEFECT000077219 High Summary: Switch panics due to a bad ipaddress in the syslog.conf file. This affects all platforms. Symptom: If for any reason an invalid ipaddress makes it into the syslog.conf file using the root level "echo" command, the raslogd init fails and causes the switch to panic. Note, This is not likely to happen in production because all user/admin interface commands already validate the ipaddress before adding it to the syslog.conf file. If it does occur, the error message is: [TRCE-1001], 4984,, WARNING, SilkWorm48000, Trace dump available (Slot 6)! (reason: PANIC). Solution: The fix changes the code to take precautions and handle the problem gracefully. Service Request# RQST00000052993 Reported in Release: V5.0.3




DEFECT000077274 High Summary: CUP Port fails to come online on a SilkWorm 7500 due to F-Rej to PLOGI to the CUP Port Symptom: CUP Port fails to come online on a SilkWorm 7500 in a FICON environment Solution: Path to embedded destination for CUP Port was missing. This was because the policy on the embedded destination node was being set to (int: dev-based, ext: port-based), forcing the path calculation to use the virtual graph. This was because the original requested policy for the destination node (int: none, ext: port-based) was missing from the supported policy capabilities for the SilkWorm 7500 platform. Adding this combination as one of the supported policies solved the problem. Probability: Medium Reported in Release: V5.2.0

DEFECT000077278 High Summary: In FICON environment, switch panics with FICON CUP daemon. Symptom: Switch will panic, indicating the following error: "kSWD: Detected unexpected termination of: ''[5]ficud:0'RfP=713" Solution: Fix code to handle zero byte payload for Chain Commands sent by FICON Channel with Command Response Request enabled. Probability: Medium Service Request# RQST00000053078 Reported in Release: V5.1.0

DEFECT000077317 High Summary: When using the SMIAgent to connect to a switch in a larger fabric containing more than 383 FDMI devices, the LoginAsUser fails and switch panics. Symptom: The switch will either reboot or HA failover depending on the platform type. Solution: The fix is to prevent a buffer overflow in an FDMI interface function. The buffer overflow corrupts the header of the following malloc'd buffer. This results in a core dump when the second buffer is freed. This problem will happen only when there are more than 383 FDMI HBA entries in the fabric. Probability: Medium Reported in Release: V5.2.0




DEFECT000077478 High Summary: Linux 2.6 iSCSI initiator is not able to discover a third-party storage array. Symptom: The initiator was disconnecting the session and reconnecting. Any target which sends out sense data (sense len % 4 == 0) will experience the problem. Solution: Add pad bytes to the response PDU to ensure it is word aligned. Probability: Medium Reported in Release: V5.2.0

DEFECT000077479 High Summary: iSCSI: Cannot fail back after performing a Path Failover with certain third-party multi-pathing software. It works with iSCSI GE ports, but not on FC. Symptom: Disruption in LAN: - Host logs in again to the target after every time2wait value (20 seconds). Disruption in SAN: - There is no disruption, as the virtual target is marked offline. Solution: Return "target is not currently operational" after a normal login to a VT that is in an offline state so the initiator can retry the login. Workaround: Log in again to the target when it comes back online. Probability: High Reported in Release: V5.2.0

DEFECT000077841 High Summary: Removing a switch from a fabric causes trace files on remote switches to generate error message. Symptom: Error messages displayed: " EVMD-1001:Event could not be sent to remote proxy. =fffc61" with FFDC triggered files. This only applies to switches running FOS v5.2.0x. Solution: When the fabric is reconfigured or the proxy goes offline, if the switch cannot send the event to the proxy, the EVMD logs an RASLOG event 1001 and unintentionally triggers the FFDC. The fix removes the FFDC trigger for this case. Service Request# RQST00000053962 Reported in Release: V5.2.0




DEFECT000078007 High Summary: Switch panics with remote procedure call daemon (rpcd) during process link incident record (LIRR). Symptom: Switch panic with error message: "[KSWD-1003], 1869,, WARNING, Silkworm24000, kSWD: Detected unexpected termination of: ''[11]rpcd:" Solution: Fix management server code to properly initialize data structure after allocation when there is no link incident record registration (LIRR) to avoid sending un-initialized data to rpcd and causing it to panic later. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Service Request# RQST00000054206 Reported in Release: V5.0.3

DEFECT000078296 High Summary: Registration to a storage array by a specific third party storage management agent fails with CRC error on the FC side. Symptom: Registration to array fails, preventing the host from discovering its devices in the storage group. Solution: This happens when write data is not 4-byte aligned. The fill bytes are added to the FC header, but total number of bytes to send does not include the fill bytes. This leads to CRC error detected on the FC side. The fix is to include fill bytes in total number of bytes to send to FC target. Probability: Medium Service Request# RQST00000054696 Reported in Release: V5.2.0




DEFECT000078298 High Summary: Third-party storage array adapter disengages when running I/O between iSCSI host and certain third-party storage array via FC4-16IP (iSCSI to FC bridge) with iSCSI Header and Data Digest enabled Symptom: Array changes into DD (Dead Director) mode when running IO with header/data digest enabled. Solution: Copy packets from iSCSI initiator to the previous buffer if it has room and send to FC target with minimum of 512 byte in each frame. Workaround: Disable header/data digest. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Medium Service Request# RQST00000054700 Reported in Release: V5.2.0

DEFECT000078456 High Summary: After an upgrade to Fabric OS v5.0.5x, dual CP switches show excessive TRCK-1004 logs though there is no change to the configuration. Symptom: Excessive raslog messages similar to "2006/11/22-21:29:38, [TRCK-1004], 42700,, INFO, SW12K_Demo1, Config file change from task:PDMIPC." There is no functional repercussion beyond the excessive raslog and syslog entries. Solution: Whether or not there are any changes to the configuration, the standby CP performs a "config update." This results in a change in the configuration file. If change tracking is enabled, the PDM (code that replicates files between CPs) sends a CONFIG_CHANGE message to the change tracking daemon. Every config update results in a raslog being generated. The fix changes the code to eliminate the config update when there is no actual configuration change. Service Request# RQST00000054539 Reported in Release: V5.0.1

DEFECT000078518 High Summary: After upgrading from one internal build to another, GE ports are in a loop and reinitializing. This is already adddressed in GA build. Symptom: Upgraded ED-48K from 5.2.0b bld04 to 5.2.0b bld05 and GigE ports have become unstable and are crashing and reinitializing. Service Request# RQST00000054950 Reported in Release: V5.2.0




DEFECT000078871 High Summary: After connecting a SilkWorm 7500 to a 3rd party switch fabric in native mode with a large zone DB- the FCR daemon (iswtichd) panicked. Symptom: Error message: "RASLOG 2006/12/19-18:02:26, [KSWD-1003], 250, FFDC, WARNING, sw75BO015, kSWD: Detected unexpected termination of: ''[28]iswitchd" Solution: Fix code to increase memory allocation to accommodate large zone database in native mode. Workaround: Limit zone names to fewer than 12800 characters, or use 3rd party open mode option. Customer Impact: Minimal, since there is a workaround. Probability: Medium Service Request# RQST00000055219 Reported in Release: V5.2.0

DEFECT000078988 High Summary: iSCSI traffic through an FC4-16IP blade has low throughput between iSCSI initiator and certain third-party storage array. Symptom: iSCSI traffic through an FC4-16IP blade has noticeably low throughput. Solution: When sending write data to FC target, check and only send minimum of 2KB or remaining bytes of the Xfer Ready. Workaround: Disable immediate data. Probability: Low Service Request# RQST00000055325 Reported in Release: V5.2.0

DEFECT000079588 High Summary: Name Server daemon crash and switch panic after a failed attempt to enable a zone configuration (D,P format). Symptom: The zone enforcement fails and the name server panics. Solution: During zone change, allocate a large enough buffer to accommodate large zone sets. Probability: Medium Reported in Release: V5.2.0




DEFECT000049275 Medium Summary: swAgtCmtyTable is empty when queried through an SNMPv3 client; when queried through SNMPv1, the table shows entries. Symptom: swAgtCmtyTable lists SNMPv1 community strings and trap targets (and their severity levels) configured at the switch. When queried using SNMPv3 client, swAgtCmtyTable shows no entries. When queried through SNMPv1, swAgtCmtyTable shows entries matching those shown by the agtcfgshow command. Workaround: The swAgtCmtyTable returns community strings and traps recipients when queried through SNMPv1. The same data is also available through the agtcfgshow command. Also, details of configured trap targets are available by querying trapRegTable (FA-MIB) using both SNMPv1 and SNMPv3 clients. Customer Impact: Minimal, since there is a workaround. Probability: High Reported in Release: V4.4.0

DEFECT000058298 Medium Summary: Switch error exception when trying to getInstance of NodeFCPort of FDMI enabled HBA attached to v3.2/v4.4.0d switch, when Fabric OS v4.4.0d or v5.0.1 switch is proxy. Symptom: The user will not get all the NodeFCPort data while trying to do getInstance of NodeFCPort in the fabric. Workaround: Issue portdisable/portenable on the switch port to which the device is attached. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V4.4.0




DEFECT000060321 Medium Summary: Part of the Switch Admin window is covered by log message text area. Symptom: When log display contains many messages, the window size covers most of the Web Tools display. Solution: Now allow the dividbar to go up to cover the screen. Workaround: Move the divide bar down and user will see the screen. No customer impact to switch operations. If the Web Tools components are covered, the user can reduce the size of the window with the resizing handles to reveal more of the Web Tools switch display. Customer Impact: No customer impact to switch operations. If the Web Tools components are covered, the user can reduce the size of the window with the resizing handles to reveal more of the WebTools switch display. Probability: Low Reported in Release: V5.0.2

DEFECT000063230 Medium Summary: Null Pointer exception at java console when clicked on SFP, Loop tabs under portinformation tab. Symptom: Java console will report a Null Pointer Exception under rare circumstances, but switch operations are unaffected Solution: Correct the error message Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.0.4

DEFECT000064319 Medium Summary: Using 3rd party chassis management software to "Restore Factory Defaults" can result in a loss of telnet connectivity. Symptom: Might lose ability to telnet. Solution: Fix creates a more robust design to handle IP address changes. Workaround: Go to the 3rd party chassis management software and click on "Save" button to save the network configuration. Customer Impact: Minimal, since there is a workaround. Probability: High Reported in Release: V5.0.3




DEFECT000065168 Medium Summary: On the SilkWorm embedded switch platforms, Fabric Watch erroneously shows internal ports as having SFP's. Symptom: Internal ports shown as having SFPs. Solution: The Fabric Watch SFP state report now accurately reflects that internal ports have no SFPs. Reported in Release: V5.0.4

DEFECT000067161 Medium Summary: SilkWorm 3014 overvoltage error: false error is reported to the log. Symptom: Error message: 2006/03/16-11:07:16, [HIL-1107], 188,, CRITICAL, SilkWorm3014, Switch faulted, 2.5V (3.29) above threshold. System preparing for reset. -- 2006/03/16-11:07:16, [EM-1041], 189,, WARNING, SilkWorm3014, Sensor values for Switch: Temperature 1 = 33 C Temperature 2 = 35 C Temperature 3 = 30 C 1.8V = 1.8 6 V 2.5V = 3.29 V 3.3V = 3.35 V 12V = 11.69 V -- 2006/03/16-11:07:16, [EM-1041], 190,, WARNING, SilkWorm3014, Sensor values for Switch: 1.2V Bloom = 1.19 V 3.3V Flash = 3.25 V Solution: Requires two consecutive readings before reporting the voltage error to the log. Probability: Low Service Request# RQST00000047143 Reported in Release: V5.0.1

DEFECT000067211 Medium Summary: Host in one edge fabric fails to see targets in another edge fabric when the routing policy is changed from exchange-based to port-based in the backbone (SilkWorm 48000). Workaround: Reboot. Probability: Medium Reported in Release: V5.1.0

DEFECT000067851 Medium Summary: FCIP Diag failed when external link went offline/online during POST. Symptom: If an external link goes offline/online during POST, the Diag program will fail. Solution: Correctly detects the link state change and an internal error so that the POST test passes. Workaround: Turn off POST Probability: High Reported in Release: V5.1.0




DEFECT000068562 Medium Summary: CPU port code is unable to achieve 2:1 throughput increase on FR4-18i or SilkWorm 7500, even when data is highly compressible. Symptom: Customer might notice that the FC side of the throughput doesn't increase at the ratio expected when GE compression is enabled. Solution: Fixed incorrect buffer return rate when compression is ON Probability: Medium Reported in Release: V5.1.0

DEFECT000069030 Medium Summary: Portshow is not displaying IPIF present in configDB for GE1 of a SilkWorm 7500. After using portcfg to configure the port, cannot configure IPIF. Error Message: "IP address already configured on Port ge1" Symptom: GE port configuration will not be validated during configdownload. Solution: Automated verification of GE port configuration will be added in a future release. Workaround: GE port configuration must be valid while applying the configuration to the ports. Any duplicate or invalid entry will not be reported, so manual verification is desirable. Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000069322 Medium Summary: During FCR 3rd party interop testing, ex_ports status does not change after enabling the ex_ports from Web Tools. Symptom: Web Tools will show the incorrect status for the ex_ports. Workaround: Use CLI to check the correct status Reported in Release: V5.1.0




DEFECT000069372 Medium Summary: Web Tools: Columns like Name, State, Reason etc, from Fabric Watch tree (Port class) are restored mistakenly after auto refresh. Symptom: End user tries to change the column according to his requirement, but the column will be restored after auto refresh. Customer Impact: Usability issue, with no impact to functionality. Probability: High Reported in Release: V5.0.5

DEFECT000069539 Medium Summary: CAM entries on a shared port should set up in conjunction with zoning on an existing AD. Currently, CAM entries are all the same across ADs. Symptom: Partial PIDs that are not part of the current AD are displayed in portcamshow output. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000071184 Medium Summary: Web Tools Events are showing wrong date when time zone is changed Symptom: Customer will see wrong date in Web Tools Events when timezone is changed till the switch is rebooted. Solution: Need a mechanism for processes starting at switch boot up to see the changed time zone. Workaround: Reboot switch after changing the timezone. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000071568 Medium Summary: core_files not being deleted by supportsave automatically Symptom: core_files will be left on the switch after the supportsave command is invoked Probability: High Reported in Release: V5.2.0




DEFECT000072123 Medium Summary: Several E-ports and unused ports on an FC4-48 port blade stuck at Mod_Val state, resulted from a slotpopweroff/on test on FC4-48 port blades after a triggered hafailover test. Symptom: If failover occurs while FC4-48 port blades are being slotpoweron, the user may see the problem that some E-ports and unused ports get stuck at Mod_Val state. Workaround: portdisable/enable the affected ports Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000072209 Medium Summary: OEM: see FFDC_data core dump generated in several switches Symptom: core file ffc_data is generated after disable/enable switch. Customer Impact: Usability issue, with no impact to functionality. Probability: High Reported in Release: V5.2.0

DEFECT000072716 Medium Summary: After configdownload and reboot, portshow displays FCIP tunnel for an IPIF that does not exist on the GE. Symptom: GE port configuration will not be validated during configdownload and any duplicate or invalid entry will not be reported. Workaround: GE port configuration must be valid while applying the configuration to the ports. Any duplicate or invalid entry will not be reported, so manual verification is desirable. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000072773 Medium Summary: "swNbEntry" under "swFabric" omits display of trunked EX-port ISLs. Symptom: No display of the trunked EX-port ISLs. Customer Impact: This problem has been observed only once and several attempts to reproduce it failed. Probability: Medium Reported in Release: V5.2.0




DEFECT000073103 Medium Summary: In a Mixed Fabric AD Context, EnumerateInstances/EnumerateInstanceNames/ExecQuery (Brocade_SwitchFcPort) is returning SwitchFCPort Instances of NON v5.2.0 FOS Switches. Symptom: In a mixed fabric(Fabric OS v5.2.0 and older switches), discovering SwitchPort returns all instances from pre-v5.2.0 FOS switches. Admin Domain filtering is not done for pre-v5.2.0 FOS switches. Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000073632 Medium Summary: supportsave: not marking all local tracedump files as FTPed consistently after supportsave transfer completed Symptom: The tracedump command normally shows status of tracedump transfers. It is reflecting supportsave transfers as "FTPed" for AP blades but not for active or standby CP tracedumps. This can be confusing to the user. Solution: Revisit the design of supportsave/Trace/ftp and multiple blades to implement in a clean, logical manner that covers all cases Workaround: check FTP files on host machine to see if files were ftp'd. Customer Impact: Minimal, since there is a workaround. Probability: High Reported in Release: V5.2.0

DEFECT000074023 Medium Summary: Porterrshow shows Slot and Port for chassis based systems. Symptom: No change to 5.1 porterrshow. It's just cosmetic. Solution: should show in slot/port format. Workaround: none. porterrshow is the same as 5.1 Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0




DEFECT000074542 Medium Summary: Since QuickLoop is not supported on Fabric OS v5.2, Web Tools should grey out buttons to create it. Symptom: Can't create quick loop zones from Web Tools. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000074562 Medium Summary: SilkWorm 7500 faults during power up. Symptom: Upon power cycle, switch fails to complete POST and faults. Reported in Release: V5.1.0

DEFECT000074744 Medium Summary: LongDistance: portenable does not enable the "Disabled (Insufficient frame buffers)" port even after portcfglongdistance command is issued to reduce the distance to get within enough buffer. Solution: Allow portenable to go through when there are sufficient buffers allocated for the port. Workaround: If customer has a port disabled due to insufficient buffers and customer changes long distance setting to allow sufficient buffers on port, the customer has to issue the following to enable the port: portdisable [slot]/<port> portenable [slot]/<port> Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000074755 Medium Summary: Upon deleting many user accounts (150) from Web Tools, Web Tools loses connection to switch Symptom: Web Tools may not respond to the user about the configuration result if the request contains a large number of deletions. Workaround: Limit deletions to a maximum of 50 accounts at one time. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0




DEFECT000074812 Medium Summary: ipperf reports higher % loss than expected over an impaired WAN link Symptom: If ipperf is running over the impaired WAN link (1% loss and ~250msec delay), ipperf will report higher (double/triple) loss rates. This kind of impairment is not that common. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: High Reported in Release: V5.2.0

DEFECT000074892 Medium Summary: Usability: High availability of transferred supportsave files is not ensured with current implementation Symptom: No direct customer symptom. Customer indirectly affected if required supportsave files were not confirmed as having been reliabily transferred to the customer FTP service. Solution: supportsave has checked the file size before FTP Workaround: Retry supportsave Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000074991 Medium Summary: Arbitrary VE ports go down and come up during FCIP IO with two tunnels of 500MB each and compression turned on. Symptom: VE port goes down and then comes up with two tunnels of 500MB each and compression turned ON. I/O is resumed when the VE port comes back up. Workaround: No workaround required because the tunnel should recover by itself and applications are generally capable of retransmitting. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0




DEFECT000075045 Medium Summary: Linux magic debug sequence "sysrq" needs to be disabled by default. Symptom: In case the keys received over serial console turn out to be magic sysrq keys, the switch may restart/kill tasks/umount disks. Solution: The solution is to make this sysrq module part of the kernel on the switch but disable it so that even if the switch receives the sysrq keys, no action is taken. If the user wants to use it, it can always be enabled from the command prompt. Probability: Medium Service Request# RQST00000051340 Reported in Release: V5.1.0

DEFECT000075536 Medium Summary: The sticky xlate domain was not persistent over hafailover Symptom: If ports are disabled from the edge fabric, the unused bb xlate domain is preserved to prevent disruption in the bb fabric. However, after HA failover the BB xlate domain is not preserved, potentially causing fabric reconfiguration in the backbone. Workaround: No workaround because the sticky xlate domain is an unused xlate domain. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Medium Reported in Release: V5.2.0

DEFECT000075776 Medium Summary: On a SilkWorm 48000, error message after setting the switch speed to 1 Gbps. Symptom: After the switch speed was changed from Auto-Negotiate to 1 Gbps on a SilkWorm 48000 the following was displayed: "Software Error:ASSERT - Failed expression: (data != NULL) && (iuRx != NULL), file = fc/fcop.c, line = 209, user mode args = 0, 26940580" on the active CP's console log. Customer Impact: This problem has been observed only once and several attempts to reproduce it failed. Probability: Low Reported in Release: V5.2.0




DEFECT000075805 Medium Summary: Using powerslotoff command to turn off a blade, switch view in Web Tools shows the slot is vacant but the temperature status still displays the temperature of the blade. Symptom: Temperature status still displays for a blade that shows as vacant in Web Tools switch view. The blade is being turned off by the slotpweroff command. Customer Impact: Usability issue, with no impact to functionality. Probability: Low Service Request# RQST00000051670 Reported in Release: V5.0.5

DEFECT000075847 Medium Summary: Unexpected messages are seen when DHCP is turned ON/OFF Symptom: Unexpected messages (lines of Bad file descriptorgiot_process_event) are seen when DHCP is turned ON/OFF, however the DHCP gets enabled/disabled. Customer Impact: Usability issue, with no impact to functionality. Probability: High Reported in Release: V5.2.0

DEFECT000075881 Medium Summary: Hosts lost access to storage due to lack of buffer credits to transmit frames over a particular VC. Symptom: A particular flow of end-to-end traffic will be affected due to a lack of buffer credit on a particular VC over an ISL link, even if another end-to-end traffic flow using a different VC on this same ISL can pass traffic without any problem. A VC buffer credit is needed for one end of the switch to transmit frames to another end, and without buffer credit, the switch will not transmit any frames. Normally, this problem is most likely to occur over a very bad ISL link. Solution: Due to either to a bad physical link or a flaky connector, the ASIC will receive more VC_RDY than it should, or lose some VC_RDY on other VCs over the same ISL. As a result, some VC's will have zero buffer credit while other VC's will have more. In this case, if the total number of VC credits on an ISL exceeds the total number of credits initially configured for that ISL, the fix ensures the code will perform a link reset to recover. This way the end-to-end traffic can be recovered due to 0 credit on a particular VC, without the user having to manually bring down the link. A buffer credit on any VC over a Brocade E-port is required for transmitting FC frames. Workaround: Manually issue "portdisable <port#>" and then "portenable <port#>" in order to recover the lost buffer credits. Service Request# RQST00000051378 Reported in Release: V5.0.3




DEFECT000075927 Medium Summary: EX_Port missing switchname in switchshow for single 3800 switch edge fabric after BB bringup. Symptom: Customer may notice EX_Port trunks attached to edge switches running v3.2.0 FW fail to recover after fastboot of FCR switches. Solution: N/A Workaround: Issue a portdisable followed by portenable on all ports that are part of the trunk group attached to the edge switch. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000076071 Medium Summary: 'switchenable' should not enable a (persistently) disabled port in FICON mode. Symptom: Persistently disabled port will be enabled after switchenable in FMS mode. Solution: The fix ensures, on switchenable, that the port information is pulled from the FICON configuration file into the active configuration file. Persistently disabled ports will no longer be enabled after switchenable in FICON mode. Workaround: Selectively disable/enable ports instead of using switchenable to ensure that persistently disabled ports are not re-enabled. Alternatively, physically unplug the cable to prevent the port connection after re-enabling. Customer Impact: Minimal, since there is a workaround. Probability: High Reported in Release: V5.2.0

DEFECT000076167 Medium Summary: Warning dialogue appears after modifying an AD and neither saving it nor applying it; then switching the AD context from the switch explorer window; then closing the expired AD window. Symptom: Even if the user clicks the "Cancel" button, the user cannot go back and apply the change, and the earlier modification is lost. Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0




DEFECT000076203 Medium Summary: non-disruptive configdownload failed due to thresh.alarmsfilter:1 invalid value. Symptom: the non-disruptive configdownload may fail if an invalid threshold alarm filter value is used. Workaround: disable the switch and download the config parameters Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000076241 Medium Summary: On a SilkWorm 48000, moving the switch PID to 2, waiting for the fabric to stabilize and then issuing hafailover causes an assert on the new active CP. Symptom: Error Message: ASSERT - Failed expression: if_is_valid (ns_portifid[port]), file = nsDead pid: 667, process: nsd_util.c, line = flags: 1540" targets. hafailover causes, "VERIFY - Failed expression: if_is_valid (ns_portifid[port]), file = ns_util.c, line = 3113, user mode". Solution: The getLocalVPort() improperly equated the area with the port. This was broken for the displaced-PID format. The fix derives the proper association between the area and the port. In addition, the FC4-16IP blade driver was improperly setting the interface property (user port) when the displaced- PID format was used. The fix retains the constant area. Workaround: Prior to enabling the switch, issue slotpoweroff; sleep 5; slotpoweron for each slot containing an FC4-16IP blade. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Medium Reported in Release: V5.2.0




DEFECT000076242 Medium Summary: Redundant message about EX port trunking resulted from checking system settings for firmwaredownload: Fabric OS v5.2.0 to v.5.1.0d. Symptom: Error message: Checking system settings for firmwaredownload...Port configuration with EX ports enabled along with trunking for port(s) 112,113,114,115,120,121,122,123,124,125,126,127,200,201,202,203,204,205,206,207, use the portcfgexport, portcfgvexport, and/or portcfgtrunkport commands to disable the port configuration. Verify that the blade is ENABLED. (error 3) Solution: The solution is to suppress the redundant error message in the code. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000076281 Medium Summary: portperfshow does not register traffic for a VE tunnel while portstatsshow displays the count. FASTWRITE is not enabled. Symptom: The portperfshow command does not display traffic on a configured VE tunnel. However, issuing the portstatsshow <port> command displays the traffic. Workaround: Use portperfshow in the remote port's switch to get the data rate OR use portstatsshow to see increasing words/frames. Customer Impact: Cosmetic issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000076337 Medium Summary: Downloading a config file that contains only Fabric Watch parameters incorrectly fails when header and termination labels are absent. Symptom: t. Workaround: Don't discard the the header and end tag, then download. Customer Impact: Minimal, since there is a workaround. Probability: High Reported in Release: V5.2.0




DEFECT000076457 Medium Summary: After a Silkworm 7500 shut down, the other side of VE links (on a Silkworm 48000) still has stale/uncompleted info of VE links on switchshow. fabricshow still has some incomplete info from the Silkworm 7500. Symptom: If the Silkworm 7500 shuts down due to overheating, the FCIP tunnel will stay up. Switchshow on the other side of the tunnel will still show VE up. The fabricshow on the other side will also have information about the switch. Workaround: Manually power down the switch. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Low Reported in Release: V5.2.0

DEFECT000076494 Medium Summary: In User Defined AD [Device members]/AD0, inconsistencies observed for GiGE Port Details. Symptom: The customer will be able to get the FCIP tunnel Details if AD0 is default Home AD. Workaround: Work in AD255, AD[ All Ports/Switch as the member] Customer Impact: Minimal, since there is a workaround. Probability: Medium Reported in Release: V5.2.0

DEFECT000076534 Medium Summary: burninstatus returns script failure after blade is replaced with brand new blade in the slots. Error message includes "/fabos/link_sbin/burninstatus: : Missing the SN data: command not found /fabos/link_sbin/burninstatus: [: =: unary opera" Symptom: burninstatus on newly installed FC4-16IP may be unreadable and script errors may be returned. Workaround: Run systemverification/burnin test and burninstatus will be corrected. Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0




DEFECT000076586 Medium Summary: Seeing timeout error message when pathinfo CLI is used in interactive mode with extended stats, reverse path and strict source route. Symptom: When pathinfo CLI is used in interactive mode with extended stats, reverse path and strict source route. It returns an error message, "ERROR, SW4100_35,Failed to get lock within 1800 m.sec. to read the path information, pathinfo.c, line: 1671, comp". This will happen only when the hop count given is less than the total hop count required to reach the destination domain plus the reverse path. Solution: Fail the CLI with proper error message. Workaround: There are two workarounds for this-: 1) Use the CLI in non-interactive mode. 2)Use the CLI in interactive mode with the default hop count. Customer Impact: None if recommended service procedures are followed carefully. Probability: Low Reported in Release: V5.2.0

DEFECT000076600 Medium Summary: All ports on an FR4-18i blade do not get persistently disabled after swapping the FR4-18i blade and an FC4-16IP blade, followed by failover, then swapping them black, followed by another failover. Workaround: First bladedisable the affected FR4-18i blade, then configure the FC ports and FCIP tunnels as desired. Customer Impact: Should not occur under normal maintenance operation; represents an unlikely user scenario. Probability: Medium Reported in Release: V5.2.0

DEFECT000076631 Medium Summary: Disabling and enabling a trunk master on a SW48K causes a SW200E on the other side of the ISL to be stuck at Port_Flt Symptom: Running stress tests using a script in a High Port Count fabric's SW48K, issuing portenable and portdisable cli's on trunk masters, may cause Port_Flt's on attached SW200E's. Workaround: Reboot the SilkWorm 200E. Customer Impact: Should not occur under normal maintenance operation; resulted from stress-to-fail testing designed to push the limits of the switch and fabric to point of failure. Probability: Low Reported in Release: V5.2.0




DEFECT000076656 Medium Summary: If the domain ID of a switch changes, iSCSI initiators may not be able to access targets due to the Name Server database containing stale iSCSI host information. Symptom: iSCSI initiators not able to access targets. Solution: Switch driver F-port LOGO processing assumes the local Domain ID to be defined by the configured Domain ID. The fix is to remove this assumption and derive the pertinent PID from the SID of the LOGO frame. Notify Nameserver of the Virtual Initiator logout with UPDSCN_DEL_AREA rather than UPDSCN_UPD_AREA since the latter is not subject to synchronization problems. Workaround: Avoid domain ID changes. Probability: Low Reported in Release: V5.2.0

DEFECT000076660 Medium Summary: Web Tools : When user tries to configure the port speed through port admin services window for FC4-48 port, the port is also getting flagged as locked G_port when observed through portcfgshow CLI Customer Impact: Usability issue, with no impact to functionality. Probability: Medium Reported in Release: V5.2.0

DEFECT000076685 Medium Summary: ISCSI: a command does not exist to report current iSNS status Symptom: A command is needed to report to the user the current iSNS status. Solution: As a part of isnsccfg --show also display whether current operational status is connected to the isns server or not. Probability: Medium Reported in Release: V5.2.0




DEFECT000076696 Medium Summary: On a SilkWorm 48000, the rtedebug cli returns error: "REASON: ROUTE_ABSENT" for a device connected to FC4-48 port blade. Symptom: Following lengthy stress tests that included, but were not restricted to, setting the switch speed to 1, 2, 4, switchdisable, switchenable, and disabling and enabling trunking on the switch, users may see ROUTE_ABSENT for a device connected to an FC4-48 port blade when user issues an internal routing debug CLI (rtedebug) to verify the routing paths. Solution: During E_Port master change, the loopback route for the E_Port on an FC4-48 share port is added automatically, which is prohibited behavior for the FC4-48 shared E_Ports. The fix is to not allow this. Probability: Low Reported in Release: V5.2.0

DEFECT000076727 Medium Summary: After extended stress testing with repeated slotpoweroff/on on two FC4-16IP blades, the ICP daemon was stopped by the watchdog software (kSWD) and the switch CP failed over. Symptom: Performance server may crash if slotpoweron, slotpoweroff is issued continuously for over 11 hours on a switch containing an FC4-16IP blade. Solution: This fix will address the memory leak in PS when PS receives ROUTE_CHANGE SCN. If it fails to get the list of domains reachable, the previously allocated memory is not getting released. This causes approximately a 12K memory leak during slotpoweroff/slotpoweron tests. Probability: Low Reported in Release: V5.2.0

DEFECT000076752 Medium Summary: When multiple API clients try to get default zone mode at the same time, core files can be generated by the common access layer deamon (cald) on the switch. Symptom: API applications should not be affected as the cald application is automatically restarted. Notification of core file existence may be observed. Customer Impact: This problem has been observed only once and several attempts to reproduce it failed. Probability: Low Reported in Release: V5.2.0




DEFECT000076836 Medium Summary: ZET showing all the resources for an AD from a fabric in a particular scenario, even though the AD has only a few resources in it. Symptom: The particular view ("Port Zoning" view) displays the info when a user refreshes AD information from the fabric. AD membership information in the ZET is broken for the refresh operation on the "Port Zoning" view. Workaround: Use Mix Zoning View to manage Port Zoning. Customer Impact: Minimal, since there is a workaround. Probability: Low Reported in Release: V5.2.0

DEFECT000076931 Medium Summary: Inserting an unsupported FC4-48 blade into a pre-Fabric OS v5.2 chassis causes FAULTY(1) error that will not clear with blade removal. Symptom: HA sync will be lost and removal of the blade will not recover the HA sync. An active CP reboot is required to recover. Solution: Detects and faults the unknown blade, allowing HA to re-gain sync once the unknown blade is removed and the standby CP is rebooted. Workaround: Schedule window to perform reboot of the active CP. Probability: Medium Reported in Release: V5.0.5

DEFECT000077113 Medium Summary: When a host PC uses the Java-plugin 1.4.2_08, the event time filter does not work properly. Symptom: When the host PC uses the Java-plugin 1.4.2_08, the event time filter is filtering on the basis of AM time when PM time is provided in 'TO' field. Solution: The fix filters the events correctly when the TO time selected has a meridian of PM. Java 1.4 has a problem in the Calendar class in the set method. This problem is fixed in Java 1.5. The fix ensures the code works properly with the problem in Java 1.4. Workaround: Use the Java 1.5 plugin. Probability: Medium Service Request# RQST00000052925 Reported in Release: V5.2.0




DEFECT000077512 Medium Summary: SupportSave fails with illegal Option -s saving CORE_FFDC Symptom: Supportsave prints out error messages and fails to collect core file and FFDC data. Solution: Rescan switchname and remove none digital, alphabetical,- and _ characters before applying the switchname to savecore command. In addition, included fabric data distribution parameters in supportShow/Save. Workaround: Change the switch name to exclude space characters. Customer Impact: Usability issue, with no impact to functionality. Probability: Low Reported in Release: V5.2.0

DEFECT000077515 Medium Summary: rcsd panic on SilkWorm 4100 due to shared memory allocation failure. Symptom: rcsd running on a SilkWorm 4100 panicked during zone activation. Solution: The fix resets the shared memory size to zero to anticipate the conditions causing the rcsd panic. Probability: Medium Service Request# RQST00000053121 Reported in Release: V4.4.2

DEFECT000077616 Medium Summary: Ports 32-47 on an FC4-48 port blade come up as G-ports for certain third-party storage arrays. Symptom: An HBA that sends a previously assigned PID in the SID of the FLOGI on ports 32 - 47 on the FC4-48 port blade will come up as a G-port. Solution: The FC driver was masking out the arbitrated loop physical access (ALPA) when an F-port received an FLOGI with SID already assigned. Fix was not to mask out the ALPA and leave the ALPA base bit, which is 0x80 in this case. Workaround: Use ports 0 - 31 on the FC4-48 blade as a workaround if the HBA uses previously assigned PID in the SID of the FLOGI. Probability: Medium Service Request# RQST00000053630 Reported in Release: V5.2.0




DEFECT000078465 Medium Summary: Some operations initiated from Fabric Manager, Web Tools, and/or SMI-A will fail if the system switch is running Fabric OS v5.2.0a for more than 30 days.. Symptom: All common access layer (CAL) based asynchronous operations initiated from Fabric Manager. Web Tools, and SMI interfaces will fail in this situation. Fabric Manager operations that fail include: SupportSave, TraceRoute, & FICON Setup/Merge wizards. SMI operations that fail includes ACL configuration. Web Tools operations that fail include all configuration related functions for AD, ACL and iSCSI. Solution: The fix prevents the generation of invalid transaction IDs. Workaround: Reboot the switch or run the operation from the CLI rather than the management tool. Reported in Release: V5.2.0

brocade fabric os v5.2 - fujitsu.com pki certificates ... the section also includes the updated...

Documents