HP 316095-B21 - StorageWorks Edge Switch 2/24 Instruction Manual

HP StorageWorks
HA-Fabric Manager release notes
Part number: AA–RUR6J–TE/958–000288–013
Eighth edition: (October 2006)

Description
These release notes contain late-breaking and supplemental information for the HA-Fabric Manager
(HAFM) 08.09.01 software.
These release notes are intended for customers who purchased HAFM software.
This information is available and periodically updated on the HP web site
http://h18006.www1.hp.com/storage/saninfrastructure/index.html.
Update recommendation
Routine
Devices supported
For the latest supported devices, see the HP StorageWorks SAN design guide,availableat
http://h18006.www1.hp.com/storage/saninfrastructure.html.
Product models
This section describes the supported switches and HAFM software.
Switches
This section describes Director and Edge Switch product model support.
Table 1 Supported switches, firmware, and software
HP name McDATA name
HP StorageWorks Edge Switch 2/12 Sphereon™ 4300 Fabric Switch
(first and second-generation)
HP StorageWorks Edge Switch 2/16 Sphereon 3216 Fabric Switch
HP StorageWorks Edge Switch 2/24 Sphereon 4500 Fabric Switch
(first and second-generation)
HP StorageWorks Edge Switch 2/32 Sphereon 3232 Fabric Switch
HP StorageWorks Director 2/64 Intrepid™ 6064 Director
HP StorageWorks Director 2/140 Intrepid 6140 Director
Surestore Director FC-64
StorageWorks SAN Director 64
ED-6064 Director
N/A Sphereon 4400 Fabric Switch
N/A Sphereon 4700 Fabric Switch
HAFM Basic (formerly, Embedded Web
Server or EWS)
EFCM Basic Edition (formerly, SANpilot)
HA-Fabric Manager (HAFM) Enterprise Fabric Connectivity Manager (EFCM)
Firmware Enterprise Operating System (E/OS)
HAFM Appliance EFCM Server
HA-Fabric Manager release notes 3

HAFM software version 08.09.01
HAFM 08.09.01 applies to the 1U rack-mount appliance only and cannot be installed on the notebook
server. All remote clients running down-level versions of HAFM are required to reinstall the client
application. You must exit HAFM before installing the latest version of HAFM. When logging in to the
HAFM appliance through the remote client, an error message is displayed stating that the new version
must be installed. Follow the instructions to install the new version of the remote client.
HAFM 08.09.01 is available on the HP web site
http://h18006.www1.hp.com/storage/saninfrastructure.html.
Upgrading to HAFM 08.09.01
If you have an HAFM notebook server or an HAFM appliance and you are upgrading to HAFM
08.09.01, you must follow the directions as detailed in the HP StorageWorks HA-Fabric Manager
Transition Guide (part number AA-RV1MD-TE), to successfully transfer your SAN configuration information
to HAFM 08.09.01. This version of this guide was written for HAFM 08.08.00, and also applies to
version 08.09.01.
Downgrading HAFM
When downgrading from HAFM 08.xx.xx to 07.xx.xx, your zoning configurations and nicknames are not
converted. HP recommends that you backup and save your 07.xx.xx configuration before upgrading to
08.xx.xx. You can then restore that backup if you downgrade to HAFM 07.xx.xx. See the HAFM Basic
user guide or EFCM Basic user guide for more information.
Operating systems
HAFM 08.09.01 clients are supported on the following operating systems:
•Windows®Server 2000 Professional, Service Pack 4
•Windows Server 2003, Service Pack 1
•Windows XP Professional, Service Pack 2
•Solaris®8and9
•Red Hat®Linux 9.0, kernel 2.4.20
•Red Hat Enterprise Linux 3, kernel 2.4.21
•HPUX 11.0
•IBM®AIX™4.3.3
•IBM AIX 5.1
Languages
English
4

Fixes
Table 2 lists issues that have been fixed in HAFM 08.09.01 since the release of HAFM 08.08.00.
Table 2 HAFM 08.09.01 fixes
Description Problem Category
The Operational Status column does not update correctly when a
McDATA 4Gb SAN Switch is taken offline. Instead of indicating
that the McDATA 4Gb SAN Switch is offline, the status in the
OperationStatuscolumnandalerticon(ProductListandTopology
view)displaysasfailed.
Display/Layout/GUI
Topology view has update problems when the principle switch is
not managed by HAFM and the fabric is segmented. The device
list adjacent to the topology may display the same switch in two
different fabrics or the switch may toggle from one fabric to another,
at a rate of approximately half of the discovery rate (approximately
eight seconds).
Display/Layout/GUI
In the product List, Domain ID column not getting sorted properly. Device List/Properties
Unable to discover switch when IP is reused in a switch swap-out
scenario.
Discovery
Changing the server time back causes the link to drop or the closes
the Element Manager. When the user attempts to reopen Element
Manager, the following message appears: The link to the
product is not available: Connection lost.
Element Manager
A Preferred Path cannot be enabled intermittently even when the
Domain ID is set to Insistent.
Element Manager
With FICON Management style, the Addresses menu item under
the Configure menu (Configure > Addresses)isnotavailable
in the 6140 Element Manager.
Element Manager
When performing a CTP switch over from either CTP 0 or 1, the entire
Hardware View window is blank until stopping and starting services.
Element Manager
The user is not notified that firmware download on single-CTP system
causes switch to boot the IPL file rather than the current active
configuration.
Element Manager
For an upgrade firmware install, the red progress bar is not updated. Element Manager
Stopping External port diagnostics in the middle of a test may cause
port to indicate failure.
Element Manager
Firmware may not update on HP-UX systems. Element Manager
The cooling fan module may display incorrect position information
in the element manager FRU list.
Element Manager
Port status alert indicator may not be removed when a NOS alert
is cleared from Ports tab
Element Manager
No log entries generated when a link between switches is
disconnected.
Event Management
User Action Event may not be generated in Master log for
modifications done in Call Home dialog.
Event Notification
The HAFM Event Log is missing system power up and standby ready
events after system power down.
Element Manager
When enabling or disabling fabric binding on muliple fabrics, the
wrong WWN displays in the Fabric Name field of the Fabric Binding
Status window.
Fabric Binding
HA-Fabric Manager release notes 5

Description Problem Category
Switches added to the fabric binding membership list may not be
able to join the fabric.
Fabric Binding
Isolated switches with fabric binding enabled may display information
from a prior fabric.
Fabric Binding
User created columns may not import using the CVS import feature. Import
HAFM does not take the NickName file from default folder and
import failed. Browsing and selecting file to import works.
Import
The cursor stays at the Yes option even though you select the No
option in the Copy Data and Settings menu.
Install
The UNIX remote client install path does not get appended in the 8.8
folder, resulting in erasing non-HAFM files during uninstall.
Install/Uninstall
The HAFM client does not automatically launch after installation; you
must manually launch the client.
Install/Uninstall
Embedded Port log may not open if a switch connection goes offline. Logs
Audit Log window may not open when you select Display in a new
window.
Logs
Events may not be displayed in Audit Log (Monitor menu). Logs
Security Event Information may be displayed in Hexadecimal format
in Security Event Log file.
Logs
Events may not be displayed in the Product Status Log window. Logs
Data Field may be empty in the Event Log window. Logs
Administrator’s password is not migrated from HAFM 7.X to 8.X.
Password gets reset to default “password” in HAFM 8.X.
Migration
The Configuration Wizard/Copy Data and Settings dialog box may
allow entry of an invalid migration path.
Migration
Call Home information is not migrated from 8.x to 8.6 during
upgrade.
Migration
Setup tools detail information is not migrated to prevent replacing
updated launch information.
Migration
Prior software versions automatically populated the label field with
adefaultvalue. Theuser-defined nickname field has replaced the
label field. Users may not have defined the label field information
in the prior software release so views may appear to lose data after
migration.
Migration
The Suppress Zoning RSCNs on zone set activations
setting may not migrate from earlier software versions.
The switch parameter may not have been saved properly in the
prior software version.
Migration
User edited startup and configuration file settings may be lost after
migration.
Migration
Persisted fabrics are not migrated from HAFM 7.X to 8.7. Migration
Discovery is not stable after migration. Discovered devices are
frequently getting changed after migrating from HAFM 8.6.
Migration
Delaymaybeexperiencedwhenimportingasortednicknamefile. Nicknames
When Port WWN and Node WWN are the same, the nickname
of node may be displayed as the Port Nickname in the configure
nicknames dialog
Nicknames
6

Description Problem Category
TheNicknamesizelimitisnotconsistentthroughoutHAFM
application.
Nicknames
No data in exported file when exporting WWNs of discovered
devices via Configure Nicknames dialog.
Nicknames
All WWNs are displayed in the configure nicknames dialog when
you select All WWNs from the display pull-down menu. The configure
nicknames dialog should show WWPN for devices that are currently
logged into a switch, or the WWNNs of switches in fabrics that are
being managed or discovered.
Nicknames
The Nickname import allowing invalid characters in the WWN field. Nicknames
High and Low critical indicators are not correct when opening Port
Performance Graph from the McDATA principal switch.
Performance
Switch label did not show up on the switch performance graph. Performance
Ports show up twice in the performance graph and a no light
port shows up in the graph.
Performance
The Errors drop-down list not available in the Performance Graph
for HAFM 8.6.
Performance
An exception is displayed when the Performance Graph menu item is
selected by right-clicking the switch.
Performance
Performance graph may display a short duration data increase
immediately after changing discovery method
Performance
Partition that is no longer part of the fabric continues to display on
the topology after setting a switch offlinetosegmentthefabric.
Physical Map
After clicking Remove after applying threshold in the same session,
the threshold is still applied.
Port Fencing
Threshold limit dialog is not triggered when fencing policy is applied
at the switch level.
Port Fencing
Security violations do not display yellow alert icon. Port Fencing
Switch threshold dialog does not open in session when a policy has
been removed from switch level during the same session.
Port Fencing
Removing switch level threshold causes ports to incorrectly inherit
policy. The switch inherits the fabric policy. All the subordinate
ports also inherit the policy regardless of any directly assigned port
policies.
Port Fencing
Nicknamescannotbeleftblank.YoumustselectCancel to do so. An
error occurs if you attempt to delete a previously added nickname.
Properties
User defined nicknames for node devices do not show in connection
properties.
Properties
Port WWN is listed in the Node WWN field of Properties dialog
box for a node device. The Properties dialog box of a node device
lists the Port WWN in the Node WWN field.
Properties
Operational Status for a switch may not display in Product list. Properties
Model field may not show correct value in Security Center. Value
may appear as 001.
SANtegrity
An Operation failed error message window appears after
HAFM Appliance shutdown.
Shutdown/Exit
Unable to enable fabricbinding for switch blade-only fabric. McDATA 4Gb SAN switches
HA-Fabric Manager release notes 7

Description Problem Category
In a blade switch fabric, an invalid attachment is not shown on
McDATA switch port. If there is an invalid attachment between 2
ports,oneonaswitchbladeandoneonaMcDATAswitch,the
invalid attachment is not displayed on the McDATA switch port.
McDATA 4Gb SAN switches
Switches may show in the same fabric with no visible connection (if
physical connection method is not supported).
Topology
BrowseusermaybeabletodeleteaGrouplog. User Management
The Copy, Delete, Delete All options may be enabled for a read
access user.
User Management
An application may not gather active zone sets from HAFM 7.1 if
invalid port number is discovered.
Zoning
HAFM as well as EWS are labeling port names as node names.
IntheHAFMSecuritytab,HAFMislabelingdeviceportnamesas
node names.
User Management
Cannot import a zone set with more than 2048 members. Zoning
All domain IDs in a fabric do not get displayed in product view. Display
HAFM 8.9. management of Directors and Edge Switches and
Directors running firmware 6.x or earlier, does not display the
port state correctly in various parts of the Element Manager:
1. Product/Fabric View: Ports that are logged-in
intermittently appear and vanish from view.
2. Performance View: Ports with established I/O do not
display the green bars or water marks. The performance
tables do not update port status after a requested "refresh".
3. Port Card view identifies all ports as "failed", with
flashing red diamond. All of the ports may be operational.
4. Node List: With devices attached, the node list fails to
populate to reflect device attachments.
Display/Layout/GUI
Export successful message displays without any file name and path. Export
No CTA event is logged in Threshold Alert log. Element Manager
FC Address fieldinNodeListViewcannotbesortedonall3bytesof
the port Address. HAFM sorts on the Domain/Area fields of the Port
Address only, not on the Port field.
Element Manager
When an Element Manager launch is attempted after IP address
is swapped between the Sphereon 4500 Fabric Switch and the
Sphereon 4700 Fabric Switch, an error message Invalid Product Type
is displayed for Sphereon 4700 Fabric Switch.
Element Manager
Users can enable the insistent domain ID feature when the preferred
domain ID and current domain ID do not match.
Element Manager
Users can enable the preferred path feature when the active domain
IDandthepreferreddomainIDdonotmatch.
Element Manager
Switch Blades and SANtegrity: Cannot activate Fabric Mode on a
switch blade-only fabric mixed-fabric in Open Systems mode.
Enterprise Fabric Mode
Switch Blade and Enterprise Fabric Mode: When attempting to
deactivate Switch Binding with Enterprise Fabric Mode enabled, no
message displays. A message should display stating that switch
binding cannot be deactivated while Enterprise Fabric Mode is
enabled.
Enterprise Fabric Mode
8

Description Problem Category
Switch Blades and Enterprise Fabric Mode: switch binding is
deactivated when Enterprise Fabric Mode is deactivated for switch
blades.
Enterprise Fabric Mode
When an ethernet cable is disconnected, the ethernet event is not
sent for switch in a discovered fabric.
Ethernet Event
No SNMP traps are displayed in the event log when sent across
through a firewall.
Event Log
Exception occurs when export- to-disk operation is performed on
aremoteclientonAIXplatform.
Export
HAFM 8.7 Fabric Log falsely reports ISLs added and removed. Fabric Persistence
Firmware load failure to a Director due to intermittent timeout. Install/Uninstall
The incorrect firmware version is displayed in fabric tree after
downgrading to firmware 7.1.3.
Install/Uninstall
FL Ports do not show a destination icon and port in performance
graphs.
Performance
The persist fabric does not work properly if the connection is lost to
a non-McDATA principal switch. The persist fabric always keys off
the principal switch. Non-McDATA switches do not display in HAFM.
When a non-McDATA switch is the principal switch and the fabric is
persisted, incorrect persist fabric updates occur if connection is lost to
the principal switch.
Persisted Fabric
A fenced port reports as disabled after a clearing problem and the
port is placed back online.
Port Fencing
The user must unblock, block then unblock again to recover a port
after a fencing policy has been invoked.
Port Fencing
Cannot enable a show route between the end devices attached to a
switch blades.
Show Route/Hide Route
Enhancements
This section describes enhancements for HAFM 08.09.01.
4Gb Fibre Channel speed support for Director 2/140
This release now provides 4 Gb Fibre Channel speed support in the Director 2/140 when using the 4
Gb/s Intrepid 6000 QPM Blade. You can purchase the 4 Gb/s Intrepid 6000 QPM Blade (QPM
blade) through an authorized McDATA Reseller.
For information about purchasing the 4 Gb/s Intrepid 6000 QPM Blade, please contact your authorized
McDATA reseller. You can findthelatestMcDATAcontactinformationbyregionontheMcDATAwebsite
http://www.mcdata.com/about/contact/index.html.
For detailed information about the 4 Gb/s Intrepid 6000 QPM Blade, see the latest versions of the
Intrepid 6140 and 6064 Directors Element Manager User Manual and the E/OS Command Line Interface
User Manual,availableontheMcDATAwebsite
http://www.mcdata.com/wwwapp/resourcelibrary/jsp/navpages/indexForSearchBy.jsp?search-
FolderName=/tpub/&resourceType=Technical%20Documents&product_line_root=tpub&re-
source=tpub_search&mcdata_category=www_resource.
HA-Fabric Manager release notes 9

PFE key support
For information about all PFE features, contact your authorized McDATA reseller.
You can find the latest McDATA contact information by region on the web site
http://www.mcdata.com/about/contact/index.html.
Compatibility and interoperability
This section describes Director or Edge Switch products compatibility.
HAFM and firmware compatibility
Table 3 lists the minimum version of HAFM that can run with the various versions of firmware for Directors,
Edge Switches, and the McDATA 4Gb SAN Switch.
Table 3 HAFM and firmware compatibility
Firmware version HAFM version
Director and Edge Switches
01.01.02 04.00.01 (HP EFCM)
01.02.02 -06 04.01.02-14 (SDCM)
01.03.00-3504.02.00-40 (HP EFCM)
01.04.00-01 04.02.00-40 (SDCM)
02.00.00-33 06.00.00-45 (HP EFCM)
02.00.02-01 06.00.02-06
04.01.02-04 06.03.01-05
05.02.00-13 07.02.00-09 (HAFM Appliance)
07.01.00-09 (Notebook Server)
05.05.00-12 None (firmware for Edge Switch 2/12 only)
06.01.00-18
06.02.00-22 08.02.00 recommended (HAFM Appliance)
07.02.00-09 minimum (HAFM Appliance)
07.01.00-09 minimum (Notebook Server)
07.00.00-84
07.01.02–4
07.01.03
08.06.00 recommended (HAFM Appliance)
07.02.00-09 minimum (HAFM Appliance)
07.01.00-09 minimum (Notebook Server)
08.01.00
08.01.01
08.02.00
08.08.00 recommended (HAFM Appliance)
07.02.00-09 minimum (HAFM Appliance)
07.01.00-09 minimum (Notebook Server)
09.00.00 08.09.01 recommended (HAFM Appliance)
07.02.00-09 minimum (HAFM Appliance)
07.01.00-09 minimum (Notebook Server)
McDATA 4Gb SAN Switch
5.2.2.15.00 08.09.01 recommended (HAFM Appliance)
08.08.08 minimum (HAFM Appliance)
Important information
This section lists information that you should consider before you using HAFM.
10

Disabling the SNMP Proxy
By default, the SNMP Proxy is enabled on all McDATA 4Gb SAN Switches. To ensure proper HAFM
functionality, disable the SNMP Proxy feature before attempting to discover the McDATA 4Gb SAN
Switches in HAFM.
To disable the SNMP Proxy:
1. Telnet to the switch.
2. Login to the switch.
3. Enter:
admin start
4. Enter:
set setup snmp
5. Set the ProxyEnable setting to False.
6. Enter Yat the following prompt:
Do you want to save and activate the snmp setup? (y/n): [n]
E-Mail filter considerations
The firmware e-mail filter overrides the HAFM e-mail filter. The firmware determines that certain events do
not generate an e-mail notification. In this situation, the e-mail message is not sent for these events even if
you add this type of event to the selected events section of the HAFM email filter dialog.
Managing McDATA 4Gb SAN Switch considerations
DO NOT use the Element Manager Fabric Binding feature on McDATA 4Gb SAN Switches. Element
Manager Fabric Binding does not synchronize with HAFM Fabric Binding. If both Fabric Binding features
are used, a disconnect can occur between the two features.
Examples
•If you activate an ISL set while a port set is already active, the ISL set takes over and vice versa.
This is working as designed.
•If you activate Switch Binding using the Element Manager, and then activate Fabric Binding using
HAFM, the McDATA 4Gb SAN Switch copies and appends the Fabric Binding to the current
Switch Binding security policy. This is working as designed. However, if you then deactivate
Fabric Binding using HAFM, it only deactivates on legacy M-Series and McDATA switches.
You must deactivate the policy using the Element Manager to deactivate it on the McDATA
4Gb SAN Switch.
If you also had Switch Binding activated and the learned Fabric Binding policy is merged, Switch
Binding becomes deactivated. A security hole exists if you do not realize this; however, it is apparent
that this happens when you look at the active policy. Activate the policy you created using the
Element Manager again to enable Switch Binding.
•If there is no Switch Binding activated in the Element Manager and Fabric Binding is activated
using HAFM, Fabric Binding can be deactivated only by using HAFM. This is working as
designed.
•If Switch Binding is activated using the Element Manager with an F-Port logged in to the McDATA
4Gb SAN Switch that is not in the Switch Binding membership list (for example, an administrator
command has not logged out the unwanted port), Fabric Binding cannot be activated using
HAFM.
HA-Fabric Manager release notes 11

Adding McDATA 4Gb SAN Switch to HAFM management
AMcDATA4GbSANSwitch is added to HAFM management the same way as other M-Series switches,
with one additional configuration requirement. The username and password for the Administrator account
of the switch mustbeconfiguredinHAFM.ThedefaultusernameandpasswordforaMcDATA4GbSAN
Switch are preconfigured for the switch in HAFM when the switch is added to HAFM management. If the
switch username and password have not been changed from the default settings, no change is required.
The default user name is admin,andthedefaultpasswordispassword.
To configuretheusernameandpasswordthoughHAFM:
1. Select Discover > Setup.
The Discover Setup dialog box is displayed.
2. Click Add.
The Address Properties dialog box is displayed.
3. Enter the description, IP address, and subnet mask in the appropriate boxes.
4. Select the Product Type and Access tab.
TheAddressPropertiesdialogboxwithProductTypeandAccesstab
is displayed.
5. Enter the switch user ID and password for the Administrator account.
6. Click OK.
Formatting is required for new backup CD-RW disks
If you use a different CD-RW disk to backup other than the one provided with the HAFM appliance, you
must access the desktop of the HAFM appliance in order to enable the formatting of the new disk.
Backups cannot be performed until the new disk is formatted.
Default zone is disabled by default
The default zone on the Directors, Edge Switches, and McDATA 4Gb SAN Switches is disabled by default.
You must cofigure zoning so that connected devices can communicate with the Director, Edge Switch, or
McDATA 4Gb SAN Switch.
Some IP addresses must be avoided
If you use HAFM to manage other M-Series Fabric Directors and Edge Switches, do not use IP addresses
in the following range:
192.168.0.0 through 192.168.0.255—This subnet is used internally by the HAFM Appliance. Using an IP
address in this range causes the Call Home feature to function incorrectly.
Logging out of Microsoft Windows after access to HAFM appliance is
recommended
After you access the HAFM appliance desktop via a web browser using the TightVNC application, HP
recommends that you log out of Microsoft Windows before disconnecting your web browser access.
This prevents unauthorized access to the HAFM appliance by someone using the TightVNC application
to access the HAFM appliance. When the new user attempts to log into the HAFM appliance, the
Welcome to Windows screen is displayed, and a Windows user name and password are required to
access the HAFM appliance Windows desktop. HAFM remote client access does not require the HAFM
appliance to be logged in to Windows.
12

Backup to network
When configuring backup to network using the Options dialog, the path for the backup output directory
must be in the folowing format:
\\Networkname\sharedfolder\subfolder
You must manually create a subfolder under a shared folder. HAFM backup does not recognize the
shared folder; only the subfolder.
For complete information for configuring backup to a network drive, see the McDATA manual, EFC
ManagerSoftwareUserManual, part number 620-000170-060, Rev A.
Known issues
This section describes the known issues related to the 1U HAFM appliance and the HAFM software.
External Loop diagnostic may show incorrect result
If an External Loop test is run on a port and the required external loopback plug is not inserted into that
port, the test result should show the test as failed. However, it shows the test as incomplete. The Event Log
correctly shows the Loopback diagnostics port failure event.
Port fencing issues
The Alert bar may not count the correct number of ports blocked by fencing. More than one port may be
blocked, however, only one is displayed.
Zoning restrictions
When you activate zoning with the maximum number of zones (2048) in a zoneset, the HAFM
application may hang.
For Solaris, Linux, and AIX remote clients, a zoneset cannot be imported from a directory not containing
aDTDfile.
Workaround
Donotactivateazonesetwith2048ormorezones.
For Solaris, Linux, and AIX remote clients, add the zoneset XML file and either remove the reference to the
DTD file or specify the full path to the DTD file.
Issue concerning HAFM remote client access to the HAFM appliance with dual LAN
configuration
When using a single public LAN connection at the HAFM appliance for all Ethernet communications, the
single LAN connection operates correctly for the following functions:
•Directors and Edge Switches that the HAFM appliance manages
•Computers seeking remote client access to the HAFM appliance
•SAN management applications such as HP OpenView SAN Area Manager
When using two LAN connections (public and private) at the HAFM appliance, Microsoft Windows
and HAFM determine the following:
•Which LAN is to be the private LAN for communication between the HAFM appliance, and the
Directors and Edge Switches that the HAFM appliance manages?
•Which LAN is to be the public LAN for communication between the HAFM appliance and
computers seeking remote client access to the HAFM appliance?
HA-Fabric Manager release notes 13

The issue arises because either LAN connection on the HAFM appliance can be the public LAN or the
private LAN. Though the Directors and Edge Switches can be managed via either LAN, the public
LAN is the only one that can support remote client access. Thus, if one attempts to access the HAFM
appliance via a remote client session and is unknowingly using what has been designated as the private
LAN,theremotesessionisnotallowed. TheIPaddressthattheHAFMappliancehasdeterminedto
be the public LAN which supports remote client access, displays HAFM which displays after selecting
SAN > Server Properties.
HAFM designates the public LAN as the first LAN detected whose IP address is not the reserved private
subnet 10.x.x.x. Thus, if neither IP address is 10.x.x.x,thefirst LAN detected by HAFM is designated
as the public LAN. This order of detection is influenced by Microsoft Windows and is not guaranteed.
For a dual LAN configuration, both LANs must be connected when the HAFM appliance is booted up. If
only one is connected, HAFM interprets this as a single LAN configuration, and the connected LAN is
designated as the LAN for remote client sessions.
Workaround
There are two ways to ensure the public and private designations of the LANs.
•If you use a private LAN IP address, this causes this LAN to be designated as the private LAN.
You must also have the public LAN connection active when the HAFM appliance is booting up, or
else HAFM interprets this as a single LAN connection configuration, and the 10.x.x.x LAN is
designated as the LAN for remote client sessions.
•You can configure a specified Ethernet interface on the HAFM appliance to be the public LAN (to
listen for remote client connections). To configure this feature, you must manually edit a file on the
HAFM appliance to explicitly specify which IP address HAFM should use as the public LAN.
For detailed instructions, see HP StorageWorks HA-Fabric Manager User Guide.
If the public LAN IP address of the HAFM appliance is ever changed, this file must be edited again to
reflect the new IP address.
Support for speed Auto-Negotiate
Auto-negotiate is supported. However, HP recommends that the port speed for E_Ports (for Interswitch
links, or ISLs) be set to a specificportspeed(1Gb/sec,2Gb/sec,or4GB/sec as appropriate for the
speed of the Directors or Edge Switches being connected) instead of to Negotiate.Usingaspecific
port speed decreases the time for a fabric build in response to some perturbation event in the fabric.
Similarly, setting a specific port speed for N_Ports also decreases fabric build time. However, setting a
specific port speed for N_Ports is not required.
There are a few older HBA devices that do not always succeed in logging in to a switch port when
the port speed is set for auto-negotiate.
Workaround
If an older HBA has difficulty logging into a switch port that has its port speed configured as Negotiate,
configure that port speed to 1Gb/sec,2Gb/sec,or4GB/sec according to the operation speed of the
HBA connected to that port.
Losing LAN connection to the HAFM appliance when logged in to HAFM
If the LAN connection to the HAFM appliance is lost while you are logged in to HAFM, the application
may stop.
Workaround
The LAN connection must be restored. Stopping HAFM has no impact on the Fibre Channel operations
of any Edge Switch or Director. Monitoring switch operations, logging events, and implementing
configuration changes are interrupted only while the LAN is not connected.
14

Effect of no LAN connection to HAFM appliance during boot up
If the HAFM appliance has no LAN connection while booting up, but it is connected after booting up, the
remote client sessions to the HAFM appliance are not allowed. Also, the IP address that is displayed
when you select SAN > Server Properties is possibly incorrect.
Workaround
This is corrected by restoring the LAN connection and rebooting the appliance. Rebooting the appliance
has no impact on the Fibre Channel operations of any switch or Director. Only monitoring switch
operations, logging events, and implementing configuration changes are interrupted.
Setting HAFM appliance LAN to use DHCP is activated to wrong LAN
When setting the IP configuration for an older HAFM appliance (delivered with HAFM server boot/restore
3.01.00 Rev A or earlier) using the LCD panel, setting DHCP configuration for LAN 1 causes LAN 2,
not LAN 1, to beconfigured for DHCP. Similarly, setting DHCP configuration for LAN 2 causes LAN 1,
not LAN 2, to be configured for DHCP.
Workaround
To configure LAN 1 for DHCP, select LAN 2 instead of LAN 1 at the start of the configuration procedure.
If you wish to configure LAN 2 for DHCP, select LAN 1 instead of LAN 2 at the start of the configuration
procedure.
New sound files are not added to Event Manager immediately
New sound files do not display in the pull down menu of the Event Manager. They cannot be selected for
inclusioninarule.Inorderforthenewsoundfiles to be available, HAFM services must be stopped and
restarted, or the HAFM appliance must be rebooted. This is not disruptive to managed switches, but
monitoring and logging functions are interrupted while the appliance is rebooting.
HP-UX parameters may need to be changed before you run the HAFM client
The following two HP-UX 11.0 kernel parameters are set too low for most Java applications. Usually
you will see this problem as a java.lang.OutOf MemoryError: unable to create new
native thread error. To resolve the issue, edit the following parameter limits:
•max_thread_proc—The maximum number of threads allowed in each process. The minimum
value (and default) is 64, which is often too low for most Java applications. Set the value of
the max_thread_proc higher than the expected maximum number of simultaneously active
threads. Themaximumvalueisthevalueofnkthread.
•nkthread—The total number of kernel threads available in the system. This parameter is similar
to the nproc tunable except that it defines the limit for the total number of kernel threads
able to run simultaneously in the system. The value must be greater than nproc. The default is
approximately twice that of nproc. The maximum is 30000. The suggested value of nkthread
is 2*max_thread_proc. If you have many Java processes running and each running process
uses many threads, you should increase this value.
Ethernet port on HAFM appliance may encounter problems
When you use TightVNC to access the HAFM appliance and you are running HAFM locally on the
HAFM appliance, the Ethernet port on the HAFM appliance can appear to hang. When the Ethernet port
encounters problems, the following may occur:
•There is an apparent loss of communication to switches being managed by HAFM.
•The appliance CPU usage and memory usage display abnormally high readings.
HA-Fabric Manager release notes 15

Workaround
If this condition persists, it may be necessary to reboot the HAFM appliance. This is not disruptive to
managed switches, but monitoring and logging functions are interrupted while the appliance is rebooting.
Disabled Show Route/Hide Route menu item
ThetheShowRoutemenuitem,fromtheRight-clickonaportintheProductTree,isdisabled(grayedout)
after installing HAFM on an HAFM Appliance that has not previously had HAFM installed. For example,
installing on a restored HAFM Appliance or a replacement HAFM Appliance.
Workaround
Activate or reactivate the zoneset to enable the Show Route menu item, making it available from the
Right-click menu.
Fibre Channel Logical Address error in the Zoning dialog box
The Fibre Channel Logical Address may appear as 000000 on the zoning panel.
Workaround
Restart the HAFM client to display the correct Fibre Channel Address.
Unconnected ports showing different speeds
In the Port List View, unconnected ports may show speeds different than what has been configured.
Workaround
Restart the HAFM client to display the correct speed.
Inaccurate Node Descriptor information
Node Descriptor information in the Node Properties dialog may occasionally be inaccurate.
Workaround
Restart the HAFM client to display the correct information.
HAFM client for Solaris does not respond to short cut keys
When using the HAFM client for Solaris, the client application does not respond to a small number
of shortcut keys.
Workaround
Choose one of the following workarounds:
•You may use an HAFM client for a different operating system.
•Forthoseshortcutsthatarenotavailable,youmustusethemouse.
Node Nickname does not appear in Port Properties dialog box
TheAttachedPortWWNfield of the Port Properties dialog box does not display the nickname of the
attached device if there is one configured to it.
Workaround
ThenicknamefortheAttachedPortWWNcanbeobservedfromtheNodePropertiesdialogbox.
16

HAFM Client cannotlogintoaserverthroughaCiscorouterwithstrictACLs
An HAFM Client cannot log in to a server when it is required to cross a Cisco router with strict ACLs.
Workaround
Remove the strict ACLs enforcement.
Topology view does not load with HAFM and a McDATA 4Gb SAN Switch
When using HAFM to manage a McDATA 4Gb SAN Switch for HP p-Class BladeSystem, if you perform
alogout instead of an exit, and if you log back as a different user, a problem is encountered and the
topology view does not load. The following error message appears:
SAN not loaded, no devices present
Workaround
Select exit from the menu instead of logout and relaunch the client.
Switch Binding dialog box does not display all the NPIV nodes
If NPIV is installed, the Switch Binding dialog box does not display all the NPIV nodes. The HAFM
8.09.01 switch binding dialog displays only the first NPIV device.
Workaround
Use HAFM Basic instead. All the logged-in NPIV devices are displayed in the Switch Binding dialog box.
Determining current version
To display the HAFM version, select Help > About HAFM from the HAFM menu bar.
Supersedes history
This version supersedes HAFM 8.08.00.
Effective date
October 2006
HA-Fabric Manager release notes 17