Important Information about Intel(R) Server Management (ISM) v5.5.7 January 5, 2004 Errata Items A. Install B. Silent Install C. Uninstall D. Documentation/User Interface E. Platform Instrumentation Control (PIC) F. ISM Console G. Platform Instrumentation (PI) H. Independent Hardware Vendor (IHV) Instrumentation I. Direct Platform Control (DPC) Console Manager J. Client System Setup Utility (CSSU) K. Enterprise System Management Console (ESMC) Integration (includes SNMP Support) L. Intelligent Chassis Management Bus (ICMB) M. LAN Alert Viewer N. One-Boot Flash Update Utility O. Serial Over LAN (SOL) Console Redirection This errata.txt document is current at product launch. Any new errata items not mentioned in this file will be documented in the monthly "Intel Server Management Specification Update" posted on http://support.intel.com. Search on the key words "Intel Server Management." ***************************************************************************** Information in this document is provided in connection with Intel products. No license, express or implied, by estoppel or otherwise, to any intellectual property rights is granted by this document. Except as provided in Intel's Terms and Conditions of Sale for such products, Intel assumes no liability whatsoever, and Intel disclaims any express or implied warranty, relating to sale and/or use of Intel products including liability or warranties relating to fitness for a particular purpose, merchantability, or infringement of any patent, copyright or other intellectual property right. (Intel makes no representations or warranties and specifically disclaims all liability as to the sufficiency, reliability, accuracy, completeness or usefulness of information in this file or in the ISM software). Intel products are not intended for use in medical, life saving, or life sustaining applications. Intel may make changes to the information contained in this file and/or the ISM Software, specifications, and product descriptions at any time, without notice, (however, Intel has no obligation to provide modifications, updates and/or support for same, nor shall Intel bear any liability whatsoever for any such changes or modifications.) This file identifies certain currently characterized errata; however, Intel does not represent or warrant either that such list is all- inclusive or the absence of errata which do not appear on this list. Intel shall not be liable for any damages whatsoever (including without limitation, lost profits, business interruption, computer failure or malfunction, or lost information) arising out of the use of or inability to use the ISM Software. Copyright (c) Intel Corporation 2004 Intel and Xeon are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. * Other names and brands may be claimed as the property of others. ***************************************************************************** A. Install: ***************************************************************************** 1. ISM setup and some of the components thereof require a pointing device and will not be driven with keyboards alone. 2. If a PI-only install is attempted on a system with a non-supported platform, the installation will not clean up after itself after determining that it should not install PI on this particular system. A PI-only install on a supported platform will work properly. ***************************************************************************** B. Silent Install: ***************************************************************************** 1. No errata. ***************************************************************************** C. Uninstall: ***************************************************************************** 1. For customers using H-P OpenView*, the files Ezrpcw32.dll, Wcdmionc.dll, Wcrap.dll, Winrpc32.dll, and Wonctcp.dll are not removed during uninstall when H-P OpenView is running. PROBLEM: Since the ISM Uninstall does not have information regarding other applications running, ISM does not know how to shutdown H-P OpenView. Therefore, there are ISM files that are being used by H-P OpenView that can not be deleted. IMPLICATION: Not all the files that are installed are removed from the system during uninstall. WORKAROUND: Stop H-P OpenView prior to ISM Uninstall. STATUS: Intel is investigating the possibility of fixing this erratum. 2. ISM objects are not being removed from the H-P OpenView database and map. PROBLEM: After uninstalling the Intel Server Console, ISM symbols are left on the H-P OpenView map. If the user double clicks on an ISM symbol, H-P OpenView has a Dr. Watson error. IMPLICATION: ISM symbols and objects created during ISM integration are not removed from the system during uninstall. WORKAROUND: Users have to copy manually "c:\Program files\Intel\ServerControl\Bin\UninstIscOv.txt" file to "c:\Openview\registration\C\UninstIscOv.txt". STATUS: Intel is investigating the possibility of fixing this erratum. 3. During a UnixWare* uninstall the following message may appear on the server: "S99ISC: Cannot open". This message is not significant, and the file s99isc will be removed once ISM uninstall is complete. ***************************************************************************** D. Documentation/User Interfaces: ***************************************************************************** 1. On internationalized releases, some text may appear in English or may not be correctly translated. PROBLEM: Some user interfaces, Help strings, or documentation may not have been translated, may contain redundant English text, or may contain incorrect translations. IMPLICATION: User may not have a clear understanding of the information being communicated by the user interface or documentation. WORKAROUND: None. STATUS: Intel is investigating the possibility of fixing this erratum. 2. On non-English operating systems, the font display may be unclear or corrupted in some portions of the user interface or help files. PROBLEM: The help and user interface may be unclear or difficult to read. IMPLICATION: User may have difficulty reading output. WORKAROUND: None. STATUS: Intel is currently investigating a fix for this erratum. 3. Pressing F1 for Help may not work. PROBLEM: On non-English operating systems, pressing F1 may not cause any Help to appear. Instead, the user may receive a message that Help does not exist. IMPLICATION: User may not have F1 access to needed Help topics. WORKAROUND: Access the Help topic through the menu selection. STATUS: Intel is investigating the possibility of fixing this erratum. 4. On non-English operating systems, the text may be cut off in the dialogs or status box, or dialogs may have formatting issues. PROBLEM: The translations may not fit within the space provided, and the user is unable to scroll or resize the dialog or status box. Dialogs may be truncated or have formatting issues which prevent the entire dialog from being viewed and accessed. IMPLICATION: User may be unable to read all the output or access all controls in a dialog. WORKAROUND: None. STATUS: Intel is investigating the possibility of fixing this erratum. 5. Context-sensitive help may not be available. PROBLEM: In some instances the Microsoft question mark icon for context sensitive help may not be available. IMPLICATION: User may not be able to access help for specific items in the interface. WORKAROUND: Look for information in the general help documentation. STATUS: Intel is currently investigating a fix for this erratum. 6. Help system may contain duplicate items in the index or have formatting issues. PROBLEM: On internationalized releases, duplicate entries and terms may be displayed in the help indices, or the help topics may have formatting problems. IMPLICATION: User may find help navigation difficult. WORKAROUND: None. STATUS: Intel is investigating the possibility of fixing this erratum. 7. On non-English operating systems, tabs may disappear when resizing windows. PROBLEM: When using PIC full screen, the right frame which displays details associated with the processor displays four tabs. When using a mid-size display, two of the tabs disappear. IMPLICATION: User may be unable to access full functional display. WORKAROUND: None. STATUS: Intel is currently investigating a fix for this erratum. 8. On non-English versions of the product, the Start Menu may contain incorrect or duplicate information. PROBLEM: Items may be displayed in the Start Menu which do not exist or may have been uninstalled, or some items may have duplicate entries. IMPLICATION: User may not get a clear representation of what has been installed and what items are available for use. WORKAROUND: None. STATUS: Intel is currently investigating a fix for this erratum. 9. The ISM Installation and User's Guide cannot be launched from the Korean Help system. PROBLEM: On a Korean system the ISM Installation and User's Guide cannot be launched from the Help system. IMPLICATION: User may not be able to view the User's Guide from within the application. WORKAROUND: The User's Guide can be launched from the Start Menu. STATUS: Intel is currently investigating a fix for this erratum. 10. Korean Help in LAN Alert Viewer is missing a section. PROBLEM: There are three sections in the LAN Alert Help, "LANAlert Viewer", "LANAlert Configuration", and "Alert Details". On Korean systems the first section, "LANAlert Viewer", is missing. IMPLICATION: Users on Korean systems will not have access to all the LAN Alert help. WORKAROUND: None. STATUS: Intel is currently investigating a fix for this erratum. 11. Installation and User's Guide bookmarks may be truncated. PROBLEM: On localized systems, the bookmarks for the "Installation and User's Guide" may be truncated. IMPLICATION: User may have difficulty identifying bookmarks. WORKAROUND: None. STATUS: Intel is investigating the possibility of fixing this erratum. 12. PIC closes while trying to configure. PROBLEM: On Korean systems, selecting the PIC configuration menu may cause the application to close. IMPLICATION: User may not be able to perform configuration actions. WORKAROUND: None. STATUS: Intel is investigating the possibility of fixing this erratum. ***************************************************************************** E. Platform Instrumentation Control (PIC): ***************************************************************************** 1. Keyboard shortcuts do not always work. PROBLEM: The ISM GUI is split into two windows or panes. The left pane is used for navigation, the right pane is used for presentation of information. If the current window focus is in the right pane, then keyboard shortcuts will not work. For example, the F5 key will not refresh the screen, the tab key will not navigate, and the Alt-A shortcut will not Apply. If the current window focus is in the left pane, then the F5 key will refresh the data displayed including the right pane. IMPLICATION: The mouse and menu must be used instead of the keyboard. WORKAROUND: Use the mouse to click in the left pane to apply the window focus. This will allow use of the F5 key for refresh. The tab key and Alt-A key do not work. STATUS: Intel is investigating the possibility of fixing this erratum. 2. The win32rap service reports several error messages if NetWare* services are not installed. PROBLEM: If ISM is installed on a console or server without installing the NetWare services, then win32rap may report the following error messages in the application event log: "failed to initialize transport 'ddetrdll'" "failed to initialize transport 'ipx' "initializing winsock for ipx failed (winsock error 0x0000273f)" IMPLICATION: These messages simply indicate that win32rap attempted to start the NetWare services. If you are using tcp/ip for server communication then they are not a concern. WORKAROUND: You can ignore the error if your network only supports tcp/ip. STATUS: Intel is not investigating this erratum. 3. The "Immediate Hardware Reset" and "Immediate Power Off" features may not work properly during initialization. PROBLEM: During the ISM GUI initialization, if the user selects the power control actions "Immediate Hardware Reset" or "Immediate Power Off", the actions may not work as documented. IMPLICATION: The power control actions "Immediate Hardware Reset" and "Immediate Power Off" may not work properly and the server may continue to run without the desired action. WORKAROUND: Users should wait until the ISM GUI has completed initialization (READY state in status bar) before attempting these actions. STATUS: Intel is investigating the possibility of fixing this erratum. 4. Paging alert actions may continue to happen after paging has been removed as an alert action. PROBLEM: If paging is configured as an alert action for a threshold and that threshold is crossed, a page will happen as desired. If paging is then removed as an alert action from the threshold, it may still occur. IMPLICATION: A page may be issued when not desired. WORKAROUND: No workaround has been identified. STATUS: Intel is investigating the possibility of fixing this erratum. 5. If both a shutdown action and a page are selected, the page may not be sent. PROBLEM: If both a shutdown action and a page are selected, the page may not be sent. IMPLICATION: A page may not be issued. WORKAROUND: Do not configure both a page and a shutdown action on the same event. STATUS: Intel is investigating the possibility of fixing this erratum. 6. Some processor temperature sensors may first show a current status of 'unknown' with a current reading of 'unsupported'. PROBLEM: The processor temperature sensors may initially show a current value of 'unknown' and a current reading of 'unsupported'. These values will be replaced by the correct data as soon as the sensor becomes active; this can take up to two minutes. It may be necessary to manually refresh the console application before the health indicators correctly reflect the new status values. IMPLICATION: The processor temperature sensor status information may not be available for approximately two minutes after the server boots. A manual refresh of the console application may be needed to correct the health indicators of a console that started before the status information was available. WORKAROUND: If a processor temperature sensor has a status of 'unknown', wait two minutes and refresh the console display. STATUS: Intel is investigating the possibility of fixing this erratum. 7. Due to the rounding of floating point values, the user can set different threshold values without variance in ISM PIC. PROBLEM: Due to the rounding of floating point values, the user can set different threshold values without any effective difference between them. This is possible only if the user tries to set different threshold values so close to each other that the hardware characteristics of the sensors are bypassed. IMPLICATION: The user may be able to set the same value for different sensor thresholds. WORKAROUND: Do not set different threshold values very close. STATUS: Intel is investigating the possibility of fixing this erratum. 8. Broadcast messages sent from NetWare to a Windows 2000 console may not be displayed. PROBLEM: When an alert action "Send Broadcast Message" is selected as one of the LRA actions for a sensor whenever it crosses a threshold, the broadcast message may not get displayed at the console if the console is running Windows 2000. This is the case when Microsoft Client for NetWare is being used at the console to access NetWare resources. IMPLICATION: It gives the impression that NetWare broadcast messages are not sent to the console at all. WORKAROUND: If the Novell client for Windows is used instead of the Microsoft client, then this problem does not happen. STATUS: Intel is investigating the possibility of fixing this erratum. 9. ISM PIC may not show all servers manageable via ICMB. PROBLEM: After the user selects the "Enable as Management Point" checkbox in the "Local ICMB Server Configuration" group box in the "ICMB Configuration" tabbed page which is accessed using the ICMB folder in the left pane sensor tree view, PIC may not discover all of the servers that support management via ICMB. IMPLICATION: User will not be able to manage all servers on the network that are manageable via ICMB. WORKAROUND: Relaunch PIC after selecting the "Enable as Management Point" checkbox. STATUS: Intel is investigating the possibility of fixing this erratum. 10. Actions for some hot swappable power supply events may not be set and handled properly. PROBLEM: ISM PIC supports control actions for events such as "Power supply failed", "Power Supply OK," and "Power Supply Likely to fail." These events include situations of the power supply failing or internal power cable coming loose, but they do not cover events such as removal and replacement of the hotswap power supply, or removal of the AC input power cord. The control actions, such as Shutdown the OS, Shutdown the OS and Power off, etc., cannot be set for these events: "Power Supply Inserted", "Power Supply Removed", "Power Supply AC lost", "Power Supply AC lost or out-of-range", and "Power Supply AC out-of-range, but present". Furthermore, after the removal or insertion of a power supply unit, ISM PI may not handle any of the power supply-related events properly. This means that if you set a control action to occur on power supply failure, and then replace a power supply without rebooting the OS, if a power supply subsequently fails, the control action may not occur. IMPLICATION: The system may not perform the requested control actions set for the power supply sensors. WORK-AROUND: None STATUS: Intel is investigating the possibility of fixing this erratum. 11. Auto refresh of health status on the ISM PIC GUI may not work when monitoring a Windows-based server. PROBLEM: Auto refresh in ISM PIC may not work when monitoring a Windows-based server across subnets. IMPLICATION: The PIC console may not accurately reflect the current health status of sensors for Windows-based servers, if the server and client systems are in different subnets. WORKAROUND: There are 2 options if this failure occurs: (1) Map a drive to the server being monitored. Then set broadcast alert messages for any sensors you wish to monitor. (2) Use F5 to refresh the ISM PIC GUI; this updates the current health status of the sensors and ISM. STATUS: Intel is investigating the possibility of fixing this erratum. 12. The PIC user interface might not correctly display current health status for servers running the OpenUNIX 8.0 operating system. PROBLEM: When monitoring an OpenUnix 8.0 based server, the colored icons representing health status on the left pane of PIC may not agree with the health status icons in the right pane of PIC. This situation only occurs in servers running OpenUNIX 8.0. In this situation, only the health icons displayed in the right pane represent the correct health status. The condition is caused by the DMI Service Provider running under OpenUNIX 8.0. IMPLICATION: The user might see conflicting health information when comparing the colored icons in the left pane with those in the right pane. For these cases, the icon displayed in the right hand pane of the PIC interface is correct. All event actions and logging of system events continue to operate correctly. WORKAROUND: If the managed server running OpenUNIX 8.0 is configured with a static IP address, the following steps will resolve the issue. a. Locate the "hosts" file on the client system running Intel Server Management PIC Console. If Windows 2000 is the operating system used on the client system and it is installed to C:\ drive, then this file will be located in the folder C:\WINNT\system32\drivers\etc b. Add the IP address of the remote server running OpenUNIX 8.0 along with the server name, if it is not already in the file. (e.g. 10.7.180.45 ismou8_45). c. Save the edits. d. Stop and restart network services on the server or reboot the server Performing these steps will display the same colored icons for health status in both the left and right hand panes of the PIC interface. An alternate method would be to add the IP address of the OpenUnix 8.0 server and its name to one of the nameservers referred to in the resolv.conf file in the server running OpenUNIX 8.0 e.g. In the server running OpenUNIX 8.0, the file /etc/resolv.conf may look like this search mydomain.com nameserver 192.168.0.100 nameserver 192.168.0.200 For example, if the IP address of the OpenUNIX 8.0 server is 192.168.0.10 and its name is ism55_ou80, edit the 'hosts' file in the server with IP address 192.168.0.100 and add the IP address and name of the OpenUNIX 8.0 server to it. 192.168.0.10 ism55_ou80 After entering and saving the changes, both the DNS server and the server running OpenUNIX 8.0 need to be restarted and so does the server running OpenUNIX 8.0 These changes will not work when the OpenUNIX 8.0 server running ISM uses DHCP to obtain its IP address. STATUS: Intel is investigating the possibility of fixing this erratum. ***************************************************************************** F. ISM Console: ***************************************************************************** NONE. ***************************************************************************** G. Platform Instrumentation (PI): ***************************************************************************** 1. BMC/BIOS will not maintain changes to dynamic IP address unless ISM PI component is running. PROBLEM: If you configure a DHCP (dynamic) IP address for a server using the SSU or Client SSU, the IP address is maintained by the Baseboard Management Controller (BMC). Ordinarily, a floating IP address will expire from time to time and the server is notified by the DHCP server of the new IP address. However, under certain circumstances the BMC/BIOS will retain the expired address, causing a network conflict with the entity that now has been assigned that IP address. This occurs if: A. The ISM PI component is not installed or is not running on the server (a PI agent keeps the BMC's IP in sync with the OS's IP address) or B. The OS is not running (e.g., the server is in standby mode or is running only DOS) Under these circumstances, the BMC may issue LAN Alerts or Gratuitous-ARP broadcasts (which are enabled by default) using the expired IP address, until the server is brought back to running the OS. This can occur on the SCB2, SDS2, or later platforms. IMPLICATION: Network traffic may be slow or exhibit access problems. WORKAROUND: Configure the DHCP server to issue only a "fixed" DHCP address to any servers using the BMC. Then the DHCP server will always issue the same IP address to those servers. You cannot use the SSU to do this, but must configure it directly on the DHCP server. Alternatively, you can use SSU to assign a Static IP address to the server, and not use DHCP. STATUS: Intel is currently investigating a fix for this erratum. 2. Unloading SENSOR.NLM may show a small number of short term memory allocations that were not free. PROBLEM: SENSOR.NLM may not free all memory associated with some functions. This will cause NetWare to display a message when it frees the memory after unloading the NLM. IMPLICATION: A small amount of memory will be used and will not be made available again until the NLM is unloaded. This should not be a significant number of bytes and should not impact server performance. WORKAROUND: No workaround has been identified. STATUS: Intel is investigating the possibility of fixing this erratum. 3. The ISM Console health indication might not correctly display current health status for servers running the Red Hat Linux operating system. PROBLEM: The colored icon representing health status displayed in the ISM Console for servers running Red Hat Linux 8.0 might not be the same as the colored icon displayed in the PIC user interface. This situation only occurs in servers running Red Hat Linux and using a static IP address. In this situation, the health icon displayed in PIC represents the correct current health status, while the icon displayed in the ISM Console represents an incorrect health status. The condition is caused by the DMI Service Provider running under Red Hat Linux. Servers running Red Hat Linux 8.0 which use DHCP and have access to a DNS server for name/address resolution, do not experience this problem. IMPLICATION: The user might see conflicting health information if comparing the colored icon in the ISM Console interface to the colored icon in the PIC interface. For these cases, the icon displayed in the PIC interface is correct. All actions and logging of system events continue to operate correctly. WORKAROUND: If the managed server running Red Hat Linux 8.0 is configured with a static IP address, the following steps will resolve the issue. a. Edit the "hosts" file (located in "/etc" folder). b. Add the IP address of the server along with the server name, if it is not already in the file. (e.g. 10.7.180.45 scb2_45). c. Save the edits. d. Restart network services on the server or reboot the server Performing these steps will display the same colored icons for health status in both the ISM Console and the PIC interface. STATUS: Intel is investigating the possibility of fixing this erratum. 4. IP address configuration in NetWare PROBLEM: If you use the Internetworking Configuration NLM (INETCFG.NLM) to configure the IP address of the NetWare server, the IP Synchronization Agent (IPSA.NLM), might not be able to synchronize the IP address configured by the INETCFG.NLM with the BMC. IMPLICATION: Network traffic may be slow or exhibit access problems. WORKAROUND: Configure the IP address in the startup script AUTOEXEC.NCF and remove network configuration from the Internetworking Configuration NLM (INETCFG.NLM). Once INETCFG.NLM configures the network board, it comments out any LOAD and BIND statements in the file AUTOEXEC.NCF and loads them using the startup script INITSYS.NCF. To restore the original configuration take the following steps: a. On the NetWare server, run INETCFG.NLM and remove any network configuration. b. Exit to the server console. c. Edit the file AUTOEXEC.NCF from the server console or from a client connected to the NetWare server. Comment the line in the file AUTOEXEC.NCF which loads the startup script INITSYS.NCF by placing a # in front of the LOAD statement as below: #LOAD SYS:ETC\INITSYS.NCF d. Add statements to load and bind the appropriate network drivers. For example, to load the Intel Gigabit Ethernet driver (CE1000.LAN) using Frame type Ethernet_II and bind it using the following information: IP Address: 169.254.11.30 Network Mask: 255.255.255.0 Gateway IP address: 169.254.11.251 you would add these lines into AUTOEXEC.NCF: LOAD CE1000.LAN FRAME=Ethernet_II NAME=CE1000_1_EII BIND IP CE1000_1_EII ADDR=169.254.11.30 MASK=255.255.255.0 GATE=169.254.11.251 e. Restart the server to verify that the network drivers are loaded and bound. STATUS: Intel is investigating the possibility of fixing this erratum. 5. Logging into a server from ISM 5.5.x Console may occasionally cause a login to server failure, due to password error. PROBLEM: While logging on to a server using ISM 5.5.x remote console, the user may encounter a login failure. An error message indicating that the failure is due to a password error, will appear as a popup message. This happens only occasionally. However, if the user tries to login again, they can do so successfully. IMPLICATION: Login failure due to password error at the first login trial, causes confusion to the user. WORKAROUND: Login a second time and it will work. User can login successfully to a server from a remote ISM 5.5.x console. STATUS: Intel is investigating the possibility of fixing this erratum. 6. While trying to load certain MIB files in ISM 5.5.x using ESMC HPOV, the user will encounter some failures and will see some error messages. PROBLEM: User will notice some messages while loading certain MIB files in ISM 5.5.x using ESMC HOPV. Specifically, there will be some errors loading Basebrd2.mib & rmtchas.mib. User may also not be able to see MIB data and may get GETNEXT errors. Unloading the MIB successfully and then reloading does not fix the issue. IMPLICATION: It is cumbersome to the user to reboot a second time for proper MIB file loading. WORKAROUND: User can do an additional reboot and reload the MIBs without an error. STATUS: Intel is currently investigating a fix for this erratum. 7. Hot-swap fan removal on SE7501BR2 servers will show as a critical fan error. The fan icon will not be removed in ISM PIC. PROBLEM: ISM 5.5.x will not recognize a hot-swap fan removal on the SE7501BR2 servers. A fan removal will show as critical error on the PIC. The icon will not be removed. When the fan is replaced, the fan icon will go back to the green healthy state. IMPLICATION: The ISM PIC health status for fans could be misleading to the user. WORKAROUND: None. Fan removal seen as fan critical error. STATUS: Intel is currently investigating a fix for this erratum. 8. Power Unit Global Table shows fully redundant for system with only two Power Supplies. PROBLEM: ISM PIC shows green healthy status for the Power Unit Global table when there are only two Power Supplies in the Power Cage that can hold three Power Supplies. It should show a non-critical error for a degraded power unit redundancy. IMPLICATION: The health status of the Power Unit Global Table could be misleading to the user on a system with only two Power Supplies. STATUS: Intel is currently investigating a fix for this erratum. 9. Incorrect message logged to SEL for hot swap power supply event on HudsonIII chassis. PROBLEM: On HudsonIII/SE7501BR2 and HudsonIII/SE7501HG2 chassis with BRP hot swap power supply unit, during power cable removal the message "power supply failure event" will be logged in the SEL instead of "power AC lost event". IMPLICATION: ISM Platform Instrumentation will not be able to perform any actions set for "power AC lost event" in GUI. WORKAROUND: In this chassis and power supply unit combination, during the event of power cable removal setting actions to "power supply failure event" in GUI will work as expected. STATUS: Intel is investigating the possibility of fixing this erratum. 10. Server Health events are not updating in PIC with OpenUnix Server. PROBLEM: The ISM Console health indication might not correctly display current health status for servers running the OpenUnix 8.0 operating system. The colored icon representing health status displayed in the ISM Console for servers running OpenUnix 8.0 might not display the correct status (Red - Critical, Yellow - Non-Critical, Green - OK) for sensors whose status changes. This situation only occurs in servers running OpenUnix 8.0. The condition is caused by the DMI Service Provider running under OpenUnix 8.0. IMPLICATION: The user might see incorrect health information status represented by the colored icon in the PIC interface. All actions and logging of system events continue to operate correctly. WORKAROUND: If the managed server running OpenUnix 8.0 is configured with a static IP address, the following steps will resolve the issue. a. Edit the "hosts" file in the client system where ISMPIC application is running. On a system where Windows 2000 is installed on C:\ drive to the folder WINNT, this file would be found in the folder C:\WINNT\system32\drivers\etc. b. Add the IP address of the server along with the server name, if it is not already in the file. (e.g. 10.7.180.45 scb2_45). c. Save the edits. d. Restart the client and connect to the OpenUnix 8.0 server running ISM services Performing these steps will make sure that the health status for sensors in PIC on the client get updated when the health status changes on the managed OpenUnix 8.0 server running ISM modules. STATUS: Intel is investigating the possibility of fixing this erratum. 11. Netware module ONCSP.NLM fails to start after it is copied manually. PROBLEM: If ONCSP.NLM is copied manually after installing TIRPC, then it may not start correctly and a message may be displayed on the NetWare server indicating that it did not start. IMPLICATION: ISM will be nonfunctional as no clients (ISM PIC) will be able to connect to the server to obtain status or set/get sensor values. WORKAROUND: Follow the steps as outlined in the Installation and User Guide to install the TIRPC package, the TIRPC updates and then ISM. STATUS: No further action is planned by Intel. 12. ISM services will not start if IPX/SPX protocols are disabled in a NetWare server PROBLEM: If you configure a NetWare server to load only the IP protocol, then ISM services do not function correctly even though the NLMs may load without errors. IMPLICATION: ISM PIC cannot connect to the NetWare server nor will health status indications be sent across the network. WORKAROUND: For ISM to function, confirm that IPX/SPX services are not disabled on the server. STATUS: Intel is currently investigating a fix for this erratum. ***************************************************************************** H. Independent Hardware Vendor (IHV) Instrumentation: ***************************************************************************** 1. The Adaptec* SCSI component reports the wrong information about Logical Unit Drives on Mylex* Host Adapters. PROBLEM: On NetWare Servers, a known incompatibility between the Adaptec and the Mylex ASPI drivers can cause the Adaptec SCSI component to report the wrong information about Logical Unit Drives on Mylex Host Adapters. IMPLICATION: The wrong information may display. WORKAROUND: The SCSI DMI instrumentation requires Adaptec ASPI drivers (ASPI32.SYS and WNASPI32.DLL). The user must load these drivers on the server. The Mylex DMI instrumentation does not require Mylex ASPI drivers. Unless some other application requires the Mylex ASPI drivers, do not load the Mylex ASPI drivers. STATUS: Intel is currently not investigating a fix for this erratum. 2. During Wired for Management (WfM) testing, the DMTF CompChk tool reports errors in the Adaptec MIF. PROBLEM: The recent discovery of MIF syntax violations in Adaptec instrumentation does not compromise the Intel(R) AD450NX MP systems standing as "WfM upgradeable" system. The SCSI instrumentation supplied by Adaptec is not required on WfM 1.1a servers. IMPLICATION: As a side effect of WfM testing, we have found cases where the Adaptec MIF does not comply with the syntax of DMTF standard groups. The WfM test suite uses the DMTF CompChk tool to scan the entire system MIF. It reports WfM pass/fail on the specific groups required by WfM, and also notifies the user about any problems it detects with respect to general DMI compliance for all other groups it finds registered with the service provider. The Adaptec problems fall into this second category. WORKAROUND: None required. The (slightly) non-compliant MIFs can be successfully browsed by the DMI Explorer browser bundled with ISM. We currently believe there should be no effect on the user's ability to access and use this data. STATUS: Intel is investigating the possibility of fixing this erratum. 3. On server systems with onboard Adaptec SCSI controllers, users will see CIO Array Management event messages as the operating system loads. PROBLEM: These messages are not reporting a problem with the system. The messages are normal indications from Self Monitoring and Reporting Technology (SMART) that the SCSI controller and hard drives are communicating normally. IMPLICATION: Some users may find these messages annoying as the operating system is loading. WORKAROUND: In Microsoft Windows operating systems the user may turn off SMART polling to eliminate these messages. This will however also prevent any other SMART messages being generated, including those predicting drive failure. To turn off SMART polling the user can modify the file IOMGR.INI in the following file folder, Program Files\Intel\ServerManagement\bin\Adaptec. Any text editor may be used to open this file. Near the end of this file is a line which reads by default... "SmartPolling = Yes". The user may modify this line to read "SmartPolling = No". STATUS: CIO Array Management messages are normal indications. 4. DMI and SNMP support are not available for SCSI controller and drives on NetWare 6.0. PROBLEM: The DMI instrumentation is not working correctly under NetWare 6.0 for servers with the Adaptec Ultra 320 SCSI controller, also prevent the SNMP support from working correctly. IMPLICATION: Neither DMI events nor SNMP traps will be available for events related to the Adaptec ultra 320 SCSI controller and attached SCSI disk drives for servers running NetWare 6.0 WORKAROUND: None available. STATUS: Intel is investigating the possibility of fixing this erratum. 5. DMI support for on-board network adapters is available for Windows Server 2003 only for Intel (R) SE7501WV2, Intel (R) SE7501HG2 and Intel (R) SE7500WV2 servers. PROBLEM: The DMI management software for on-board Intel network adapters integrated with ISM 5.5.7 works only with a limited number of servers when Windows Server 2003 is installed. On an unsupported server when using ISM to manage one of the affected systems, there will be no Intel Ethernet LAN Adapter(s) folder on the Platform Instrumentation Control console under the Third Party folder. IMPLICATION: When managing an unsupported server no alerts will be generated for the onboard Intel network adapters. WORKAROUND: None available. STATUS: Intel is investigating resolving this issue. ***************************************************************************** I. Direct Platform Control (DPC) Console Manager: ***************************************************************************** 1. Redirect doesn't support green (teal) color. PROBLEM: Console redirect does not display green shades. Currently, teal is displayed as cyan (or aqua). IMPLICATION: The Adaptec Option ROM BIOS displays white text on teal background. DPC Console Manager is displaying this as white text on cyan, which is extremely hard to read. WORKAROUND: None identified. STATUS: Intel is investigating the possibility of fixing this erratum. 2. "Graceful Shutdown has failed." or "Graceful reset has failed" dialog appears when you know the OS is up and Platform Instrumentation is running. PROBLEM: On rare occasions (less than 2% of the time), when you attempt to do a reset, shutdown or boot to Service Partition, the Platform Instrumentation service may begin performing this operation, but DPC may not receive notification of this action. Under this circumstance, DPC will present a dialog indicating that this 'graceful' action failed, would you like to 'force' it. IMPLICATION: Unexpected behavior, user thinks the reset, power-down or reboot to Service Partition action failed, when in fact it is successfully being carried out. WORKAROUND: If you see in the status of DPC that the OS is detected, or if you are certain this OS is running and the Platform Instrumentation service is active, and this dialog appears, press 'Cancel' when prompted for a 'Force' operation. Wait 5 minutes and try again. STATUS: Intel is currently investigating a fix for this erratum. ***************************************************************************** J. Client System Setup Utility (CSSU): ***************************************************************************** 1. Pointing device needed to run CSSU. PROBLEM: Most dialogs in CSSU do not have keyboard support; for example, pressing the Enter key will not cause any action to occur in dialogs. IMPLICATION: Ordinary keystrokes, hotkeys or keyboard shortcuts may not work. WORKAROUND: Use a pointing device to access CSSU functions. STATUS: Intel is investigating the possibility of fixing this erratum. 2. Invalid characters appear for the server name. PROBLEM: When dialed in from an English operating system to a non-English server, the server name may display invalid characters in the server box at the bottom of the Client SSU container. IMPLICATION: User may be unable to identify server. WORKAROUND: None. STATUS: Intel is investigating the possibility of fixing this erratum. 3. User and Administrator Passwords in SSU/CSSU/BIOS PROBLEM The Password Manager Help in CSSU describes that an Administrator password must be entered before you can set or change a User password. On the SSH4/SPSH4/SRSH4 series of baseboards this is not true. The BIOS on these baseboards allows you to set or change the User password whether there is an Administrator password or not. IMPLICATION: On these baseboards a user may set user and administrator independently, unlike how it is described in CSSU Help. WORKAROUND: If any passwords will be set, always set an administrator password first, then a user password. If any passwords will be cleared, always clear the user password first, then the administrator password. STATUS: Intel enforces this policy in BIOS on baseboards beginning with the SHG2 baseboard. 4. Keyboard input associated with the console redirection window during a CSSU remote client reboot does not function. PROBLEM : The console redirection window that is displayed while CSSU is connecting to a remote server does not accept the key. IMPLICATION: This becomes a problem when the user invokes the boot device selection menu or the F2 BIOS setup program on the server. Any operation that requires the use of the ENTER key to select options or to exit will not work. Also, if a password on boot is enabled and set, the user will not be able to press the enter key to invoke to server to accept the password and continue the boot process. WORKAROUND : The user should avoid any keyboard activity while the CSSU is rebooting the remote server into the service partition. If the user inadvertently uses keystrokes that prevent completing the boot process, the CSSU connection can be terminated and a new connection re-established. When CSSU reconnects it will cause the remote server to reboot. If the user wishes to remotely run the F2 BIOS setup or other console redirection operations, using Direct Platform Control is recommended. 5. When EMP terminal mode is activated, heartbeat characters take exclusive control of the remote console. PROBLEM: When EMP terminal mode is activated, heartbeat characters from the EMP take exclusive control of the remote console effectively disabling the console redirection. These handshake characters are all that can be seen on the remote console beginning shortly after the option ROM. IMPLICATION: This prohibits the ability to access any PXE boot menus or user login prompts, etc. WORKAROUND: None. STATUS: Intel is investigating the possibility of fixing this erratum. ***************************************************************************** K. Enterprise System Management Console (ESMC) Integration ***************************************************************************** 1. SNMP Support: Some MIB files will not manually load in some versions of H-P OpenView. PROBLEM: Some of the MIBs provided with ISM will not load in some versions of H-P OpenView. IMPLICATION: H-P OpenView will not provide as rich information when displaying an SNMP trap from ISM. WORKAROUND: No workaround has been identified. STATUS: Intel is investigating the possibility of fixing this erratum. 2. Unable to launch ISM applications from CA-TNG map. PROBLEM: After ISM gets installed and while some ISM servers are being discovered and added to the CA-TNG map, right-clicking on an ISM server and selecting "Launch ISM" fails to launch an ISM application. IMPLICATION: User will be unable to launch ISM applications from the CA-TNG map. WORKAROUND: This problem can be easily fixed by first stopping and then restarting the Intel TNG-ISM AutoDiscovery service. If that fails to fix the problem, a system reboot may be necessary. After the second reboot everything should work properly. STATUS: Intel is investigating the possibility of fixing this erratum. 3. Unable to see ISM objects on the CA-TNG map. PROBLEM: After ISM gets installed, the user is unable to see ISM objects on the CA-TNG map. IMPLICATION: User will be unable to see ISM objects on the CA-TNG map. WORKAROUND: This problem can be easily fixed by first stopping and then restarting the Intel TNG-ISM AutoDiscovery service. If that fails to fix the problem, a system reboot may be necessary. After the second reboot everything should work properly. STATUS: Intel is investigating the possibility of fixing this erratum. 4. Must reboot console twice following ISM install to discover ISM nodes. PROBLEM: User must reboot Windows XP and Windows 2000 consoles twice following ISM installation in order to discover H-P OpenView ISM nodes. After installing H-P OpenView ver 6.2 and ISM the console is rebooted. When the user opens H-P OpenView, servers with Platform Instrumentation are discovered, but no ISM nodes are available for selection. If the applications are closed and the system is reboot the ISM Nodes are available the next time H-P OpenView is launched. IMPLICATION: User may not be able to launch ISM applications and not know why. WORKAROUND: None. STATUS: Intel is investigating the possibility of fixing this erratum. ***************************************************************************** L. Intelligent Chassis Management Bus (ICMB) ***************************************************************************** 1. Immediate Power Off and Immediate Hardware Reset actions are not functional in ISM 5.x using ICMB. PROBLEM: After a successful ICMB connection using Platform Instrumentation Control (PIC), if the user tries to do an Immediate Power Off Server, this error message is displayed: "The attempt to power off server failed". Selecting Immediate Hardware Reset action does nothing. Immediate Power Off and Immediate Hardware Reset do not function as desired. IMPLICATION: Immediate Power Off and Immediate Hardware Reset actions will not work as desired. WORKAROUND: None STATUS: Intel is investigating the possibility of fixing this erratum. 2. ISM Health folder for ICMB chassis always indicates green "OK" status. PROBLEM: Using PIC, the user can access the ICMB folder. If the user selects the "Enable as Management Point" checkbox under "ICMB Configuration" tab, after the discovery period, PIC will discover any ICMB chassis. If the user goes to ICMB menu to choose "View Managed Server(s)..." and selects a remote server, PIC will show the remote ICMB server selected. However, the Health folder for this remote server will always show green "OK" status, even if the Chassis status for that remote server shows red "Critical" error. IMPLICATION: ISM Health folder for ICMB chassis does not function correctly. WORKAROUND: None STATUS: Intel is investigating the possibility of fixing this erratum. ***************************************************************************** M. LAN Alert Viewer ***************************************************************************** 1. No LAN Alert message is generated for Power Supply Removed or Power supply AC lost in ISM 5.5.x PROBLEM: When a power supply unit is removed or when the AC power is lost for a power supply unit, no LAN Alert is generated for the power supply events in LAN Alert Viewer. The following SEL events are logged: Power Supply Removed (Power Supply AC Lost) Power Supply Redundancy Lost but there is no SEL log that indicates: Power Supply Failure In SSU, LAN Alerts can only be set in one case for power supply: Power Supply Failure. So, these LAN Alerts can no longer be set from SSU. WORKAROUND: None. STATUS: Intel is currently investigating a fix for this erratum. ***************************************************************************** N. One-Boot Flash Update Utility ***************************************************************************** 1. Updating the FRUs on a system with a Hot Swap Controller (HSC) will cause the One-Boot Flash Update utility to appear to hang (~45sec). PROBLEM: The One-Boot Flash Update utility takes a long time to update the FRUs on the HSC and to the user, the utility may appear to be hung. IMPLICATION: User may think the utility is not functioning. WORKAROUND: Wait for the utility to complete the task (~45sec). STATUS: Intel is investigating the possibility of fixing this erratum. 2. Downloading files from a FTP/HTTP server that is running an Operating System that is different from the Operating System on the target server may fail. PROBLEM: One-Boot Flash Update utility may fail to download files if the utility is running in an OS that is different than the one running on the FTP/HTTP server. For example, if the FTP/HTTP server is running Windows 2000, and the One-Boot Flash Update utility is executing on a system that is running Red Hat Linux 8.0, then the utility may not be able to download the files from the server. IMPLICATION: User may think that this feature is not functioning. WORKAROUND: (1) Download the update package and update files to a local directory prior to running the utility; or (2) use an FTP/HTTP server that is running the same OS as the target server on which the utility will be executed. STATUS: Intel is investigating the possibility of fixing this erratum. 5. Utility can't download the update files if the URL contains spaces. PROBLEM: One-Boot Flash Update can not download the update files if the URL contains spaces. For example, the command: "flashupdt -i -u http:/www.examplesite.com/Flash Update Package" contains spaces in the URL portion of the -u command line option. Instead, the URL should be specified with the spaces replaced by "%20", as follows: "flashupdt -i -u http:/www.examplesite.com/Flash%20Update%20Package" IMPLICATION: User may think that this feature is not functioning. WORKAROUND: Replace each space with "%20". STATUS: Intel is investigating the possibility of fixing this erratum. 6. Utility allows a BMC HEX file to be updated onto the server even if the HEX file is not intended for that server. PROBLEM: The One-Boot Flash Update utility allows the user to update the server with a HEX file that is intended for a different server, i.e. the utility does not validate that the HEX file is intended for the platform prior to updating the FW. IMPLICATION: User may think that the FW has been successfully updated with the HEX file however the FW will fail to update after the system reset, and will rollback to the previous version. WORKAROUND: The "PROBE SYSTEM" command should be used in the Configuration (CFG) file to detect the platform prior to updating the BMC. By using the "PROBE SYSTEM" command the CFG file can guard against updating a HEX file to a server for which it is not intended. Please note however that even without this added check there is no risk of corrupting the FW with an incompatible version, since the FW will not allow the update to take place after the system reset; i.e. the FW will detect that the new image is not intended for the platform and it will rollback to the previous version. STATUS: Intel is investigating the possibility of fixing this erratum. ***************************************************************************** O. Serial Over LAN (SOL) Console Redirection ***************************************************************************** 1. From a Red Hat Linux operating system based client, some keyboard operations do not respond correctly when executed remotely during a Serial Over LAN connection. PROBLEM: After switching to Serial Over LAN console redirection from a Command Line Interface connection, some keyboard operations do not respond correctly. Arrow direction keys, Enter key and some function keys may fail to operate as expected. IMPLICATION: Users might not be able to move around the console screen and could possibly encounter a situation where they cannot exit or close a remote activity. WORKAROUND: From either a KDE or GNOME terminal window, use “xterm” to open a second terminal window. From that window, start the Command Line Interface connection as before using “telnet localhost 623”. STATUS: Intel is investigating the possibility of fixing this erratum. ***************************************************************************** End of errata file *****************************************************************************