Installation of Linux on PRIMERGY TX150 S8 and SX150 S8


Contents

Linux drivers for PRIMERGY servers

An overview of all avaliable Linux documentation you can find here:
http://partners.ts.fujitsu.com/com/products/servers/primergy/opsys/linux
and here:
http://ts.fujitsu.com/products/standard_servers/os

An overview of all avaliable Linux readmes you can find here: http://www.fujitsu.com/fts/primergy-linux-docs.
For a more detailed description on the specifics of Linux drivers, the tool primesetup and errata kernels please have a look at the elaborate "PRIMERGY Linux Drivers Readme" document. You can find this information offline in the /Unix/Linux/Doc folder on the FUJITSU ServerView SuiteTM DVD-ROM. Here you can find the newest published version of this document: http://www.fujitsu.com/fts/primergy-linux-docs.
For a more detailed description on the specifics PRIMERGY server you find an overview here:
http://manuals.ts.fujitsu.com
and here:
http://support.ts.fujitsu.com/manuals

General hints

The ACPI states S1 and S4 are currently not supported under every Linux OS.

BIOS Configuration/Setup Utility

General BIOS setup [F2]


Released Distributions


SUSE Linux Enterprise Server 10 x86 & x86_64 (Service Pack 4 and newer)

Please read first the "Release Notes for SUSE LINUX Enterprise Server 10".

NOTE
i586 and i686 Machine with more than 16 GB of Memory
Depending on the workload, i586 and i686 machines with 16GB-48GB of memory can run into instabilities. Machines with more than 48GB of memory are not supported at all. To run on such a machine, lower the memory with the mem= kernel boot option.
In such memory scenarious we strongly recommend to use a x86_64 system with 64-bit SLES10 and run the x86 (32bit) applications on it.
Please see the "Release Notes for SUSE LINUX Enterprise Server 10 for x86".

Installation with FUJITSU drivers

NOTES

NOTE
If you plan to add additional storage devices to your system _after_ the OS installation, we strongly recommend to use persistent devicenames for all storage devices during installation. The installer by default uses the kernel device names.
Please see also the Release Notes for SUSE Linux Enterprise Server 10, which you can find on the first CD of SLES 10.

How to proceed:

During installation, enter the partitioner. For each partition, select "Edit" and go to the "FStab Options" dialog. Any mount option except "Device name" provide you persistent devicenames.

To switch an already installed system to using persistent device names, proceed as described above for all existing partitions. In addition, rerun the boot loader module in YaST to switch the bootloader to using the persistent device name also. Just start the module and select "Finish" to write the new proposed configuration to disk. This needs to be done before adding new storage devices.

For further information please look at http://en.opensuse.org/SDB:Persistent_storage_device_names.

NOTE
Systems using storage attached to Fibre Channel host bus adapters (FC HBAs) may encounter issues during installation due to PCI device load ordering.
If you install on a system with FC HBAs with attached storage devices please double-check the disk you have selected as your boot disk.
If you install on a system with FC HBAs without attached storage devices please check the driver order in the file /etc/sysconfig/kernel after installation. The driver for your boot controller has to be the first one. Afterwards rerun /sbin/mkinitrd and do a reboot. Otherwise your system will not boot after attaching storage devices to your FC HBAs.

NOTE
SLES10 SP2 x86 doesn't support kdump for memory > 4GB. See Novell support TID 7002347 http://www.novell.com/support/documentLink.do?externalID=7002347.

Install update kernel

SUSE Patch Support Database (PSDB)

A way to update your SLES installation is to download the necessary patches and errata kernels from the SUSE Patch Support Database located at
http://support.novell.com/linux/psdb
For a first access you have to register your business product by using the 14 digits code printed on the product CD box. After registering you can select the necessary RPMs as follows: ( example for kernel patch Release:20100826)

Install these RPMs as described in the "Installation notes" on this page.
Download the kernel version you need from this page.

Install the update kernel following your Linux distributor's instructions. Use primesetup to install the required FUJITSU drivers for the update kernel before rebooting your system. If you use the Novell Partner Linux Driver Process (PLDP) it is not necessary to install the required FUJITSU drivers again after a kernel update.
You find a detailed description of the tool primesetup in the document "PRIMERGY Linux Drivers Readme" and a description of the PLDP in the document "PRIMERGY Novell PLDP README" (http://www.fujitsu.com/fts/primergy-linux-docs ).

Hints and restrictions

Failure during start of XServer
After adding a PCI controller to an installed system it could happen that the XServer doesn't start.
This happens because on some systems the PCI device ID's will be changed after adding an additional controller to the system. The PCI ID for the graphic device is registered in the file /etc/X11/xorg.conf and this is not changed automatically.
There are two workarounds for this problem:
Method 1:
Press Alt-F2 and login into the ANSI terminal. Start the configuration tool "yast" and do a rescan for the graphic card and the monitor. After rescan , reboot the system.
Method 2:
Have a look at the error logfile which is created in the error situation. Look for the line "No matches Device section for instance (BusID PCI: x:xx:x)." Note the item (BusID PCI: x:xx:x).
Press Alt-F2 and login into the ANSI terminal. Edit the file /etc/X11/xorg.conf and change in the section "Device" the item in the line "BusID" with the item from the error logfile. Start the XServer with the command "startx".

Remote Deploy
Deployment of SUSE SLES 10 is only possible if the Mountby-Option "Device Name" is used instead of "Device ID".
During native installation the customer must set these option for every partition by hand.

Bad LAN performance
Is there a bad LAN performance with SLES 10 SP4 this workaround could solve the problem:
It is recommended to decrease the receive buffer in the Operating System. Please use these commands:
echo 4096 8192 16384 > /proc/sys/net/ipv4/tcp_wmem
echo 0 > /proc/sys/net/ipv4/tcp_window_scaling

NOTE
If you want to install with the PLDP Driver Kit you can not use the internal DVD-ROM drive. Please use an external USB CD/DVD-ROM drive.

Released host bus adapters

The following host bus adapters are released:

Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD. If there is no driver offered on the Support web page please use the native driver.

For configuration and administration of the Emulex FC HBA's please use the tool HBAnyware, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD.


Onboard LSI SAS/SATA SW RAID (driver megasr)

For the installation of the onboard LSI SAS/SATA SW RAID (MegaSR) controller the driver megasr is needed.

With a SAS key installed, this controller supports SATA as well as SAS disks. Without this key, only SATA disks are supported.

NOTE
Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerstartCD.

For the installation of the Intel Gigabit Adapters, the driver ixgbe is needed. Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://ts.fujitsu.com/support/downloads.html or on the FUJITSU ServerView Suite DVD. If there is no driver offered on the Support web page please use the native one.

For performance improvement the following actions are recommended:

For the installation of the Intel Gigabit Adapters, the driver ixgbe is needed. Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://ts.fujitsu.com/support/downloads.html or on the FUJITSU ServerView Suite DVD. If there is no driver offered on the Support web page please use the native one.

For performance improvement the following actions are recommended:


Graphic Adapter NVIDIA Quadro NVS 300 (driver nvidia)

General:

This graphic card NVIDIA Quadro NVS 300 is only released for SLES 10 SP4 with the Nvidia driver version 304.55 and higher.
None of the Linux Operating Systems supports the Nvidia NVS 300 with a native driver.
The implemented vesa driver allows the installation of the Operating System.
To benefit from the advantages and features of the Nvidia NVS 300 graphic card, a closed source driver is required and can be installed afterwards.

FUJITSU releases a closed source driver which can be downloaded from the FUJITSU support page:
http://support.ts.fujitsu.com/download

Nvidia releases frequently newer drivers on their support page:
http://www.nvidia.com/Download/index.aspx

The customer is welcomed to download the latest driver, Nvidia has to offer. However these drivers will not be supported by FUJITSU.

The NVIDIA driver doesn't support the XEN kernel.

How to install the OEM driver:
First make sure you have the correct package, which corresponds to the architecture of the Linux distribution.
To install the driver, switch to runlevel 3 (enter "init 3" in a terminal).
Execute the *.run package and follow the instructions on the screen.

X-Server Configuration:
Once the OEM driver is installed, The Linux X-Server configuration tools don't support modifications regarding resolution, monitor detection, DualView.
Please use the OEM application "NVIDIA X Server Settings" which is installed with the driver package.
e.g. execute the command "nvidia-settings" in a shell.

The Graphic card NVIDIA Quadro NVS 300 is only released in PCIe Slot 3, 5 and 6.


iSCSI-Boot

iSCSI-Boot is released for SLES 10 SP3 and newer.
iSCSI-Boot is not released for for the XEN kernel.
Installation option: The boot option for iSCSI-Boot is: withiscsi=1.
Kdump is not possible with iSCSI-Boot.
iSCSI-Boot does not support channel bonding and/or MultiPath SW.
iSCSI-Boot is released for the Onboard-LAN controller (LOM).

On using iSCSI and Emulex driver lpfc you will get error messages like:

Emulex LightPulse Fibre Channel SCSI driver 8.2.0.22
Copyright(c) 2004-2008 Emulex. All rights reserved.
ACPI: PCI Interrupt 0000:0b:00.0[A] -> GSI 16 (level, low) -> IRQ 169
PCI: Setting latency timer of device 0000:0b:00.0 to 64
scsi3 : on PCI bus 0b device 00 irq 169
irq 169: nobody cared (try booting with the "irqpoll" option)
Call Trace: <IRQ>
<ffffffff8015e118>__report_bad_irq+48
<ffffffff8015e321>note_interrupt+444
<ffffffff8015dbf4>__do_IRQ+191
<ffffffff8010d569>do_IRQ+59
<ffffffff80109e85>mwait_idle+0
<ffffffff8010b35e>ret_from_intr+0 <EOI>
<ffffffff80109e85>mwait_idle+0
<ffffffff80109ec6>mwait_idle+65
<ffffffff80109e57>cpu_idle+160
<ffffffff803ee832>start_kernel+626
<ffffffff803ee2a2>_sinittext+674
handlers:
<ffffffff880c0f7b>(lpfc_intr_handler+0x00x57e [lpfc])
Disabling IRQ #169
ACPI: PCI Interrupt 0000:0b:00.1[B] -> GSI 17 (level, low) -> IRQ 177
PCI: Setting latency timer of device 0000:0b:00.1 to 64

Workaround for this entry in /var/log/messages is:
Edit the /etc/sysconfig/kernel file and delete at the section INITRD_MODULES the "lpfc" driver name, save it and do "mkinitrd" to build a new initrd.
Only the build of new initrd without the driver lpfc will help. The driver lpfc will be loaded on hotplug and udev mechanismen.


SUSE Linux Enterprise Server 11 x86 & x86_64 (Service Pack 2 and newer)

Please read first the "Release Notes for SUSE Linux Enterprise Server 11": https://www.suse.com/releasenotes/x86_64/SUSE-SLES/11-SP4/.

NOTE
i586 and i686 Machine with more than 16 GB of Memory
Depending on the workload, i586 and i686 machines with 16GB-48GB of memory can run into instabilities. Machines with more than 48GB of memory are not supported at all. To run on such a machine, lower the memory with the mem= kernel boot option.
In such memory scenarious we strongly recommend to use a x86_64 system with 64-bit SLES 11 and run the x86 (32bit) applications on it.
Please see the "Release Notes for SUSE LINUX Enterprise Server 11".

Installation with FUJITSU drivers

Install update kernel

SUSE Patch Support Database (PSDB)

A way to update your SLES installation is to download the necessary patches and errata kernels from the SUSE Patch Support Database located at
http://support.novell.com/linux/psdb
For a first access you have to register your business product by using the 14 digits code printed on the product CD box. After registering you can select the necessary RPMs as follows: ( example for kernel patch Release:20100722)

Install these RPMs as described in the "Installation notes" on this page.
Download the kernel version you need from this page.

Install the update kernel following your Linux distributor's instructions. Use primesetup to install the required FUJITSU drivers for the update kernel before rebooting your system. If you use the Novell Partner Linux Driver Process (PLDP) it is not necessary to install the required FUJITSU drivers again after a kernel update.
You find a detailed description of the tool primesetup in the document "PRIMERGY Linux Drivers Readme" and a description of the PLDP in the document "PRIMERGY Novell PLDP README" (http://www.fujitsu.com/fts/primergy-linux-docs ).

Remote Deploy
Deployment of SUSE SLES 11 is only possible if the Mountby-Option "Device Name" is used instead of "Device ID".
During native installation the customer must set these option for every partition by hand.

Enabling the power capping functionality
To make the power meter device fully functional on PRIMERGY under SLES 11, use the kernel parameter "acpi_power_meter.force_cap_on=1".
If the sysfs files power_cap, power_cap_min, etc. are visible in sysfs, power capping is available.

Hints and restrictions

Installation using Persistent Device names
The installer uses persistent device names by default. If you plan to add additional storage devices to your system after the OS installation, we strongly recommend you use persistent device names for all storage devices.

To switch to persistent device names on a system that has already been installed, use the YaST2 partitioner. For each partition, select "Edit" and go to the "FStab Options" dialog. Any mount option except "Device name" provides you persistent device names. In addition, rerun the boot loader module in YaST to switch the bootloader to using the persistent device name. Just start the module and select "Finish" to write the new proposed configuration to disk. This needs to be done before adding new storage devices.

For further information please visit:
http://en.opensuse.org/SDB:Persistent_storage_device_names.

NOTE
Systems using storage attached to Fibre Channel host bus adapters (FC HBAs) may encounter issues during installation due to PCI device load ordering.
If you install on a system with FC HBAs with attached storage devices please double-check the disk you have selected as your boot disk.
If you install on a system with FC HBAs without attached storage devices please check the driver order in the file /etc/sysconfig/kernel after installation. The driver for your boot controller has to be the first one. Afterwards rerun /sbin/mkinitrd and do a reboot. Otherwise your system will not boot after attaching storage devices to your FC HBAs.

SNMP NOTE
The default setting for the logging of the SNMP daemon is "debug". This configuration is done via the parameter "SNMPD_LOGLEVEL" in the file /etc/sysconfig/net-snmp.
This configuration will increase the log file /var/log/net-snmpd.log.
To avoid the increasing of the log file edit the file /etc/sysconfig/net-snmp and change the logging from "d" (debug) to "n" (notice) or to "w" (warning).

Memory error handling: Disable the logging of correctable errors
It is recommended to use the kernel parameter "mce=ignore_ce" under SLES 11 SP2 and newer.
Otherwise the system may freeze if a broken DIMM continuously emits correctable error events.

Released host bus adapters

The following host bus adapters are released:

Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD. If there is no driver offered on the Support web page please use the native driver.

For configuration and administration of the Emulex FC HBA's please use the tool OneCommandManager (OCM), which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD.


Onboard Intel RST (driver isci)

Support for internal SAS backup devices (LTOx)

For the installation of the onboard Intel SAS controller(Intel Rapid Storage Technology RST) the driver isci is needed.
Please insure that in the system BIOS the Onboard-SAS/SATA (SCU) is enabled and the Onboard SAS/SATA OpROM Intel RSTe is enabled
A SAS key has to be installed on the Motherboard to support internal SAS backup devices (LTOx).


Onboard LSI SAS/SATA SW RAID (driver megasr)

For the installation of the onboard LSI SAS/SATA SW RAID (MegaSR) controller the driver megasr is needed.

With a SAS key installed, this controller supports SATA as well as SAS disks. Without this key, only SATA disks are supported.

NOTE
LSI OPROM has to be enabled in BIOS Setup to use the onboard LSI SAS SW Raid!
Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerstartCD.

NOTE Virtualization
Running virtual machines under the KVM and XEN hypervisors is unsupported with the megasr driver 16.x and later.


Graphic Adapter NVIDIA Quadro NVS 300 (driver nvidia)

General:

This graphic card NVIDIA Quadro NVS 300 is only released for SLES 11 SP1 and above.
None of the Linux Operating Systems supports the NVIDIA NVS 300 with a native driver.
The implemented vesa driver allows the installation of the Operating System.
To benefit from the advantages and features of the NVIDIA NVS 300 graphic card, a closed source driver is required and can be installed afterwards.

FUJITSU releases a closed source driver which can be downloaded from the FUJITSU support page:
http://support.ts.fujitsu.com/download

NVIDIA releases frequently newer drivers on their support page:
http://www.nvidia.com/Download/index.aspx

The customer is welcomed to download the latest driver, NVIDIA has to offer. However these drivers will not be supported by FUJITSU.

The NVIDIA driver doesn't support the XEN kernel.

How to install the OEM driver:
First make sure you have the correct package, which corresponds to the architecture of the Linux distribution.
To install the driver, switch to runlevel 3 (enter "init 3" in a terminal).
Execute the *.run package and follow the instructions on the screen.

X-Server Configuration:
Once the OEM driver is installed, The Linux X-Server configuration tools don't support modifications regarding resolution, monitor detection, DualView.
Please use the OEM application "NVIDIA X Server Settings" which is installed with the driver package.
For example execute the command "nvidia-settings" in a shell.

The Graphic card NVIDIA Quadro NVS 300 is only released in PCIe Slot 3, 5 and 6.


Graphic Adapter PGRA CP NVS 1GB VGA PCI-e x16 (NVIDIA NVS 315) (driver nvidia)

General:

This graphic card NVIDIA NVS 315 is only released for SLES 11 SP3 and above.
None of the Linux Operating Systems supports the NVIDIA NVS 315 with a native driver.
The implemented vesa driver allows the installation of the Operating System.
To benefit from the advantages and features of the NVIDIA NVS 315 graphic card, a closed source driver is required and can be installed afterwards.

FUJITSU releases a closed source driver which can be downloaded from the FUJITSU support page:
http://support.ts.fujitsu.com/download

NVIDIA releases frequently newer drivers on their support page:
http://www.nvidia.com/Download/index.aspx

The customer is welcomed to download the latest driver, NVIDIA has to offer. However these drivers will not be supported by FUJITSU.

The NVIDIA driver doesn't support the XEN kernel.

How to install the OEM driver:
First make sure you have the correct package, which corresponds to the architecture of the Linux distribution.
To install the driver, switch to runlevel 3 (enter "init 3" in a terminal).
Execute the *.run package and follow the instructions on the screen.

X-Server Configuration:
Once the OEM driver is installed, The Linux X-Server configuration tools don't support modifications regarding resolution, monitor detection, DualView.
Please use the OEM application "NVIDIA X Server Settings" which is installed with the driver package.
For example execute the command "nvidia-settings" in a shell.

The Graphic card NVIDIA NVS 315 is only released in PCIe Slot 4.


iSCSI-Boot

iSCSI-Boot is released for SLES 11 SP1 and newer.
iSCSI-Boot is not released for for the XEN kernel.
iSCSI-Boot does not support channel bonding and/or MultiPath SW.
Kdump is not possible with iSCSI-Boot.
iSCSI-Boot is released for the Onboard-LAN controller (LOM).
uEFI iSCSI boot for SLES11 SP3 is not possible


SUSE Linux Enterprise Server 12 x86_64

Please read first the "Release Notes for SUSE Linux Enterprise Server 12": https://www.suse.com/releasenotes/x86_64/SUSE-SLES/12/.

Installation with FUJITSU drivers

Install update kernel

SUSE Patch Support Database (PSDB)

A way to update your SLES installation is to download the necessary patches and errata kernels from the SUSE Patch Support Database located at
http://download.novell.com/patch/psdb/
For a first access you have to register your business product by using the 14 digits code printed on the product CD box. After registering you can select the necessary RPMs as follows:

Install these RPMs as described in the "Installation notes" on this page.
Download the kernel version you need from this page.

Install the update kernel following your Linux distributor's instructions. Use primesetup to install the required FUJITSU drivers for the update kernel before rebooting your system. If you use the Novell Partner Linux Driver Process (PLDP) it is not necessary to install the required FUJITSU drivers again after a kernel update.
You find a detailed description of the tool primesetup in the document "PRIMERGY Linux Drivers Readme" and a description of the PLDP in the document "PRIMERGY Novell PLDP README" (http://www.fujitsu.com/fts/primergy-linux-docs ).

Hints and restrictions

NOTE Problem with SNMP and 10 Gb LAN adapter D2755
If more than one port of the 10Gb LAN controller card D2755 is in use without an Ethernet SFP+ module or without a "direct attach" cable, the snmpd hangs in a loop and doesn't answer any more questions.
That means the system can no longer be managed with ServerView.

The root couse is in a ixgbe driver problem which certain ioctl() calls delayed by more than 1s.

A useful workaround is to ensure that at most one port without SFP+ module or "direct attach" cable is running.
Optionally, one must summarize the connections to as few cards and do not install the cards that are not required.

NOTE Video configuration
The SLES 12 video driver for iRMC S3/iRMC S4 (Kronos3/4 Matrox/MGA-G200e) works in 24bpp mode.
Therefore the highest resolutions (>1440x900) aren't available under SLES 12 with the Video Redirection via the iRMC.

Workaround:
Use the kernel parameter "mgag200.preferreddepth=16"

NOTE ServerView
Please update the openssl packages to be able to connect to SVOM / Remote Connector Service with updates provided by SUSE, e.g.:

libopenssl1_0_0-1.0.1i-2.12.7637.0.PTF.902714.x86_64.rpm
openssl-1.0.1i-2.12.7637.0.PTF.902714.x86_64.rpm.

Released host bus adapters

The following host bus adapters are released:

Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD. If there is no driver offered on the Support web page please use the native driver.

For configuration and administration of the Emulex FC HBA's please use the tool OneCommandManager (OCM), which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD.


Onboard Intel RST (driver isci)

Support for internal SAS backup devices (LTOx)

For the installation of the onboard Intel SAS controller(Intel Rapid Storage Technology RST) the driver isci is needed.
Please insure that in the system BIOS the Onboard-SAS/SATA (SCU) is enabled and the Onboard SAS/SATA OpROM Intel RSTe is enabled
A SAS key has to be installed on the Motherboard to support internal SAS backup devices (LTOx).


Onboard LSI SAS/SATA SW RAID (driver megasr)

For the installation of the onboard LSI SAS/SATA SW RAID (MegaSR) controller the driver megasr is needed.

With a SAS key installed, this controller supports SATA as well as SAS disks. Without this key, only SATA disks are supported.

NOTE
LSI OPROM has to be enabled in BIOS Setup to use the onboard LSI SAS SW Raid!
Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerstartCD.

NOTE driver update not possible
SLES 12 does not boot anymore after updating the MegaSR driver.

There are two workarounds for this problem available:
Workaround 1:
After the update, run "mkinitrd" manually.

Workaround 2:
Before the update, change /etc/zypp/zypp.conf as follows:
Replace the line
multiversion = provides:multiversion(kernel)
by
multiversion = kernel

NOTE Virtualization
Running virtual machines under the KVM and XEN hypervisors is unsupported with the megasr driver 16.x and later.


Graphic Adapter PGRA CP NVS 1GB VGA PCI-e x16 (NVIDIA NVS 315) (driver nvidia)

General:

None of the Linux Operating Systems supports the NVIDIA NVS 315 with a native driver.
The implemented vesa driver allows the installation of the Operating System.
To benefit from the advantages and features of the NVIDIA NVS 315 graphic card, a closed source driver is required and can be installed afterwards.

FUJITSU releases a closed source driver which can be downloaded from the FUJITSU support page:
http://support.ts.fujitsu.com/download

NVIDIA releases frequently newer drivers on their support page:
http://www.nvidia.com/Download/index.aspx

The customer is welcomed to download the latest driver, NVIDIA has to offer. However these drivers will not be supported by FUJITSU.

The NVIDIA driver doesn't support the XEN kernel.

How to install the OEM driver:
First make sure you have the correct package, which corresponds to the architecture of the Linux distribution.
To install the driver, switch to runlevel 3 (enter "init 3" in a terminal).
Execute the *.run package and follow the instructions on the screen.

X-Server Configuration:
The display settings configured in "nvidia-settings" do not persist.
For persistant modifications regarding resolution and monitor detection use the configuration tools shipped with SLES 12, e.g. Gnome Tools.

The Graphic card NVIDIA NVS 315 is only released in PCIe Slot 4.


iSCSI-Boot

iSCSI-Boot is released for SLES 12.
iSCSI-Boot is not released for for the XEN kernel.
iSCSI-Boot does not support channel bonding and/or MultiPath SW.
Kdump is not possible with iSCSI-Boot.
iSCSI-Boot is released for the Onboard-LAN controller (LOM).


Red Hat Enterprise Linux 5 x86 & x86_64 (Update 8 and newer)

Please read first the "Release Notes for Red Hat Enterprise Linux 5": https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/.

NOTE
During the Red Hat Enterprise Linux 5 installation, it can be challenging to identify individual storage devices in system configurations that include multiple storage adapters. This is particulary true for systems containing Fibre Channel adapters, because in many instances it is desirable to install Red Hat Enterprise Linux on local storage. The Red Hat Enterprise Linux 5 installation program addresses this issue by delaying the loading of the following modules until after all other SCSI devices have been loaded:

This results in any locally-attached SCSI device names starting with /dev/sda, /dev/sdb, and so on, with the FC-attached storage following.

NOTE for RHEL 5.2 and newer
If you are using an Installation Number when installing Red Hat Enterprise Linux 5.2 on a system with SATA boot controller (driver megasr, megaide, ft3xx), be sure to deselect the Virtualization package group during the installation. The Virtualization package group option installs the kernel-xen kernel.

Installation with FUJITSU drivers

NOTE for RHEL 5.2 and newer
Systems currently deployed with Red Hat Enterprise Linux 5.0 or 5.1 that have PCI-e Intel Network devices (e1000) may see a change in the enumeration of their network adapters if they reinstall or upgrade to 5.2 and newer on the same hardware.
With 5.2 came the PCI-e device driver for Intel based network adapters (e1000e), effectively splitting the e1000 into two different kernel modules.
When performing a fresh install, e1000 type devices will be detected first followed by e1000e. PCI based e1000 devices would be enumerated first followed by PCI-e (e1000e) devices. On board NICs in many of the new systems are PCI-e, therefore your previous eth0 device would be discovered later in the hardware discovery process.

Systems being upgraded to 5.2 and newer via yum or an upgrade-install would not be affected if the HWADDR statement is properly defined in /etc/sysconfig/network-scripts/ifcfg-ethx.
This parameter maps a specific hardware address to a specific ethN device number, and it is included by default with new installations of Red Hat Enterprise Linux.

When re-building previously deployed systems without HWADDR defined, users will need to modify their /etc/sysconfig/network-scripts/ifcfg-ethx files either after installation or in %post of their kickstart configuration to properly define their devices as desired. Otherwise, the NIC ordering will change.

NOTE for RHEL 5 x86
During the installation process, Anaconda will automatically choose the kernel package to be installed. The kernel selected by default does not allow Red Hat Enterprise Linux 5 to detect more than 4GB of RAM. As such, if your system has more than 4GB of RAM, you need to install the kernel-PAE variant of the kernel after installation.
Note that this does not apply when a virtual install is performed.

Install update kernel

NOTE
Red Hat and FUJITSU recommend updating the installed kernel with the latest errata kernel (patch kernel). Newer errata kernels will be released on a regular basis and should be applied.

Errata kernels for Red Hat Enterprise Linux 5 are available for registered users through the Red Hat Network located at https://access.redhat.com/.

Install the update kernel following your Linux distributor's instructions. Use primesetup to install the required FUJITSU drivers for the update kernel before rebooting your system.
For a more detailed description on the specifics of Linux drivers, the tool primesetup and errata kernels please have a look at the elaborate "PRIMERGY Linux Drivers Readme" document. You can find this information offline in the /Unix/Linux/Doc folder on the ServerView Suite TMDVD-ROM. Here you can find the newest published version of this document: http://www.fujitsu.com/fts/primergy-linux-docs.

Hints and restrictions

NOTE
If your monitor is connected to a FUJITSU KVM adapter S3-1621 during installation of RHEL 5.1 and newer no monitor is detected.
Please install in text mode in this configuration or connect the monitor directly to the system. After installation in text mode you can configure the graphical console with the configuration tool "system-config-display".

NOTE
The system configuration tool "system-config-display" does not work correctly on some PRIMERGY systems.
If the output on your graphical console is not correct, please edit the file /etx/X11/xorg.conf with an editor:
Please change in the section "Screen", Subsection "Display" the Color Depth to 16 with "Depth 16" and set the resolution with: Modes "1024x768"

NOTE
Very sporadically the system can hang during shutdown with the latest message:
cpu frequency scaling
This is a genuine RHEL 5 and RHEL 5.1 problem which will be fixed with the next Red Hat updates. This is a workaround for the problem:
Turn off the cpu power management via: chkconfig cpuspeed off

NOTE
Very sporadically the system can hang during boot with the latest message:

Starting udev:

This problem can be circumvented by setting the following options in the file /etc/modprobe.conf:

options intel-rng no_fwh_detect=-1

NOTE
The order of Network Interface Cards is wrong if there are more than 10 network devices in a system.
A workaround for this problem is described in the KBase article which is available at this URL: http://kbase.redhat.com/faq/FAQ_103_11673.
This problem is fixed with RHEL5.2.

NOTE
On some PRIMERGY system you can find messages like this in /var/log/messages:

usb 1-2: device descriptor read/all, error -71
or this:
usb 1-2: device not accepting address 2, error -71

The message is just a warning. The device is an USB 2.0 device that works only with the USB 2.0 driver. The error message is generated when the kernel tries to configure the device with USB 1.1. Device works fine with the USB 2.0 driver later.

NOTE
There are a lot of messages from the snmpd in the /var/log/messages if the ServerView agents are installed and if the systems is monitored by a ServerView S2 frontend / ServerView Operation Manager. The number of messages can be reduced by changing the log level of the snmpd.
Two configuration files have to be corrected:
in /etc/snmp/snmpd.conf please add the line
dontLogTCPWrappersConnects 1

and
in /etc/init.d/snmpd please remove the # at the beginning of the line and the option "-a"
#OPTIONS="-Lsd -Lf /dev/null -p /var/run/snmpd.pid -a"
This line has to appear then as:
OPTIONS="-Lsd -Lf /dev/null -p /var/run/snmpd.pid"

Finally restart the SNMP-Daemon with: "service snmpd restart"

NOTE for RHEL5 x86_64
If you want to use the kernel parameter "kernel_unknown_nmi_panic=1" to configure your system to react to the NMI button you have to disable the NMI watchdog with the kernel parameter "nmi_watchdog=0". Otherwise the system does not boot.
To disable the NMI watchdog is not required for RHEL5 x86.
Up from RHEL 5.5 it is no longer required to use the parameter "nmi_watchdog=0".

NOTE for RHEL 5.3
The snmpd writes the file /var/log/messages full with
"Received SNMP packet(s)..."

Workaround:
Modify in the file /etc/syslog.conf the line

*.info;mail.none;authpriv.none;cron.none /var/log/messages

to

*.notice;mail.none;authpriv.none;cron.none /var/log/messages

NOTE for RHEL 5.3
The CPU consumption rate grows about 60-100 percent periodically (every 10 sec) if the ServerView agents are installed on RHEL 5.3.
The kipmi0 process is consuming 60-100% CPU time in this configuration.

The observation is already known for some time. It was reported by several other departments and individuals. You might want to search the Internet with the keywords "kipmi0 cpu" or "kipmi0 cpu usage" to see a discussion of this issue. The "kipmi0" thread is a part of the system's native IPMI kernel module, it is _not_ part of SV Agents Linux. In RHEL 5, SV Agents Linux use per default the system' IPMI kernel module.

The SV Agents eecd process is an important and frequent user of the IPMI kernel module. eecd does trigger the observation, but is not the root cause for the obserevation.

The Linux community is said to have responded as follows: "[The consumed CPU time of kipmi0 does not present a problem]

http://lists.us.dell.com/pipermail/linux-poweredge/2008-October/037636.html

kipmi0 is effectively a very fast polling loop. When an IPMI command has been issued to the controller, kipmi0 polls on completion of that command, and then retrieves the bytes of the answer. Unfortunately, the hardware interface (KCS) has no interrupt one could use to avoid polling. [...] kipmi0 only runs when there are outstanding IPMI commands issued, and even then runs at a low CPU priority, so only uses "idle" cycles. While it appears strange, it's the best we can do with the hardware interface we've got."

In PRIMERGY with SV Linux Agents operating, there is regular IPMI traffic for various reasons, including periodical SEL scanning, status inquiry etc. The issue is triggered by SV Linux Agent (i.e. eecd) operation. It should also appear, if one stops eecd and runs regularly the systems's "ipmitool" to retrieve some IPMI information.

NOTE for kdump
On systems with Maxmem configurations the default configuration of the crashkernel:
crashkernel=128M@16M
does not work.

The base address should be increased to 32 nor 64 for Maxmem configurations, e.g.
crashkernel=384M@64M
The reason is that otherwise the "bootmem" memory map may overlap with the crashkernel area.
See also: https://access.redhat.com/kb/docs/DOC-33301.

NOTE for RHEL 5.4 XEN
The pci=nomsi kernel parameter is recommended for all RHEL5.4 XEN installations.
Background: MSI is only supported under RHEL5 XEN since 5.4. But in some situations the MSI support may cause problems such as system hangs. The problem is under investigation at Red Hat.

NOTE for kdump
Booting of 2nd kernel hangs up sporadically during kdump under high load.
This situation can be circumvented by the following workaround:
Append 8250.nr_uarts=0 to KDUMP_COMMANDLINE_APPEND= in /etc/sysconfig/kdump

Released host bus adapters

The following host bus adapters are released:

Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD. If there is no driver offered on the Support web page please use the native driver.

For configuration and administration of the Emulex FC HBA's please use the tool OneCommandManager (OCM) and for configuration and administration of the QLogic FC HBA's please use the tool SANsurfer, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD.


Onboard Intel RST (driver isci)

Support for internal SAS backup devices (LTOx)

For the installation of the onboard Intel SAS controller(Intel Rapid Storage Technology RST) the driver isci is needed.
Please insure that in the system BIOS the Onboard-SAS/SATA (SCU) is enabled and the Onboard SAS/SATA OpROM Intel RSTe is enabled
A SAS key has to be installed on the Motherboard to support internal SAS backup devices (LTOx).


Onboard LSI SAS/SATA SW RAID (driver megasr)

For the installation of the onboard LSI SAS/SATA SW RAID (MegaSR) controller the driver megasr is needed.

With a SAS key installed, this controller supports SATA as well as SAS disks. Without this key, only SATA disks are supported.

NOTE
Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerstartCD.

For the installation of the Intel Gigabit Adapters, the driver ixgbe is needed. Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://ts.fujitsu.com/support/downloads.html or on the FUJITSU ServerView Suite DVD. If there is no driver offered on the Support web page please use the native one.

For performance improvement the following actions are recommended:

For the installation of the Intel Gigabit Adapters, the driver ixgbe is needed. Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://ts.fujitsu.com/support/downloads.html or on the FUJITSU ServerView Suite DVD. If there is no driver offered on the Support web page please use the native one.

For performance improvement the following actions are recommended:


Graphic Adapter NVIDIA Quadro NVS 300 (driver nvidia)

General:

This graphic card NVIDIA Quadro NVS 300 is only released for RHEL5.6 and above.
None of the Linux Operating Systems supports the NVIDIA NVS 300 with a native driver.
The implemented vesa driver allows the installation of the Operating System.
To benefit from the advantages and features of the NVIDIA NVS 300 graphic card, a closed source driver is required and can be installed afterwards.

FUJITSU releases a closed source driver which can be downloaded from the FUJITSU support page:
http://support.ts.fujitsu.com/download

NVIDIA releases frequently newer drivers on their support page:
http://www.nvidia.com/Download/index.aspx

The customer is welcomed to download the latest driver, NVIDIA has to offer. However these drivers will not be supported by FUJITSU.

The NVIDIA driver doesn't support the XEN kernel.

How to install the OEM driver:
First make sure you have the correct package, which corresponds to the architecture of the Linux distribution.
To install the driver, switch to runlevel 3 (enter "init 3" in a terminal).
Execute the *.run package and follow the instructions on the screen.

X-Server Configuration:
Once the OEM driver is installed, The Linux X-Server configuration tools don't support modifications regarding resolution, monitor detection, DualView.
Please use the OEM application "NVIDIA X Server Settings" which is installed with the driver package.
For example execute the command "nvidia-settings" in a shell.

The Graphic card NVIDIA Quadro NVS 300 is only released in PCIe Slot 3, 5 and 6.


iSCSI-Boot

iSCSI-Boot is released for RHEL 5.5 and newer.
iSCSI-Boot is not released for the XEN kernel.
iSCSI-Boot does not support channel bonding and/or MultiPath SW.
Kdump is not possible with iSCSI-Boot.
iSCSI-Boot is released for the Onboard-LAN controller (LOM).
iSCSI-Boot is released only for LOM 1/2 (eth0/eth1). This is a limitation only for RHEL 5.5 and RHEL 5.6. The problem is fixed with RHEL 5.7

It is not possible to switch from eth0 to eth1 if the network is configured with dhcp.
Workaround for RHEL 5.5 and 5.6 (eth0 --> eth1):
- create a new initrd with the following options: "mkinitrd --net-dev=eth0 --net-dev=eth1"
- change the boot port in FW
- boot with the new initrd
Workaround for RHEL 5.7:
Edit the file "/etc/sysconfig/network-scripts/ifcfg-ethx" and build a new initrd.

It is not possible to switch from eth0 to eth1 if a static network configuration is used.
Workaround for RHEL 5.5 and 5.6 (eth0 --> eth1):
- change the network configuration to DHCP
- reboot the server
- switch from eth0 to eth1 (see workaround above)
- change the network configuration from DHCP to static
Workaround for RHEL 5.7:
Edit the file "/etc/sysconfig/network-scripts/ifcfg-ethx" and build a new initrd.


Red Hat Enterprise Linux 6 x86 & x86_64 (Update 3 and newer)

Please read first the "Release Notes for Red Hat Enterprise Linux 6": https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/.

Installation with FUJITSU drivers

Install update kernel

NOTE
Red Hat and FUJITSU recommend updating the installed kernel with the latest errata kernel (patch kernel). Newer errata kernels will be released on a regular basis and should be applied.

Errata kernels for Red Hat Enterprise Linux 6 are available for registered users through the Red Hat Network located at https://access.redhat.com/.

Install the update kernel following your Linux distributor's instructions.

Hints and restrictions

NOTE memory poisoning
RHEL6 supports memory poisoning (HWPOISON, aka uncorrectable recoverable memory errors), but only with the kernel parameter "transparent_hugepage=never".

NOTE FUJITSU ServerView software
The tools BIOS ASP's, ServerView Agents and ServerViewRAID have dependencies to the 32bit versions of the packages openssl, compat-libstdc++-33 and pam.
It is not possible to choose any 32bit packages during the system installation of the 64bit version of RHEL 6.
So you have to install the packages openssl.i686, compat-libstdc++-33.i686 and pam.i686 after the system installation.

NOTE Network configuration
With the NetworkManager which is default enabled in RHEL 6 it is not possible to configure a network configuration like channel bonding or bridges. And there is no other tool to do this.
So you have to disable the Networkmanager and to configure by hand-editing the ifcfg* files.
This is a limitation for RHEL 6.
But:
If adding a new LAN Card to an already existing OS installation it's easier to make the detection via NetworkManager because it automatically detects the new HW and script files are made.

NOTE Video configuration
There is only very limited support for switching color depth and resolution under RHEL6.
The following Red Hat knowledge base articles describe the procedure to change color depth if necessary:
How to create xorg.conf:
https://access.redhat.com/knowledge/node/5263
How to change the display depth of desktop:
https://access.redhat.com/kb/docs/DOC-43178
How do I add desired display mode in Red Hat Enterprise Linux 6:
https://access.redhat.com/knowledge/solutions/86603

Using this method has disadvantages, because the resulting static Xorg configuration will prevent plug-and-play operation of the X server (e.g. adding monitors on the fly).
Because Kronos 2 offers 1600x1200 pixel only at 16bpp, and X operates at 24/32 bpp by default, 1600x1200 can only be achieved by following the above procedure to change the color depth. Because this is cumbersome for users, so it is recommended to restrict Kronos2 support to max. 1280x1024 for RHEL6.

NOTE KDUMP configuration
Using the system-config-kdump tool is not recommended for configuring kdump, because the GUI is irritating and sometimes makes wrong settings. Hand-editing /etc/kdump.conf is easier and less error-prone.
This is a limitation for RHEL6.0.

When a "raw" partition is entered in /etc/kdump.conf, kdump does write a dump to the partition but the dump is not automatically recovered at the next reboot.
When raw partitions are used for dumping, the dump must be recovered manually after reboot using "makedumpfile -R" (e.g. makedumpfile -R vmcore < /dev/sdb).

The parameter "crashkernel=auto" as normally used in RHEL 6 is deprecated in 6.1.
The crashkernel size should neither be too small nor too big. Crashkernel=256M or crashkernel=384M seems to be a reasonable size for large servers.
See also https://access.redhat.com/kb/docs/DOC-59801
(this article recommends crashkernel=768M which may be too large for some systems).

Up from RHEL 6.2
"crashkernel=auto" is now recommended again by Red Hat for RHEL 6.2.

NOTE Automated Bug Reporting Tool (abrt)
The automated bug reporting tool (abrt), which automatically generates crash reports, is still quite immature.
This has been reported to Red Hat. It will not be fixed in RHEL6.0.

NOTE MultiBitError
After a MultiBitError during running RHEL6 x86_64, the iRMC detects an IERROR if KDUMP is enabled.
Workaround:
Add the kernel parameter "mce=off" to KDUMP_COMMANDLINE_APPEND in /etc/sysconfig/kdump.
Root cause: The kdump kernel clears the MCE status registers when it starts. This prevents the BIOS from determining the cause of the machine check event after reboot. "mce=off" disables the kernel MCE handler.

RESTRICTION driver megasr
The ACPI mode "standby" does not work , if the driver megasr is used for the boot controller.

NOTE RHEL 6.2 driver update
PROBLEM:
Upgrading RHEL6.0 or RHEL6.1 to RHEL6.2 kernel there are problems if using FUJITSU driver modules. Currently we have to take care for the modules:
smbus
megasr
be2net
be2iscsi

REASON / CAUSE
When a module compiled for RHEL6.0 or RHEL6.1 is loaded into a RHEL6.2 kernel, the data structures in the module image in memory can be corrupted. Depending on the memory layout of the module, this may lead to serious data corruption and possibly a kernel crash. FUJITSU is providing its modules custom-compiled for RHEL6.2, so that people who installed 6.2 from the start should be well off.

WORKAROUND:
To be sure that the current RHEL 6.2 kernel version matches the version of the kabi-modules - please check:

# rpm -q --provides kmod-smbus | grep kabi
kabi-modules = 2.6.32-220.el6.x86_64
#
# uname -r
2.6.32-220.el6.x86_64
#

This is an example above for smbus modules, for megasr use "--provides kmod-megasr"

FUJITSU is providing its modules custom-compiled for RHEL6.2, so that people who installed 6.2 from the start should be well off. Every user who upgrades from 6.1 to 6.2 must also update all kmods to those compiled cleanly against.

Red Hat is preparing a kernel update which will solve this problem. This kernel update will probably be released in February 2012.
For user with Red Hat Account: please refer the Knowledge Base Article https://access.redhat.com/kb/docs/DOC-69556.

NOTE RHEL 6.3 problem with the driver megasr
The "tboot" (Trusted Boot) package under RHEL 6.3 sets the kernel parameter "intel_iommu=on" by default. This may cause problems with the megasr driver which doesn't fully support the IOMMU (this has been reported to LSI).

Workarounds:
a) deselect the "tboot" package during installation
b) add the additional boot parameter "iommu=pt" on the GRUB kernel command line at the first reboot and later in /boot/grub/menu.lst.

Enabling the power capping functionality
To make the power meter device fully functional on PRIMERGY under RHEL 6, use the kernel parameter "acpi_power_meter.force_cap_on=1".
If the sysfs files power_cap, power_cap_min, etc. are visible in sysfs, power capping is available.

When trying to boot from RHEL 6.4 Installation DVD, a dark screen is displayed
The problem only exists if a NVidia NVS300 is installed in PRIMERGY systems and if BIOS setting "Above 4G Decoding=enabled".

The BIOS Post is displayed correctly, but instead of getting the initial graphical Installation Boot Menu of RHEL 6.4, a dark screen is displayed.

Workaround:
1. Presss ESC when dark screen is displayed, the boot prompt is displayed then.
2. Enter "linux", then installation of RHEL6.4 starts and graphic display will be available again!

The BIOS setting "Above 4G Decoding=enabled" currently is only necessary for systems with GPGPU/Intel Xeon Phi Coprocessor cards installed.
NVidia NVS300 together with GPGPU/Intel Xeon Phi Coprocessor cards is not a preferred configuration.

RESTRICTION RHEL6-U2 with D2745 adapter
When a 4-port LAN adapter D2745 is installed in the system, the OS installation of RHEL6-U2 is hanging upon software download. To accomplish the OS installation with installed adapter, following kernel option must be used: pci=nomsi
RedHat is working on this issue and will provide a kernel update soon.

Memory error handling: Disable the logging of correctable errors
It is recommended to use the kernel parameter "mce=ignore_ce" under RHEL6.2 and newer.
Otherwise the system may freeze if a broken DIMM continuously emits correctable error events.

Released host bus adapters

The following host bus adapters are released:

Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD. If there is no driver offered on the Support web page please use the native driver.

For configuration and administration of the Emulex FC HBA's please use the tool OneCommandManager (OCM) and for configuration and administration of the QLogic FC HBA's please use the tool SANsurfer, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD.


Onboard Intel RST (driver isci)

Support for internal SAS backup devices (LTOx)

For the installation of the onboard Intel SAS controller(Intel Rapid Storage Technology RST) the driver isci is needed.
Please insure that in the system BIOS the Onboard-SAS/SATA (SCU) is enabled and the Onboard SAS/SATA OpROM Intel RSTe is enabled
A SAS key has to be installed on the Motherboard to support internal SAS backup devices (LTOx).


Onboard LSI SAS/SATA SW RAID (driver megasr)

For the installation of the onboard LSI SAS/SATA SW RAID (MegaSR) controller the driver megasr is needed.

With a SAS key installed, this controller supports SATA as well as SAS disks. Without this key, only SATA disks are supported

NOTE
Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerstartCD.

NOTE RHEL 6.3 problem with the driver megasr
The "tboot" (Trusted Boot) package under RHEL 6.3 sets the kernel parameter "intel_iommu=on" by default. This may cause problems with the megasr driver which doesn't fully support the IOMMU (this has been reported to LSI).

Workarounds:
a) deselect the "tboot" package during installation
b) add the additional boot parameter "iommu=pt" on the GRUB kernel command line at the first reboot and later in /boot/grub/menu.lst.

NOTE Virtualization
Running virtual machines under the KVM hypervisor is unsupported with the megasr driver 16.x and later.


Graphic Adapter NVIDIA Quadro NVS 300 (driver nvidia)

General:

None of the Linux Operating Systems supports the NVIDIA NVS 300 with a native driver.
The implemented vesa driver allows the installation of the Operating System.
To benefit from the advantages and features of the NVIDIA NVS 300 graphic card, a closed source driver is required and can be installed afterwards.

RHEL 6 should be installed with the kernel parameter "nomodeset".
Otherwise the dmesg and /var/log/messages are filled with error messages regarding false EDID data.

FUJITSU releases a closed source driver which can be downloaded from the FUJITSU support page:
http://support.ts.fujitsu.com/download

NVIDIA releases frequently newer drivers on their support page:
http://www.nvidia.com/Download/index.aspx

The customer is welcomed to download the latest driver, NVIDIA has to offer. However these drivers will not be supported by FUJITSU.

How to install the OEM driver:
RHEL 6 is shipped with a "nouveau" driver, which has to be uninstalled, before installing the driver package.

How to uninstall the "nouveau" driver and install the OEM driver:

A detailed description can be found on the NVIDIA support page:
ftp://download.nvidia.com/XFree86/Linux-x86_64/340.58/README/commonproblems.html

X-Server Configuration:
Once the OEM driver is installed, the Linux X-Server configuration tools don't support modifications regarding resolution, monitor detection, DualView.
Please use the OEM application "NVIDIA X Server Settings" which is installed with the driver package.
For example execute the command "nvidia-settings" in a shell.

The Graphic card NVIDIA Quadro NVS 300 is only released in PCIe Slot 3, 5 and 6.


Graphic Adapter PGRA CP NVS 1GB VGA PCI-e x16 (NVIDIA NVS 315) (driver nvidia)

General:

None of the Linux Operating Systems supports the NVIDIA NVS 315 with a native driver.
The implemented vesa driver allows the installation of the Operating System.
To benefit from the advantages and features of the NVIDIA NVS 315 graphic card, a closed source driver is required and can be installed afterwards.

RHEL 6 should be installed with the kernel parameter "nomodeset".
Otherwise the dmesg and /var/log/messages are filled with error messages regarding false EDID data.

FUJITSU releases a closed source driver which can be downloaded from the FUJITSU support page:
http://support.ts.fujitsu.com/download

NVIDIA releases frequently newer drivers on their support page:
http://www.nvidia.com/Download/index.aspx

The customer is welcomed to download the latest driver, NVIDIA has to offer. However these drivers will not be supported by FUJITSU.

How to install the OEM driver:
RHEL 6 is shipped with a "nouveau" driver, which has to be uninstalled, before installing the driver package.

How to uninstall the "nouveau" driver and install the OEM driver:

A detailed description can be found on the NVIDIA support page:
ftp://download.nvidia.com/XFree86/Linux-x86_64/340.58/README/commonproblems.html

X-Server Configuration:
Once the OEM driver is installed, the Linux X-Server configuration tools don't support modifications regarding resolution, monitor detection, DualView.
Please use the OEM application "NVIDIA X Server Settings" which is installed with the driver package.
For example execute the command "nvidia-settings" in a shell.

The Graphic card NVIDIA NVS 315 is only released in PCIe Slot 4.


iSCSI-Boot

iSCSI-Boot is released for RHEL 6.1 and newer.
Please set the kernel parameter "ip=ibft netboot=iscsi iscsi_firmware" during the installation.
Disable I/OAT in system BIOS, it has no performance benefit anyway.
The NetworkManager has to be disabled.
iSCSI-Boot does not support channel bonding and/or MultiPath SW.
Kdump is not possible with iSCSI-Boot.
iSCSI-Boot is released for the Onboard-LAN controller (LOM).


Red Hat Enterprise Linux 7 x86_64

Please read first the "Release Notes for Red Hat Enterprise Linux 7": https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/.

Installation of RHEL 7.x
On some systems during the installation of RHEL 7.x this System Event Log (SEL) entry is created:
"Generic IPMI event: GenId '00', SensorType 'OS Boot', EventType '6F', SensorName '', EventData: 07 00 00"

This message can be ignored.

Installation with FUJITSU drivers

Install update kernel

NOTE
Red Hat and FUJITSU recommend updating the installed kernel with the latest errata kernel (patch kernel). Newer errata kernels will be released on a regular basis and should be applied.

Errata kernels for Red Hat Enterprise Linux 7 are available for registered users through the Red Hat Network located at https://access.redhat.com/.

Install the update kernel following your Linux distributor's instructions.

Hints and restrictions

NOTE Video configuration
The RHEL 7 video driver for iRMC S3/iRMC S4 (Kronos3/4 Matrox/MGA-G200e) works in 24bpp mode.
Therefore the highest resolutions (>1440x900) aren't available under RHEL 7 with the Video Redirection via the iRMC.

Workaround:
Use the kernel parameter "mgag200e.preferreddepth=16"

Released host bus adapters

The following host bus adapters are released:

Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD. If there is no driver offered on the Support web page please use the native driver.

For configuration and administration of the Emulex FC HBA's please use the tool OneCommandManager (OCM) and for configuration and administration of the QLogic FC HBA's please use the tool SANsurfer, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerView Suite DVD.


Onboard Intel RST (driver isci)

Support for internal SAS backup devices (LTOx)

For the installation of the onboard Intel SAS controller(Intel Rapid Storage Technology RST) the driver isci is needed.
Please insure that in the system BIOS the Onboard-SAS/SATA (SCU) is enabled and the Onboard SAS/SATA OpROM Intel RSTe is enabled
A SAS key has to be installed on the Motherboard to support internal SAS backup devices (LTOx).


Onboard LSI SAS/SATA SW RAID (driver megasr)

For the installation of the onboard LSI SAS/SATA SW RAID (MegaSR) controller the driver megasr is needed.

With a SAS key installed, this controller supports SATA as well as SAS disks. Without this key, only SATA disks are supported.

NOTE Virtualization
Running virtual machines under the KVM hypervisor is unsupported with the megasr driver 16.x and later.

NOTE
Please use the latest driver offered by FUJITSU, which you can find on the Support web page under http://support.ts.fujitsu.com/download or on the FUJITSU ServerstartCD.


Graphic Adapter PGRA CP NVS 1GB VGA PCI-e x16 (NVIDIA NVS 315) (driver nvidia)

General:

None of the Linux Operating Systems supports the NVIDIA NVS 315 with a native driver.
The implemented vesa driver allows the installation of the Operating System.
To benefit from the advantages and features of the NVIDIA NVS 315 graphic card, a closed source driver is required and can be installed afterwards.

FUJITSU releases a closed source driver which can be downloaded from the FUJITSU support page:
http://support.ts.fujitsu.com/download

NVIDIA releases frequently newer drivers on their support page:
http://www.nvidia.com/Download/index.aspx

The customer is welcomed to download the latest driver, NVIDIA has to offer. However these drivers will not be supported by FUJITSU.

How to install the OEM driver:
RHEL 7 is shipped with a "nouveau" driver, which has to be uninstalled, before installing the driver package.

How to uninstall the "nouveau" driver and install the OEM driver:

A detailed description can be found on the NVIDIA support page:
ftp://download.nvidia.com/XFree86/Linux-x86_64/340.58/README/commonproblems.html

X-Server Configuration:
The display settings configured in "nvidia-settings" do not persist.
For persistant modifications regarding resolution and monitor detection use the configuration tools shipped with RHEL 7, e.g. Gnome Tools.

The Graphic card NVIDIA NVS 315 is only released in PCIe Slot 4.


iSCSI-Boot

iSCSI-Boot is released for RHEL 7.
The NetworkManager has to be disabled.
iSCSI-Boot does not support channel bonding and/or MultiPath SW.
Kdump is not possible with iSCSI-Boot.
iSCSI-Boot is released for the Onboard-LAN controller (LOM).


XEN with RHEL 5 and SUSE SLES 10

XEN is only released on systems with more than 4 Gigabyte main memory.

SATA SW RAID (driver: megasr) is not supported with XEN.

If you are planning to install any fully virtualized guest OS, please check, if your system supports this mode. If it is supported you must enable the VT mode in BIOS setup. After this change you also have to power down the system and power it up again. Just rebooting is not sufficient on most systems to be able to install the fully virtualized guest OS.

Hints and restrictions

Wake on LAN with RHEL 5.7 and newer
It is not possible to poweron a system by "Wake on LAN" via an Intel LAN-Controller with a driver e1000, e1000e or igb if the XEN network bridge is configured for the LAN port you want to use.

Workaround:
Stop the network-bridge with the command
"/etc/xen/scripts/network-bridge stop" before poweroff the system.

CPU scaling with SLES 10 XEN
Cpufreq is not supported with SLES 10 XEN. The xenpm tool is not implemented in SLES 10 XEN and so the P-state switching is not working.


XEN with SUSE SLES 11

XEN is only released on systems with more than 4 Gigabyte main memory.

SATA SW RAID (driver megasr) is supported with SLES 11 XEN.

If you are planning to install any fully virtualized guest OS, please check, if your system supports this mode. If it is supported you must enable the VT mode in BIOS setup. After this change you also have to power down the system and power it up again. Just rebooting is not sufficient on most systems to be able to install the fully virtualized guest OS.

Hints and restrictions

Kdump with SLES 11 XEN
The kdump with SLES 11 XEN works only after changing the default value of the parameter "KDUMP_DUMPFORMAT".
Please change the value of the parameter from "compressed" to "ELF".

The parameter can be changed in the config file /etc/sysconfig/kdump or with YAST/kdump/Dump Filtering.

Kdump with SLES 11 SP3 XEN
The kdump with SLES 11 SP3 XEN works only after changing the default value of the parameter "KDUMP_DUMPLEVEL".
Please change the value of the parameter from "31" to "0".

The parameter can be changed in the config file /etc/sysconfig/kdump or with YAST/kdump/Dump Filtering.

In addition, the parameter "KDUMP_DUMPFORMAT" needs to be changed, please see above.

Kdump with SLES 11 XEN: system hangs
Sometimes it happens that a kdump is not written (local or via network).
If this issue occurs the system hangs and a manual reset is necessary.

Workaround:
The problem can be solved by setting the "nox2apic" parameter in /etc/sysconfig/kdump.

/etc/sysconfig/kdump:
KDUMP_COMMAND_LINE_APPEND="nox2apic"

CPU scaling with SLES 11 XEN
Cpufreq under Xen 3.3 isn't done by Dom0 but by the Hypervisor itself, using the xen command line option "cpufreq=xen". This option is appended to the "kernel" line in /boot/grub/menu.lst:

kernel /xen.gz-3.3.0 cpufreq=xen

Up from SLES 11 SP3 with Xen 4.2 the parameter "cpufreq=xen" is no longer required.

Cpufreq Control is done with the "xenpm" tool which is part of the xen-tools package under SLES 11.

Memory support for dom0 with SLES 11 XEN
Under SLES11, there is a general limit for Dom0 memory of 500GB.
Please see:
http://www.novell.com/docrep/2009/09/SUSE_Linux_Enterprise_Server_Virtualization_Technology_Support_en.pdf.

.


© FUJITSU
All rights reserved.
LINUX is a trademark of Linus Torvalds.
All hardware and software names are trademarks of their respective manufactures.
Linux_Readme_TX150S8_v2_7.html

Last update: June 20, 2016