|C H A P T E R 1|
Important Information About the Sun Fire T2000 Server
These product notes contain important and late-breaking information about the Sun Fire T2000 server.
Refer to the following sections for details:
Instructions for installing, administering, and using the Sun Fire T2000 server are provided in the Sun Fire T2000 server documentation set. The entire documentation set is available for download from the following web site:
Information described in these product notes supersedes the information in the Sun Fire T2000 documentation set.
The product notes for the Sun Fire T2000 server are presented in the following categories:
Start by reviewing the general information in this chapter, and then review the notes in the chapter that apply to your server based on its part number.
1. Gain Access to the ALOM CMT system controller prompt (sc>).
At the Sun Fire T2000 console, type #. (Pound Period).
2. Perform the showfru command as follows:
3. Use the Sun_Partnumber from Step 2 and TABLE 1 to determine which notes apply to your server.
The remaining sections in this chapter describe information that applies to all Sun Fire T2000 Servers.
When PCI-E cards with X1, X2, or X4 lane widths have been installed in PCI-E slots (this problem does not affect X8 PCI-E cards), intermittent failures to recognize the cards can occur in Sun Fire T2000 servers. These failures are intermittent and occur during device training.
Normally, PCI-E devices are recognized by OpenBoot firmware as:
When these failures occur, OpenBoot firmware does not show some of the PCI-E devices in the device tree (displayed using the show-devs command at the ok prompt). After the Solaris OS boots, the cards will be missing from the output of the prtdiag -v command, as well. The system may also report a generic FMA message:
Workaround: Without the patches described below, reboot the system repeatedly until the system can see all devices (usually 1 or 2 reboots are necessary).
The issue is resolved on the following platforms:
Instructions for installing DIMMs are provided in the Sun Fire T2000 Server Service Manual.
4 GByte DIMMs may not be mentioned in the service manual, but the DIMM installation instructions apply to all supported DIMMs (512 MB, 1 GB, 2 GB, and 4 GB).
System Firmware 6.3.0 includes ALOM CMT v1.3. There are several new features in ALOM CMT v1.3:
For further information on the new features of ALOM CMT v1.3, refer to the Advanced Lights Out Management (ALOM) CMT v1.3 Guide (819-7981-10).
Changes to the networking framework in upcoming software releases might require system administrators or developers to update references to ipge interfaces. To prepare for this change, note the locations of all references to the names of networking frameworks. For example, if you reference the name of an ipge interface in a system configuration file, note the location now. Alternatively, you might choose to limit the number of applications explicitly configured to use this interface.
This section describes mandatory /etc/system file entries that must be listed in this file to ensure the optimal functionality of the server. These entries resolve CRs 6274126* and 6344888 (see Chapter 3, TABLE 3-3).
The following entry must be in the /etc/system file:
If you have a Sun Fire T2000 Server with part number 501-6843 and it is running the Solaris 10 3/05 HW2 Operating System, you must also have the following entry:
Perform this procedure in the following circumstances:
1. Log in as superuser.
2. Check the /etc/system file to see if the mandatory lines are in the file.
set segkmem_lpsize=0x400000 <--See footnote
3. If the entries are not there, add them.
Use an editor to edit the /etc/system file and add both lines.
Reboot the server.
Over time, different motherboards were manufactured for this server, and not all motherboards are interchangeable. If you replace the motherboard, you must replace it with a motherboard that has the same part number or an approved alternative motherboard (approved alternative part numbers are listed on the Sun Services Substitution List). The part number of the motherboard can be determined by visual inspection of the part number label on the motherboard or by using the showfru SC command.
When running Sun Explorer 5.2 or greater, you must specify the Tx000 option to collect the data from the ALOM CMT commands on the Sun Fire T2000 server. The script is not run by default. The following example shows how to run the script.
For more details, refer to the troubleshooting document, Using Sun Explorer on the Tx000 Series Systems. This document is available on the SunSolve web site:
Coolthreads servers running SunVTS cpu tests have encountered Solaris watchdog timeouts leading to system shutdown.
Workaround: Set the ALOM CMT sys_autorestart variable to none while running SunVTS, so that ALOM CMT issues a warning message but does not reset the server.
POST error messages regarding unsupported memory configurations can be misleading. In situations where memory rank 0 (zero) is fully populated, the following message can be ignored safely.
ERROR: Using unsupported memory configuration
The implementation of the Solaris Predictive Self-Healing (PSH) software provided on this release of Sun Fire T2000 systems causes most systems to display a few erroneous error messages.
The following messages usually occur two or three times when the system is booted. These errors are logged and can be viewed with the fmdump command as shown in the following example:
These errors are not an indication of faulty devices. Once you confirm that your messages match the example shown, you can ignore them. If you see different error messages, contact your Sun Service representative for support.
If you boot from the hard drive (boot disk) after booting from the network
(boot net), and your server is using System Firmware Version 6.1.9, PSH fault messages might be displayed.
Disregard these messages. You can clear the messages from the PSH fault logs by following the instructions in the Sun Fire T2000 Server Service Manual (819-2548).
Example of the fault messages displayed at boot time:
Example of displaying the messages with the fmdump command:
Example of displaying the System Firmware Version from the Service Controller:
The Solaris PSH facility automatically detects the replacement of the motherboard and DIMMs. However, erroneous fault messages might be displayed when the system is booted, and these messages can mislead you to think that a problem persists when it is actually fixed. To correct this situation, you must install the Sun Fire T2000 mandatory patch, 119578-2
If you perform a Solaris JumpStart installation of a Sun Fire T2000 server, the server will display erroneous PSH fault messages at boot time. To correct this situation, you must install the Sun Fire T2000 mandatory patches and make changes to the /etc/system file. In addition, you should also clear the PSH and ALOM CMT fault logs to prevent the erroneous messages from being reported again. The steps to do this are described in To Configure the System After a JumpStart Installation.
Example of Erroneous Fault Messages at boot time:
Example of displaying the messages with the fmdump command:
This procedure describes how to configure the Sun Fire T2000 server after a jumpstart installation so that erroneous fault messages are not reported.
1. Install the mandatory patches on the server.
2. Update the /etc/system file.
See Mandatory /etc/system File Entries.
3. Use the fmadm faulty command to list the UUID of each erroneous fault.
4. Clear each fault that was listed in the preceding step.
5. Clear the persistent logs as shown in the following example.
6. Reset the Solaris PSH modules as shown.
7. Reset the faults at the ALOM CMT prompt:
a. Gain access to the ALOM CMT sc> prompt.
Refer to the Advanced Lights Out Management (ALOM) CMT v1.3 Guide for instructions.
b. Run the showfaults -v command to see the UUID of any faults.
c. Run the clearfault command with the UUID provided in the showfaults output:
8. If faults continue to be reported, the server might have a faulty component. Refer to the Sun Fire T2000 Server Service Manual for diagnostic procedures.
There is an error in the documentation of the showdate command in published versions of the ALOM CMT guide. The erroneous text follows:
Displays the ALOM CMT date. The Solaris OS and ALOM CMT time are synchronized, but ALOM CMT time is expressed in Coordinated Universal Time (UTC) rather than local time.
The correct text should be:
Displays the ALOM CMT date. ALOM CMT time is expressed in Coordinated Universal Time (UTC) rather than local time. The Solaris OS and ALOM CMT time are not synchronized.
There might be a typographical error in the translated versions of the Sun Fire T2000 Server Installation Guide. The error is not present in the English version.
The error is located in Chapter 2, in the section titled, "To Boot the Solaris Operating system, in the example in Step 2.
The incorrect example shows the following:
There is a space after the first / that should not be there.
The following line shows the correct example: