Sun SPARC Enterprise M4000/M5000 Servers Product Notes |
This document includes these sections:
The following firmware and software versions are supported in this release:
If you plan to boot your Sun SPARC Enterprise M4000/M5000 server from a Solaris WAN boot server on the network, you must upgrade the wanboot executable. See Booting From a WAN Boot Server for details.
Note - For the latest information on supported firmware and software versions, see Software Resources. |
The following patches are mandatory for Sun SPARC Enterprise M4000/M5000 servers running Solaris 10 11/06 OS. These patches are not required for servers running Solaris 10 8/07 OS.
Note - The patches include a revision level, shown as a two-digit suffix. Check SunSolve.Sun.COM for the latest patch revision. See Software Resources for information on how to find the latest patches. |
Install the patches in the following order:
1. 118833-36
After installing patch 118833-36, reboot your domain before proceeding.
2. 125100-08
Install version 125100-08 at minimum. See the 125100-08 README file for a list of other patch requirements.
3. 123839-07
4. 120068-03
5. 125424-01
6. 118918-24
7. 120222-21
8. 125127-01
After installing patch 125127-01, reboot your domain before proceeding.
9. 125670-02
10. 125166-05
This section describes known hardware and software issues in this release.
Caution - For dynamic reconfiguration (DR) and hot-plug issues, see TABLE 9, Solaris Issues and Workarounds. |
This section describes hardware-specific issues and workarounds.
TABLE 1 lists known issues for which a defect change request ID has been assigned. The table also lists possible workarounds.
U320 PCIe SCSI card, part numbers 375-3357-01/02, is not supported in PCI cassettes for Sun SPARC Enterprise M4000/M5000 servers. Customers must use
375-3357-03 or later.
DIMMs are cold FRU replacement components. The entire server must be powered off and the power cords disconnected to replace the DIMMs.
You can mount up to 4 memory boards on the Sun SPARC Enterprise M4000 server and up to 8 memory boards on the Sun SPARC Enterprise M5000 server. The DIMMs on the memory board are grouped into group A and group B.
The following changes belong in the Sun SPARC Enterprise M4000/M5000 Servers Service Manual.
Caution - Do not forcethe PCI cassette into a slot. Doing so can cause damage to the cassette and server. |
1. Align the PCI cassette on the gray plastic guide and install it into the slot.
2. Lock the lever into place to seat the cassette.
3. Connect all cables to the PCI cassette and reconnect the cable management arm, if necessary.
The following information belongs in the Sun SPARC Enterprise Equipment Rack Mounting Guide.
After securing the cable management arm (CMA) to the Sun SPARC Enterprise M4000/M5000 server, attach the provided end caps to the rails.
1. Secure the CMA to the server.
Refer to the Sun SPARC Enterprise Equipment Rack Mounting Guide for information on installing the CMA to the server.
2. Attach the end caps onto the slide rails.
Note - If the CMA is not used, attach all end caps to the rails of the server. The SPARC Enterprise M4000 server uses two end caps. The SPARC Enterprise M5000 server uses four end caps. |
FIGURE 1 Installing End Caps on the Sun SPARC Enterprise M4000 Slide Rails
FIGURE 2 Installing End Caps on the Left Rear of the Sun SPARC Enterprise M5000 Slide Rails
3. Connect the power cables to the rear of the server and secure them with the cable retention clamps.
Caution - Do not connect the power cables to a power source at this time. |
4. Run the power cables beneath the CMA and secure them in place with tie wraps.
The power cables and infiniband cables should hang loosely in a service loop behind the server or the CMA might not be able to fully retract.
Note - If additional attachment points are required to route the cables, install the optional bracket kit. See Installing the Extra Brackets (Optional). |
5. Ensure that the server can slide in and out of the equipment rack without dislodging the power cables.
FIGURE 3 shows how the CMA extends and retracts.
FIGURE 3 CMA Extended and Retracted on the Sun SPARC Enterprise M5000 Server
6. Slide the server into the equipment rack.
7. Tighten the four (4) captive screws at the front of the server to secure the server in the equipment rack.
8. Replace the rack stabilizer to its original position.
The following information belongs in the Sun SPARC Enterprise Equipment Rack Mounting Guide.
If additional attachment points are required to route the cables, you can install the extra brackets that are in the bracket kit. The bracket kit contains the following:
These brackets can be used with or without the CMA for the Sun SPARC Enterprise M4000/M5000 servers.
1. Extend the rack stabilizer.
2. Slide the server out of the rack several inches for access to the rear of the Sun Rack.
3. Position the cage nuts behind the threaded ears of the Sun Rack and insert the two (2) screws through the bracket and rack ear (FIGURE 4).
Brackets should be positioned near the top level of the server or slightly below it.
Note - Brackets can be installed one per side, one only (right or left side), or two on one side, as desired for convenience in cable management. |
4. Twist the cage nuts onto the screws from behind the rack ears.
The flat edges of the cage nuts should be aligned with the rack post to prevent the server from scraping against it.
FIGURE 4 Installing the Extra Brackets in a Sun Rack 1000
5. Insert velcro strips in the desired slots of the bracket to hold back cables.
Built-in cutouts along the sides of the Sun Rack can also be used to insert velcro strips to hold back cables, as desired.
6. Slide the server into the equipment rack.
7. Replace the rack stabilizer to its original position.
The following information belongs in the Sun SPARC Enterprise Equipment Rack Mounting Guide.
To ensure redundant power sourcing, use the provided wiring configurations for the Sun SPARC Enterprise M4000/M5000 servers in a Sun Rack 1000 38/42.
The Sun Rack 1000-38/42 can fit up to two modular power supplies (MPS). Each MPS is two rack units tall. The MPS must be installed into the bottom of the rack.
Note - The numbering in a Sun Rack reads from bottom to top and right to left. |
FIGURE 5 Sun Rack 1000 With Six Sun SPARC Enterprise M4000 Servers and One MPS
Note - The numbering in a Sun Rack reads from bottom to top and right to left. |
FIGURE 6 Sun Rack 1000 With Three Sun SPARC Enterprise M5000 Servers and Two MPS
The following changes belong in the Sun SPARC Enterprise M4000/M5000 Servers Site Planning Guide and the Sun SPARC Enterprise M4000/M5000 Servers Service Manual.
This section contains late-breaking hardware information that became known after the documentation set was published.
TABLE 7 lists known documentation updates.
External I/O Expansion Unit -- A rackmountable device to add on PCI slots. It is connected to the system’s I/O unit through the PCIe connection and contains one or two I/O boats. I/O boat -- An I/O unit in the External I/O Expansion Unit. The I/O boat connects to a PCI-Express (PCIe) slot through a PCIe switch or a PCI-X bridge on the I/O boat and offers either six PCI-X slots or six PCIe slots. |
||
Sun SPARC Enterprise M4000/M5000 Servers Site Planning Guide |
TABLE 1-3 “Midrange Servers Physical Specifications“ Correct numerical value of “Depth” is 810mm/31.9 in. for the Sun SPARC Enterprise M4000/M5000 servers. |
|
TABLE 2-2 “Midrange Servers Electrical Specifications“ See Electrical Specifications in these Product Notes for the changes. |
||
8.1.3, “Installing the PCI Cassette” See Installing the PCI Cassette in these Product Notes for the changes. |
||
See DIMM Replacement in these Product Notes for the changes. |
||
TABLE C-5 “Power Supply Feature” See Electrical Specifications in these Product Notes for the changes. |
||
Sun SPARC Enterprise M4000/M5000 Servers Installation Manual |
3.3, “Connecting the Administration Console”. The RJ-11 connector at the top of Figure 3-1 was not labelled. The RJ-11 connector is not for connection to TNV circuits. Do not use this connector. |
|
In Table 1-1, “Main Unit Specification” The following information will be added. |
This section describes specific software and firmware issues and workarounds.
TABLE 8 lists known XCP issues and possible workarounds.
TABLE 9 lists Solaris issues and possible workarounds.
Using the cfgadm -c disconnect command on the following cards might hang the command:
|
Do not perform cfgadm -c disconnect operation on the affected cards. |
|
The DAT72 internal tape drive might time out during tape operations. The device might also be identified by the system as a QIC drive. |
Add the following definition to /kernel/drv/st.conf: |
|
If you create a Solaris Flash archive on a non-Sun SPARC Enterprise M4000/M5000 sun4u server and install it on a Sun SPARC Enterprise M4000/M5000 sun4u server, the console’s TTY flags will not be set correctly. This can cause the console to lose characters during stress. |
Just after installing Solaris OS from a Solaris Flash archive, telnet into the Sun SPARC Enterprise M4000/M5000 server to reset the console’s TTY flags as follows: # sttydefs -r console
|
|
On-board Gigabit Ethernet NVRAM corruption could occur due to a race condition. |
If the NVRAM is corrupted, the device is not recognized as a network device. Contact your service representative to replace the FRU. |
|
The use of a PCIe Dual-Port Ultra320 SCSI controller card (SG-(X)PCIE2SCSIU320Z) in IOU Slot 1 on a Sun SPARC Enterprise M4000/M5000 server might result in a system panic. |
Do not use this card in IOU Slot 1 on a Sun SPARC Enterprise M4000/M5000 server. |
|
Using the DR deleteboard command while psradm operations are running on a domain might cause a system panic. |
There is no workaround. Check for the availability of a patch for this defect. |
|
A large number of spurious PCIe correctable errors can be recorded in the FMA error log. |
To mask these errors, add the following entry to the /etc/system file and reboot the system: |
|
When using the PCIe Dual-Port Ultra320 SCSI controller card (SG-(X)PCIE2SCSIU320Z), a PCIe correctable error causes a Solaris panic. |
Add the following entry to /etc/system to prevent the problem: |
|
Set the maximum size of the ZFS ARC lower. For detailed assistance contact Sun Service. |
||
The showhardconf(8) command on the XSCF cannot display PCI card information that is installed in the External I/O Expansion Unit, if the External I/O Expansion Unit is configured using PCI hot-plug. |
There is no workaround. When each PCI card in the External I/O Expansion Unit is configured using PCI hotplug, the PCI card information is displayed correctly. |
|
DR addboard command can hang. Once problem is observed, further DR operations are blocked. Recovery requires reboot of the domain. |
There is no workaround. Check for the availability of a patch for this defect. |
|
The error message network initialization failed can appear repeatedly after boot net installation. |
||
Make sure you have the correct /etc/system parameter and reboot the system:
|
||
There is a low probability of a domain panic during reboot when the Sun Quad GbE UTP x8 PCIe (X4447A-Z) card is present in a domain. |
There is no workaround. Check for the availability of a patch for this defect. |
|
Do not use the following I/O cards for network access when you are using the boot net install command to install the Solaris OS: |
When running Solaris 10 11/06, use an alternate type of network card or onboard network device to install the Solaris OS via the network. |
|
There is no workaround. Check for the availability of a patch for this defect. |
||
If the system has detected Correctible Memory Errors (CE) at power-on self-test (POST), the domains might incorrectly degrade 4 or 8 DIMMs. |
Increase the memory patrol timeout values used via the following setting in /etc/system and reboot the system: |
|
The system panics when running hot-plug (cfgadm) and DR operations (addboard and deleteboard)on: |
There is no workaround. Check for the availability of a patch for this defect. |
|
The system panics when running hot-plug (cfgadm) to configure a previously unconfigured card. The message "WARNING: PCI Expansion ROM is not accessible" will be seen on the console shortly before the system panic. The following cards are affected by this defect: |
DO NOT use cfgadm -c unconfigure to disconnect the I/O card. Perform cfgadm -c disconnect to completely remove the card. After waiting at least 10 seconds, the card might be configured back into the domain using the cfgadm -c configure command. |
|
Messages of the form nxge: NOTICE: nxge_ipp_eccue_valid_check: rd_ptr = nnn wr_ptr = nnn will be observed on the console with the following cards: |
||
The system panics when DiskSuite cannot read the metadb during DR. This bug affects the following cards: |
Panic can be avoided when a duplicated copy of the metadb is accessible via another Host Bus Adaptor. Or you can apply patch
|
|
Hot-plug operation with the following cards might fail during back-to-back disconnect and connect operations: |
After disconnecting a card, wait for a few seconds before re-connecting. |
|
Hot-plug operations on Sun Crypto Accelerator (SCA) 6000 cards can cause Sun SPARC Enterprise M4000/M5000 servers to panic or hang. |
Hot-plug does not work for the SCA6000 when running version 1.0 of the SCA6000 driver and should not be attempted. Version 1.1 of the SCA6000 driver and firmware supports hot-plug operations after the required bootstrap firmware upgrade has been performed. |
|
Performing a DR deleteboard operation on a board which includes Permanent Memory when using the following network cards results in broken connections: |
Re-configure the affected network interfaces after the completion of the DR operation. For basic network configuration procedures, refer to the ifconfig man page for more information. |
|
After a successful CPU DR deleteboard operation, the system panics when the following network interfaces are in use: |
Add the following line to /etc/system and reboot the system: |
|
Use of the following cards have been observed to cause data corruption in stress test under laboratory conditions: |
Add the following line in /etc/system and reboot the system: |
|
Do not start an XSCF failover while a DR operation is running. Wait for a DR operation to finish before starting the failover. If you start the failover first, wait for the failover to finish before starting the DR operation. |
||
After using the addfru or replacefru command to hotplug a CMU, further DR operations might fail with a misleading message regarding the board being unavailable for DR. |
When performing the addfru and replacefru commands, it is mandatory to run diagnostic tests. If you forget to run the diagnostic tests during addfru/addfru then either run testsb to test the CMU or remove the CMU/IOU with the deletefru command and then use the addfru command with the diagnostic tests. |
|
The DR addboard command might cause a system hang if you are adding a Sun StorageTek Enterprise Class 4Gb Dual-Port Fibre Channel PCI-E HBA card (SG-XPCIE2FC-QF4) at the same time that an SAP process is attempting to access storage devices attached to this card. The chance of a system hang is increased if the following cards are used for heavy network traffic: |
There is no workaround. Check for the availability of a patch for this defect. |
|
Unsuccessful DR operation leaves memory partially configured. |
To recover, add the board back to the domain with an addboard -d command and then retry the deleteboard command. |
The following step must be completed prior to upgrading:
Delete any accounts named admin.
Use the showuser -lu command to list all XSCF accounts. Any accounts named admin must be deleted prior to upgrading to XCP 1050. This account name is reserved in XCP 1050 and higher. Use the deleteuser command to delete the account.
Note - For more information on admin accounts, see TABLE 10, Software Documentation Updates. |
Note - LAN connections are disconnected when the XSCF resets. Use the XSCF serial connection to simplify the XCP upgrade procedure. |
1. Log in to the XSCF on an account with platform administrative privileges.
2. Verify that there are no faulted or deconfigured components by using the showstatus command.
The showstatus prompt will return if there are no failures found in the System Initialization. If anything is listed, contact your authorized service representative before proceeding.
4. Confirm that all domains are stopped:
5. Move the key position on the operator panel from Locked to Service.
6. Collect an XSCF snapshot to archive the system status for future reference.
7. Upload the XCP 1050 upgrade image by using the command line getflashimage.
The BUI on the XSCFU can also be used to upload the XCP 1050 upgrade image.
8. Update the firmware by using the flashupdate (8) command.
XSCF> flashupdate -c update -m xcp -s1050 |
Specify the XCP version to be updated. In this example, it is 1050.
9. Confirm completion of the update.
Confirm no abnormality happens while updating the XSCF.
10. Confirm that both the current and reserve banks of the XSCFU display the updated XCP versions.
If the Current and Reserve banks on the XSCF do not indicate XCP revision 1050, contact your authorized service representative.
11. Confirm the newly introduced ’servicetag’ facility is enabled.
When a system is upgraded from XCP 104x to XCP 1050, the newly introduced ’servicetag’ facility is not automatically enabled.
a. Check the ’servicetag’ facility status by using the showservicetag CLI.
b. If it is currently disabled, you must enable it.
c. An XSCF reboot is required for the ’servicetag’ facility to be enabled.
Note - Service tags are used by Sun Service. Fujitsu customers cannot enable service tags. |
d. Wait until XSCF firmware reaches the ready state.
This can be confirmed when the READY LED of the XSCF remains lit, or the message ’XSCF Initialize complete’ appears on the serial console.
12. Turn off all of the server’s power switches for 30 seconds.
13. After 30 seconds, turn the power switches back on.
14. Wait until the XSCF firmware reaches the ready state.
This can be confirmed when the READY LED of the XSCF remains lit.
15. Log in on to the XSCFU using a serial connection or LAN connection.
16. Confirm no abnormality occurred by using showlogs error -v and showstatus commands.
If you encounter any hardware abnormality of the XSCF contact your authorized service representative.
18. Log in to the XSCFU and confirm all domains start up properly.
19. Check that there are no new errors.
20. Move position of the key switch on the operator panel from Service to Lock.
2. Type the following command:
The following example shows a display of the showdevices -d command where 0 is the domain_id.
The entry for column 4 perm mem MB indicates the presence of permanent memory if the value is non-zero.
The example shows permanent memory on 00-2, with 1674 MB.
If the board includes permanent memory and executes the deleteboard command or moveboard command, the following notification is displayed:
To support booting the Sun SPARC Enterprise M4000/M5000 server from a WAN boot server:
1. Install the Solaris 10 11/06 OS on the WAN boot server.
2. Copy the wanboot executable from that release to the appropriate location on the install server. If you need further instructions, refer to the Solaris 10 Installation Guide: Network-Based Installations or refer to:
http://docs.sun.com/app/docs/doc/817-5504/6mkv4nh65?a=view
3. Create a WAN boot miniroot from the Solaris 10 11/06 OS. If you need further instructions, refer to:
http://docs.sun.com/app/docs/doc/817-5504/6mkv4nh63?a=view
If you do not upgrade the wanboot executable, the Sun SPARC Enterprise M4000/M5000 server will panic, with messages similar to the following:
krtld: load_exec: fail to expand cpu/$CPUkrtld: error during initial load/link phasepanic - boot: exitto64 returned from client program
See http://docs.sun.com/app/docs/doc/817-5504/6mkv4nh5i?a=view for more information on WAN boot.
In XCP 105x, the command getflashimage is available, which can be used to download firmware images in place of the XSCF Web.
This section contains late-breaking information on the software documentation that became known after the documentation set was published.
Copyright © 2007, Sun Microsystems, Inc. All Rights Reserved.