Go to main content

Sun Server X4-2L Product Notes

Exit Print View

Updated: May 2019
 
 

Linux Operating Systems and Virtual Machine Open Issues

Network port name and onboard NIC port name are not accurate after system reboots

Bug ID 22849478

Issue:

When installing Oracle Linux 7.x or Red Hat Enterprise Linux 7.x using Oracle System Assistant, the network port name is named “ethX,” based on the Legacy naming policy. If the server is configured with one to four Sun Quad Port GbE PCIe 2.0 Low Profile Adapter UTP cards (7100477), the port name for this card and the onboard NIC might become inaccurate after the system reboots. For example, if the onboard NIC0 is named “eth0” and the Low Profile Adapter UTP card port0 is named “eth8” at first boot, after a few reboots, the onboard NIC0 might become named “eth8” and the Low Profile Adapter UTP card port0 might become named “eth0.” The name might change again after more reboots. A similar issue can occur on the other onboard NIC and Low Profile Adapter UTP card port.

Affected software:

  • Oracle Linux 7.x

  • Red Hat Enterprise Linux 7.x

  • Platform software release 1.5 and later.

Workaround:

Change the kernel parameter from net.ifnames=0 to net.ifnames=1 to keep the consistent network device naming policy enabled. Do this in the kernel parameter according to the Legacy BIOS mode or UEFI mode you are using:

  • For Legacy BIOS mode, change the kernel parameter in /boot/grub2/grub.cfg

  • For UEFI mode, change the kernel parameter in /boot/efi/EFI/redhat/grub.cfg

When this change is made, the network naming will follow consistent network device naming policy and the issue will be resolved.

VMware ESXi 5.0 Update 2 virtual machine software does not support network connections when using the onboard network interface controller (NIC) unless an ixgbe driver or a compatible network interface card is installed

BugID: None

Issue:

The driver needed for the 10 Gigabit Ethernet (10GbE) Controller (X540) as used in the Sun Server X4-2L is not available in the general availability release of VMware ESXi 5.0 Update 2.

Affected software:

  • VMware ESXi 5.0 Update 2 and Update 3

  • Platform software releases 1.0 and later.


Note -  This issue does not affect VMware ESXi 5.1 Update 1 or 5.5.

Workarounds:

Perform either of the following tasks:

  • Add the required driver to the ESXi 5.0 Update 2 installation ISO image.

  • Install a PCIe network interface controller (NIC) card in the server that is compatible with the standard download ESXi 5.0 Update 2 ISO image.

For instructions for performing the above tasks, see “Configuring the VMware ESXi Software or the Server Hardware to Support Network Connections” in the Sun Server X4-2L Installation Guide for VMware ESXi.

When the operating system boots and loads the mpt2sas driver, errors might appear in dmesg

BugID: 15824191 (formerly CR 7205850)

Issue:

When the operating system boots and initializes the mpt2sas driver, several AER (application error reporting) errors might appear in dmesg. dmesg displays the content of the Linux system message buffer.

Affected software:

  • Oracle VM 3.1.1

  • SUSE Linux Enterprise Server (SLES) 11 SP2 and SP3

  • Platform software releases 1.0 and later.

Workaround:

None. These errors are harmless and can be safely ignored.

If you use the preinstalled Oracle VM 3.2 software on a server that has 26 storage drives, the Oracle VM software might panic and display error messages

BugID: 17162275

Issue:

If the server is configured with 26 storage drives (24 front and 2 rear), the preinstalled Oracle VM software might panic with the following message:

mount: could not find filesystem '/dev/root'

Affected hardware and software:

  • Any server configured with 26 storage drives

  • Oracle VM 3.2

  • Platform software releases 1.0 and later.

Workaround:

To fix this problem, perform the following steps:

  • Reset the Oracle VM software.

  • When GRUB (boot loader) is loaded, edit the line with "vmlinuz..." to remove the "sync_console" parameter.

  • After Oracle VM boots successfully, edit the "/boot/grub/grub.conf" file and remove any occurrences of the “sync_console" string from the boot sections.

Operating system does not boot if the Sun Storage Dual 16 Gb Fibre Channel PCIe Universal HBA FCoE card is installed and its option ROM is enabled in UEFI BIOS mode

BugID: 16721610

Issue:

When the Sun Storage Dual 16 Gb Fibre Channel PCIe Universal HBA (host bus adapter) FCoE (Fibre Channel over Ethernet) card is installed. and its option ROM is enabled in Unified Extensible Firmware Interface (UEFI) BIOS mode, the system boots to the GRUB menu and starts to load the system kernel, and then it fails to load the Linux operating system.

Affected hardware and software:

  • Sun Storage Dual 16 Gb Fibre Channel PCIe Universal HBA, QLogic

  • Sun Storage Dual 16 Gb Fibre Channel PCIe Universal HBA, Emulex

  • Oracle Linux 6.3 and 6.4

  • SUSE Linux Enterprise Server (SLES) 11 SP2 and SP3

  • Platform software releases 1.0 and later.

Workaround:

  1. Boot the system in UEFI BIOS mode.

  2. From the system BIOS, disable the Sun Storage Dual 16 Gb Fibre Channel PCIe Universal HBA card option ROM.

  3. Save changes and exit from the system BIOS.

  4. Install the Linux operating system.

After an Oracle Linux 5.9 or 6.4 OS boots when a power limit is set, the processor remains at low speed after commands are issued to dismiss the power limit

BugID: 16728705/17181067

Issue:

After the OS is booted or the server is power cycled, the Oracle Linux 5.9 or 6.4 operating system (OS) fails to execute _PSS, and _PPC commands to dismiss the power limit. As a result, the processor fails to run at maximum speed.

Affected software:

  • Oracle Linux 5.9 and 6.4

  • Platform software releases 1.0 and later.

Workaround:

  1. Acquire the CPU ID to be modified. Enter: # cat /proc/cpuinfo | grep processor

  2. Obtain the maximum supported CPU frequency. Enter: # cat /sys/devices/system/cpu/cpu<N>/cpufreq/cpuinfo_max_freq

    Where <N> is the CPU ID to be modified from Step 1.

  3. Check that the scaling_max_freq for each CPU ID is less than the maximum supported CPU frequency obtained in Step 2. Enter: # cat /sys/devices/system/cpu/cpu<N>/cpufreq/scaling_max_freq

  4. Tell the kernel the new scaling maximum frequency for each CPU ID you have, where Step 3 shows a reduced value compared with Step 2. Enter: #echo $max_frequency > /sys/devices/system/cpu/cpu<N>/cpufreq/scaling_max_freq

    Where $max_frequency is the determined maximum frequency shown in Step 2.

  5. Verify the new scaling maximum frequency. Enter: # cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo/scaling_max_freq

Attempts to install Oracle Linux 6.3 or 6.4 to an iSCSI target HDD fail

BugID: 15807672

Issue:

If you select UEFI BIOS Boot Mode when preparing either the Oracle Linux 6.3 or 6.4 operating system for installation, and then attempt to install the OS to an iSCSI target hard disk drive (HDD), the installation fails.

Affected software:

  • Oracle Linux 6.3 and 6.4

  • Platform software releases 1.0 and later.

Workaround:

Set the Oracle Linux 6.3 or 6.4 OS boot mode to Legacy BIOS Boot Mode, and then do the installation.

After enabling Option ROM for the Sun Storage Dual 16 Gb Fibre Channel PCIe Universal HBA QLogic card in UEFI BIOS Boot Mode, the SLES 11 ISO image fails to boot

BugID: 16817765

Issue:

The SUSE Linux Enterprise Server (SLES) 11 ISO image fails to boot when the Sun Storage Dual 16 Gb Fibre Channel PCIe Universal HBA QLogic card is installed in the server, Option ROM is enabled, and the UEFI BIOS Boot Mode is selected.

Affected hardware and software:

  • Sun Storage Dual 16 Gb Fibre Channel PCIe Universal HBA, QLogic (7101674)

  • SUSE Linux Enterprise Server 11 SP2 and SP3

  • Platform software releases 1.0 and later.

Workaround:

In the BIOS settings, disable the OpROM option of the PCIe slot that contains the Sun Storage Dual 16 Gb Fibre Channel PCIe Universal HBA, QLogic.

On x86 servers running Oracle Linux 6.4 with Intel Xeon E5-2600 V2 series processors installed, the processors never enters some C-states

BugID: 16870068

Issue:

On x86 servers running Oracle Linux 6.4 with Intel Xeon E5-2600 V2 series processors installed, the processors never enters C3 and C6 processor C-states.

Affected hardware and software:

  • Intel Xeon E5-2600 V2 series processors

  • Oracle Linux 6.4

  • Platform software releases 1.0 and later.

Workaround:

Append kernel parameter "intel_idle.max_cstate=0" to the/boot/grub/menu.lst kernel line, and reboot the server.

Kernel panic occurs with BurninTest 3.1 on a RHEL 6.4 OS or Oracle Linux 6.4 OS that is running the Oracle Unbreakable Enterprise Kernel Release 2 for Linux

BugID: 17047864

Issue:

After running PassMark BurninTest Linux V3.1 on a Red Hat Enterprise Linux (RHEL) 6.4 or Oracle Linux 6.4 operating system, when you reboot or shutdown the system, the kernel might panic at the last stage of shutdown procedure. The kernel panic does not affect the operation of the system. You can ignore the error message and power cycle or power off the server.

Affected hardware and software:

  • Intel 100GB and 400GB SSDs

  • Oracle Linux 6.4 running Oracle Unbreakable Enterprise Kernel Release 2 for Linux

  • Oracle Linux 6.5 and 7.0 running Oracle Unbreakable Enterprise Kernel Release 3 for Linux

  • Red Hat Enterprise Linux 6.4 running Oracle Unbreakable Enterprise Kernel Release 2 for Linux

  • Red Hat Enterprise Linux 6.5 and 7.0 running Oracle Unbreakable Enterprise Kernel Release 3 for Linux

  • BurnInTest3.1

  • Platform software releases 1.0 and later.

Workaround:

Use PassMark BurninTest Linux V2.1 for a system with this configuration.

InfiniBand Host Channel Adapter for PCIExpress card might fail after Oracle Linux 6.5 software is installed

17899109

Issue:

After installing Oracle Linux 6.5, the InfiniBand Host Channel Adapter for PCIExpress card might stop working.

Affected hardware and software:

  • Sun Dual Port QDR InfiniBand Host Channel Adapter for PCIExpress Gen 3 (7104073 and 7104074)

  • Oracle Linux 6.5

  • Platform software releases 1.0 and later.

Workaround:

Do the following:

  1. Install the Remote Direct Memory Access (RDMA) package from yum or from the operating system installation disk.

  2. Add or unmask "MLX4_LOAD=yes" in the /etc/rdma/rdma.conf file.

GRUB fails to boot with Oracle Linux 5 and RHEL 5 and the Flash Accelerator PCIe F80 card

BugID: 18271022

Issue:

GRUB fails to boot when running Oracle Linux 5.x or RHEL 5.x with the Flash Accelerator PCIe F80 Card.

Affected hardware and software:

  • Flash Accelerator PCIe F80 Card: 800GB, eMLC (7107091 and 7107092)

  • Oracle Linux 5.x

  • Red Hat Enterprise Linux 5.x

  • Platform software releases 1.0 and later.

Workaround:

For the workaround for this bug, see Workaround for BugDB 18271022 .

The RDMA communication might fail when using the RHEL 6.5 inbox OFED package

BugID: 17933299

Issue:

If you use the RHEL 6.5 inbox OpenFabrics Enterprise Distribution (OFED) package, the Remote Direct Memory Access (RDMA) in RHEL 6.5 might fail to run.

Affected hardware and software:

  • Sun Dual Port QDR InfiniBand Host Channel Adapter for PCIExpress Gen 3 (7104073 and 7104074)

  • Red Hat Enterprise Linux 6.5

  • Platform software releases 1.1 and later.

Workaround:

None.

Update the Mellanox OFED package. To download the update package, go to http://www.mellanox.com/page/products_dyn?product_family=26, scroll down the page and click the MLNX_OFED 2.X, and download MLNX_OFED_LINUX-2.1-1.0.0-rhel6.5-x86_64.iso.

To install the Mellanox OFED update package, perform the following steps:

  1. Mount the ISO image, and run the ./mlnxofedinstall --without-fw-update command. Note - You must use the --without-fw-update qualifier to prevent the installation from flashing the firmware on the Sun Dual Port QDR InfiniBand Host Channel Adapter for PCIExpress Gen 3 card; otherwise, the installation fails.

  2. Reinstall RDMA package from yum or from the RHEL 6.5 installation ISO. Modify the /etc/rdma/rdma.conf file to change RDS_LOAD=no to RDS_LOAD=yes and save the file.

  3. Start/restart the RDMA service.


Note -  If you want to use "rds-tools" to manage or test RDMA performance, ensure that "rds-tools" is updated to a minimum version of 2.0.7.

Cannot install VMware ESXi 5.x on system configured with Sun Storage 6 Gb SAS PCIe HBA, Internal when non-RAID volumes exist

BugID: 16373075

Issue:

ESXi 5.0/5.1/5.5 and their subsequent updates cannot be installed on a configuration with the Sun Storage 6 Gb SAS PCIe HBA, Internal when non-RAID volumes exist.

Affected hardware and software:

  • Sun Storage 6 Gb SAS PCIe HBA, Internal

  • VMware ESXi 5.0, 5.1, and 5.5 and subsequent updates

  • Platform software releases 1.0 and later.

Workaround:

Use only RAID volumes when installing ESXi on a system configured with the Sun Storage 6 Gb SAS PCIe HBA, Internal.

Oracle Linux 7.0 with UEK Release 3 does not boot if the Sun Storage 10 Gb FCoE HBA card is installed and its option ROM is enabled in UEFI BIOS mode

BugID: 19521738

Issue:

When the Sun Storage 10 Gb FCoE (Fibre Channel over Ethernet) HBA (host bus adapter) card is installed and its option ROM is enabled in Unified Extensible Firmware Interface (UEFI) BIOS mode, the system boots to the GRUB menu and starts to load the system kernel, and then it fails to load the Oracle Linux 7.0 operating system.

Affected hardware and software:

  • Sun Storage 10 Gb FCoE Short Range Optics, QLogic (7101677 and 7101678)

  • Oracle Linux 7.0 with Unbreakable Enterprise Kernel Release 3

  • Platform software releases 1.2 and later.

Workaround:

  1. Boot the system in UEFI BIOS mode.

  2. From the system BIOS, disable the Sun Storage 10 Gb FCoE HBA card option ROM.

  3. Save changes and exit from the system BIOS.

Oracle Linux 7.0 installation reports an unknown error message if the installation destination disk is partitioned more than once during the operating system installation

BugID: 19140366

Issue:

Oracle Linux 7.0 installation reports an unknown error message if the installation destination disk is partitioned more than once during the operating system installation. The error causes the installation to fail and only allows the user the choice to either quit the installation process or report the bug.

The following error message is reported:

An unknown error has occurred This program has encountered an unknown error. You may report the bug below or quit the program.

Affected hardware and software:

  • Oracle Linux 7.0

  • Platform software releases 1.2 and later.

Workaround:

Do not partition the installation destination disk more than once when installing the Oracle Linux 7.0 operating system.

Linux operating systems might not install when using Oracle System Assistant

BugID: 19274609, 19232280, and 19044611

Issue:

Linux operating systems, such as Oracle Linux, SUSE Linux Enterprise Server (SLES), and Red Hat Enterprise Linux (RHEL) might not install when using Oracle System Assistant with the System BIOS set to UEFI mode.

Affected hardware and software:

  • Oracle Linux 6.3, 6.4, 6.5 and 7.0

  • SUSE Linux Enterprise Server 11 SP2 and SP3

  • Red Hat Enterprise Linux 6.4, 6.5, and 7.0

  • Platform software releases 1.2 and later.

Workaround:

Perform the following procedures to reset the System BIOS to the factory default:

  1. Log in to the Oracle ILOM web interface.

  2. Click System Management > BIOS > Settings > Reset to Defaults > Factory, and then click Save.

  3. Reboot the system to load BIOS.

  4. Change the Boot mode to UEFI.

Linux operating systems can now be installed using Oracle System Assistant.

"dmarxd_poisoned_data_from_dp_stat" message appears in Oracle ILOM Fault Manager after installing RHEL 7.0 or Oracle Linux 7.0

BugID: 19293318

Issue:

After you install Red Hat Enterprise Linux (RHEL) 7.0 or Oracle Linux 7.0, and you perform a warm reset on the server, the Oracle ILOM Fault Manager reports the following error:

ereport.io.intel.ilo.dmarxd_poisoned_data_from_dp_stat

Affected hardware and software:

  • Red Hat Enterprise Linux (RHEL) 7.0

  • Oracle Linux 7.0

  • Platform software releases 1.2 and later.

Workaround:

Before you install RHEL 7.0 or Oracle Linux 7.0, disable the Intel I/O Acceleration Technology (IOAT) in the server BIOS Setup Utility. To disable IOAT, perform the following steps:

  1. Launch the BIOS Setup Utility. For instructions, refer to the Oracle X4 Series Servers Administration Guide.

  2. In the the BIOS Setup Utility, navigate to the IO menu.

  3. Select IOAT.

  4. Select Intel I/OAT, and then select Disabled.

  5. Press F10 to save your changes and exit the BIOS Setup Utility.