Oracle Solaris OS Issues

This section describes issues related to the Oracle Solaris OS in this release.

Static/Dynamic Input/Output Not Currently Supported

The SPARC T3-4 server does not support the SDIO feature of Oracle VM Server for SPARC without a fix for 6983964. Please do not use the SDIO feature of Oracle VM Server for SPARC until a fix is available.

See SDIO Policy Violations Might Cause the Primary Domain To Panic During Boot (CR 6983964) for more information.

Oracle Solaris OS Has Changed How It Specifies Logical Device Names

The Oracle Solaris OS now uses SAS 2.0 World Wide ID (WWID) in place of the tn (target ID) field in logical device names. This change will affect how you identify the target disk when downloading the OS over a network. The following points are key to understanding the impact of this change:

• When downloading the OS over a network, you should specify the disk in HDD slot 0 as the download destination. This is the disk that OBP uses as the default boot device.

• Before the change to using WWIDs, this disk would be known to the OS by the logical name c0t0d0s0.

• With the change, the device identifier for the default boot device is now referred to as c0tWWIDd0s0, where WWID is a hexadecimal value. This WWID value does not map in a predictable way to the physical ID of the disk in HDD slot 0.

To reliably specify HDD slot 0 for the OS download operation, you must determine the correspondence between the WWID value for that disk and its physical location. You can do this by running probe-scsi-all and reading the output.

In the probe-scsi-all output, look for the following disk identifiers:

• SASDeviceName – This is the disk WWID that the Oracle Solaris OS recognizes.

• SASAddress – This is the disk WWID that the OBP references.

• PhyNum – This is the physical HDD slot that the disk occupies. It is also expressed as a hexadecimal value.

• VolumeDeviceName – This number is the RAID volume's WWID that the Oracle Solaris OS recognizes.

• VolumeWWID – This is the RAID volume's WWID that OBP references.

A SPARC T3-4 server has two on-board SAS controllers, each controlling four connected drives. The following example probe-scsi-all output is for a SPARC SPARC T3-4 with eight drives.

ok probe-scsi-all
/pci@700/pci@1/pci@0/pci@0/LSI,sas@0
/pci@400/pci@1/pci@0/pci@0/LSI,sas@0
FCode Version 1.00.54, MPT Version 2.00, Firmware Version 5.00.17.00
FCode Version 1.00.54, MPT Version 2.00, Firmware Version 5.00.17.00
Target 9
Target 9
Unit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GB
Unit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GBSASDeviceName 5000cca00abcede0  SASAddress 5000cca00abcede1  PhyNum 0
SASDeviceName 5000cca00abc5cc8  SASAddress 5000cca00abc5cc9  PhyNum 0Target aTarget aUnit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GB
Unit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GB
SASDeviceName 5000cca00abaf620  SASAddress 5000cca00abaf621  PhyNum 1
SASDeviceName 5000cca00abc51a8  SASAddress 5000cca00abc51a9  PhyNum 1Target b
Target bUnit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GBUnit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GBSASDeviceName 5000cca00abcec4c  SASAddress 5000cca00abcec4d  PhyNum 2
SASDeviceName 5000cca00abce89c  SASAddress 5000cca00abce89d  PhyNum 2
Target c
Target cUnit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GB
Unit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GBSASDeviceName 5000cca00abc5218  SASAddress 5000cca00abc5219  PhyNum 3SASDeviceName 5000cca00abc5354  SASAddress 5000cca00abc5355  PhyNum 3
/pci@400/pci@1/pci@0/pci@8/pci@0/usb@0,2/hub@2/hub@3/storage@2
{0} ok
Unit 0   Removable Read Only device    AMI     Virtual CDROM   1.00
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

The following probe-scsi-all example output shows a RAID configuration. The RAID volume's VolumeDeviceName is 3c2f959213c8a292.

ok probe-scsi-all
/pci@700/pci@1/pci@0/pci@0/LSI,sas@0
FCode Version 1.00.54, MPT Version 2.00, Firmware Version 5.00.17.00
 
Target 9
Unit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GB
SASDeviceName 5000cca00abc5cc8  SASAddress 5000cca00abc5cc9  PhyNum 0
Target a
Unit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GB
SASDeviceName 5000cca00abaf620  SASAddress 5000cca00abaf621  PhyNum 1
Target 37e Volume 0
Unit 0   Disk   LSI      Logical Volume   3000    1167966208 Blocks, 597 GB
VolumeDeviceName 3c2f959213c8a292  VolumeWWID 0c2f959213c8a292
 
/pci@400/pci@1/pci@0/pci@8/pci@0/usb@0,2/hub@2/hub@3/storage@2
Unit 0   Removable Read Only device    AMI     Virtual CDROM   1.00
 
/pci@400/pci@1/pci@0/pci@0/LSI,sas@0
 
FCode Version 1.00.54, MPT Version 2.00, Firmware Version 5.00.17.00
 
Target 9
Unit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GB
SASDeviceName 5000cca00abcede0  SASAddress 5000cca00abcede1  PhyNum 0
Target a
Unit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GB
SASDeviceName 5000cca00abc51a8  SASAddress 5000cca00abc51a9  PhyNum 1
Target b
Unit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GB
SASDeviceName 5000cca00abce89c  SASAddress 5000cca00abce89d  PhyNum 2
Target c
Unit 0   Disk   HITACHI  H103030SCSUN300G A2A8    585937500 Blocks, 300 GB
SASDeviceName 5000cca00abc5354  SASAddress 5000cca00abc5355  PhyNum 3
 
{0} ok
 

Oracle Solaris Jumpstart Examples

The following Oracle Solaris Jumpstart profile example shows how to use the WWID syntax when installing the OS on a specific disk drive. The SASDeviceName is taken from the previous six-drive configuration listing.

#
install_type flash_install
boot_device c0t5000C5001CB4A637d0s0 preserve
 
archive_location nfs 129.148.94.249:/export/install/media/solaris/builds/s10u9/flar/latest.flar
 
# Disk layouts
#
partitioning explicit
filesys rootdisk.s0         free /
filesys rootdisk.s1         8192 swap

The following Oracle Solaris Jumpstart profile example shows how to use the WWID syntax when installing the OS on a RAID volume. The VolumeDeviceName is taken from the previous RAID probe-scsi-all example.

#
install_type flash_install
boot_device c0t3CE534E42C02A3C0d0s0 preserve
 
archive_location nfs 129.148.94.249:/export/install/media/solaris/builds/s10u9/flar/latest.flar
 
# Disk layouts
#
partitioning explicit
filesys rootdisk.s0         free /
filesys rootdisk.s1         8192 swap

Interactive Installation Example

In an interactive install, you will be asked to specify one or more disks as the targets for the OS installation. The purpose of this step is to ensure that enough disk capacity is being provided for the installation. For this step, specify the disk with the WWID value corresponding to the drive on which you want to install the software.

These WWID values are illustrated in the following interactive example, which is based on the same six-disk environment used in the previous examples. The drive selected as the install target is located in HDD slot 0—the default OBP location.

_ Select Disks_________________________________________________________________
 
On this screen you must select the disks for installing Solaris software. Start 
by looking at the Suggested Minimum field; this value is the approximate space 
needed to install the software you've selected. Keep selecting disks until the 
Total Selected value exceeds the Suggested Minimum value.
NOTE: ** denotes current boot disk
 
Disk Device                                             Available Space
=============================================================================
  [ ]    c0t5000CCA00ABAF620d0                            286090 MB
 [X] ** c0t5000CCA00ABC51A8d0                            286090 MB
______________________________________________________________________________
[ ]    c0t5000CCA00ABC5218d0                            286090 MB[ ]    c0t5000CCA00ABC5354d0                            286090 MB
  Esc-2_Continue    F3_Go Back    F4_Edit    F5_Exit    F6_Help[ ]    c0t5000CCA00ABC5CC8d0                            286090 MB
[ ]    c0t5000CCA00ABCE89Cd0                            286090 MB
[ ]    c0t5000CCA00ABCEC4Cd0                            286090 MB
[ ]    c0t5000CCA00ABCEDE0d0                            286090 MB
 
Total Selected: 286090 MB
Suggested Minimum:   5009 MB
 
 
 
 
 
 

The cfgadm -al Command Takes a Long Time to Print Output (CR 6937169)

The cfgadm(1M) command for configuring or unconfiguring hot-plug devices takes a long time to complete. For example, the cfgadm -al command could take more than five minutes before it lists the attachment points for all the hot-plug devices.

Workaround:

Use the hotplug(1M) command to manage PCIe hotplug devices.

• Use the hotplug list -l command to list the status of all hot-plug PCIe slots. For example:

  # hotplug list -l | grep PCI-EM
  /pci@400/pci@1/pci@0/pci@4 [PCI-EM2] (EMPTY)
  /pci@400/pci@2/pci@0/pci@1 [PCI-EM0] (EMPTY)
  /pci@400/pci@2/pci@0/pci@2 [PCI-EM1] (EMPTY)/pci@400/pci@2/pci@0/pci@3 [PCI-EM3] (ENABLED)
  /pci@500/pci@1/pci@0/pci@1 [PCI-EM8] (EMPTY)
  /pci@500/pci@1/pci@0/pci@2 [PCI-EM10] (ENABLED)
  /pci@500/pci@2/pci@0/pci@2 [PCI-EM9] (ENABLED)
  /pci@500/pci@2/pci@0/pci@3 [PCI-EM11] (EMPTY)
  /pci@600/pci@1/pci@0/pci@4 [PCI-EM4] (EMPTY)
  /pci@600/pci@1/pci@0/pci@5 [PCI-EM6] (ENABLED)
  /pci@600/pci@2/pci@0/pci@0 [PCI-EM7] (EMPTY)
  /pci@600/pci@2/pci@0/pci@5 [PCI-EM5] (EMPTY)
  /pci@700/pci@1/pci@0/pci@4 [PCI-EM14] (EMPTY)
  /pci@700/pci@2/pci@0/pci@3 [PCI-EM12] (ENABLED)
  /pci@700/pci@2/pci@0/pci@4 [PCI-EM13] (EMPTY)
  /pci@700/pci@2/pci@0/pci@5 [PCI-EM15] (EMPTY)
   
   
   

• Use the hotplug disable command to disable a PCIe card.

  For example, to disable the EM card in PCI-EM3 and confirm that it is no longer enabled:

  # hotplug disable /pci@400/pci@2/pci@0/pci@3 PCI-EM3
  # hotplug list -l | grep PCI-EM3/pci@400/pci@2/pci@0/pci@3 [PCI-EM3] (POWERED)

• Use the hotplug poweroff command to power off a PCIe card.

  For example, to power off the EM card in PCI-EM3:

  # hotplug poweroff /pci@400/pci@2/pci@0/pci@3 PCI-EM3
  # hotplug list -l | grep PCI-EM3/pci@400/pci@2/pci@0/pci@3 [PCI-EM3] (PRESENT)

  You may now physically remove the EM card.

• Use the hotplug list command to verify that a card is removed.

  For example:

  # hotplug list -l | grep PCI-EM
  ...
  /pci@400/pci@2/pci@0/pci@3 [PCI-EM3] (EMPTY)
  ...
   
   

• Use the hotplug poweron command to power on a PCIe card.

  For example, to power on the EM card in PCI-EM3 and confirm that it has moved to the POWERED state:

  # hotplug poweron /pci@400/pci@2/pci@0/pci@3 PCI-EM3
  # hotplug list -l | grep PCI-EM3/pci@400/pci@2/pci@0/pci@3 [PCI-EM3] (POWERED) 

• Use the hotplug enable command to enable a PCIe card.

  For example, to enable the EM card in PCI-EM3 and confirm that it has moved to the ENABLED state:

  # hotplug enable /pci@400/pci@2/pci@0/pci@3 PCI-EM3
  # hotplug list -l | grep PCI-EM3/pci@400/pci@2/pci@0/pci@3 [PCI-EM3] (ENABLED)

False nxge Warning Messages (CR 6938085)

During the normal operation of your server, you might see warning messages like the following in the system console:

date time machinename nxge: [ID 752849 kern.warning] WARNING: nxge0 : nxge_hio_init: hypervisor services version 2.0

These messages are not true warning messages. These Gigabit Ethernet driver (nxge) messages display the version number of the hypervisor since the driver can operate on multiple hypervisor versions. These messages should labeled as a INFO or NOTICE messages instead of WARNING messages.

Workaround:

You can safely ignore these messages.

Spurious Interrupt Message in System Console (CR 6963563)

During the normal operation of the server, and when running the Oracle VTS system exerciser, you might see the following message in the system console:

date time hostname px: [ID 781074 kern.warning] WARNING: px0: spurious interrupt from ino 0x4
date time hostname px: [ID 548919 kern.info] ehci-0#0
date time hostname px: [ID 100033 kern.info]

Workaround:

You can safely ignore this message.

The prtpicl Command Does Not Display Drive Information (CR 6963594)

On previous systems, the prtpicl -v command displayed the state, location, and device paths for system drives under a disk_discovery heading. On SPARC T3 systems, the prtpicl command no longer displays this drive information.

Workaround:

Use the OpenBoot probe-scsi-all command. For example output, see Oracle Solaris OS Has Changed How It Specifies Logical Device Names.

Missing Interrupt Causes USB Hub Hotplug Thread to Hang, Resulting In Process Hangs (CR 6968801)

When running the Oracle VTS software on SPARC T3 series servers, it is possible (although rare) for a Oracle VTS test to hang. If this test process hangs, the hung process might cause other processes and commands to hang, including the fault management configuration tool (fmadm) and the print system configuration command (prtconf). These hung processes cannot be killed.

Workaround:

Reboot the system. If the problem repeats, contact your service representative for assistance. Avoid running the Oracle VTS software in production environments.

Long Local Console Delays During Login or Logout of Oracle Solaris (CR 6971884)

You might experience long latency times (up to three minutes) when logging in or logging out of the Solaris OS using a local console or local keyboard, mouse and monitor.

Workaround:

Use a network connection to access the system, and comment out the lines containing USB console device names (/dev/usb/*) in the /etc/logindevperm configuration file:

# /dev/console    0600 /dev/usb/hid[0-9]+      # hid devices should have the same permission with conskbd and consms
# /dev/console    0600 /dev/usb/[0-9a-f]+[.][0-9a-f]+/[0-9]+/* driver=scsa2usb,usb_mid,usbprn,ugen     #libusb/ugen devices
...

Spurious Error Message During Initial Oracle Solaris OS Installation (CR 6971896)

The miniroot is a bootable root file system that includes the minimum Oracle Solaris OS software required to boot the server and configure the OS. The miniroot runs only during the installation process.When the server boots the miniroot for the initial configuration, you might see the following messages in the system console:

Fatal server error:
InitOutput: Error loading module for /dev/fb
 
giving up.
/usr/openwin/bin/xinit:  Network is unreachable (errno 128):  unable to connect to X server
/usr/openwin/bin/xinit:  No such process (errno 3):  Server error.

The messages indicate that the Xsun server in the Oracle Solaris OS miniroot cannot find a supported driver for the AST graphics device in the service processor. These messages are fully expected, as the miniroot contains only the Xsun environment, and the AST framebuffer (astfb) is supported only in the Xorg environment. The Xorg environment is included in the installed system, so the graphics device may be used when running the installed Oracle Solaris OS.

Workaround:

You can safely ignore this message.

SDIO Policy Violations Might Cause the Primary Domain To Panic During Boot (CR 6983964)

If you have a system using Oracle VM Server for SPARC and you have one or more guests configured to use the Static/Direct Input/Output (SDIO) feature, it is possible for the primary domain to panic when the system has just been powered on and the primary domain is booting.

The panic message appears similar to the following:

panic[cpu6]/thread=2a101283ca0: Fatal error has occured in: PCIe fabric.(0x1)(0x43)
 
000002a101283700 px:px_err_panic+1ac (702cec00, 7bf57000, 43, 2a1012837b0, 1, 0)
%l0-3: 0000009980001602 00000000702cf000 0000000000000000 0000000000000001
%l4-7: 0000000000000000 00000000018af000 0000000000000001 0000000000000000
000002a101283810 px:px_err_fabric_intr+1c0 (6002c594480, 1, 702cf000, 1, 43, 200)
%l0-3: 0000000000000200 0000000000000001 00000000702cf2c0 00000000702cf000
%l4-7: 00000000702cf2b8 00000000702cf000 0000000000000001 000006002c4d4790
000002a101283980 px:px_msiq_intr+1e8 (6002c447bf0, 30002df1908, 7bf495d0, 0, 1, 6002c6f01f0)
%l0-3: 000006002c639220 000006002b561e40 0000030002df1908 0000000000000000
%l4-7: 0000000000000000 0000000003a60000 000002a101283a80 0000000000000030

This problem can occur while the primary domain is booting; it does not happen to already running systems.

Workaround:

There is no workaround currently available for this issue. Please contact your authorized service provider for a fix.

On-Board Ethernet Devices Fail to Connect After a Faulty CPU Reconfigures Back to the Host (CR 6984323)

When rebooting the server after a failed or disabled CPU reconfigures back to the host, the onboard Gigabit Ethernet connections will not connect to network. The following example messages will display on the system console:

igb0: DL_ATTACH_REQ failed: DL_SYSERR (errno 22)
igb0: DL_BIND_REQ failed: DL_OUTSTATE
igb0: DL_PHYS_ADDR_REQ failed: DL_OUTSTATE
igb0: DL_UNBIND_REQ failed: DL_OUTSTATE
Failed to plumb IPv4 interface(s): igb0

Workaround:

Reboot the server two additional times. If the problem persists, contact your service representative for assistance.

hostconfig Command Does Not Update CPU Serial Number in the Physical Resource Inventory Machine Descriptor (PRI MD) (CR 6989166)

When a CPU thread faults, the record of that fault is retained when the system reboots, even if the cause of the fault has been corrected. The presence of this record will cause the CPU thread to be off-lined when the system reboots.

Recovery:

Enable the faulted thread(s) explicitly.

Oracle Enterprise Manager Process Hangs and Becomes Unkillable (CR 6994300)

The Oracle Enterprise Manager Java process can hang and become unkillable on the server. When the Enterprise Manager process hangs, it continues to listen on its Web UI port, which makes the process unkillable. This problem has been seen on servers running both the Java SE 5.0 version that is bundled with Oracle Database software and with the most recent downloadable Java SE 6 Update 22 version.

Workaround:

Reboot the system. If the problem repeats, contact your service representative for assistance.

Gigabit Ethernet (nxge) Driver Not Loading on Systems With Oracle Solaris 10 10/09 OS and Solaris 10 9/10 Patch Bundle (CR 6995458)

If you installed the Oracle Solaris 10 10/09 OS and the Solaris 10 9/10 Patch Bundle on the server, the 10/1 Gigabit Ethernet (nxge) driver will not attach to 10 GbE devices.

Workaround:

Add the following line to the /etc/driver_aliases file:

nxge "SUNW,niusl-kt"

Then reboot the server and configure the Gigabit Ethernet devices normally.

Diagnosis Engine (eft) is Disabled on Memory Unrecoverable Errors Reportedly Due To Exceeding Module Memory Limit (CR 7000649)

Under certain circumstances, a memory error can cause an internal error in the FMA diagnosis engine that prevents proper diagnosis of the memory error.This failure might be present if fmadm faulty command output includes the fault FMD-8000-2K in the list of faulty components.

For example:

# fmadm faulty
...
--------------- ------------------------------------  -----------------------
TIME            EVENT-ID                              MSG-ID         SEVERITY--------------- ------------------------------------  -----------------------
Nov 16 12:02:01 865e378b-27a5-ebf7-c550-e34179d57241  FMD-8000-2K    Minor
 
Host        : xxxxxxx
Platform    : ORCL,SPARC-T3-4   Chassis_id  :
Product_sn  :
 
Fault class : defect.sunos.fmd.module
Affects     : fmd:///module/eft faulted and taken out of service
FRU         : None
faulty
 
Description : A Solaris Fault Manager component has experienced an error that
required the module to be disabled.  Refer to
http://sun.com/msg/FMD-8000-2K for more information.
 
Response    : The module has been disabled.  Events destined for the module
will be saved for manual diagnosis.
 
Impact      : Automated diagnosis and response for subsequent events associated
with this module will not occur.
 
Action      : Use fmdump -v -u <EVENT-ID> to locate the module.  Use fmadm
reset <module> to reset the module.
 
 
 

To confirm that you have encountered CR 70000649, use the fmdump -eV command and look for an ereport.fm.fmd.module ereport that contains the words “eft's allocation of XX bytes exceeds module memory limit” where XX is any number.

For example:

# fmdump -eV...
Nov 16 2010 12:02:01.052061745 ereport.fm.fmd.module
nvlist version: 0
version = 0x0
class = ereport.fm.fmd.module       detector = (embedded nvlist)
nvlist version: 0
version = 0x0
scheme = fmd
authority = (embedded nvlist)
nvlist version: 0
version = 0x0
product-id = ORCL,SPARC-T3-4
server-id = xxxxxxxx
(end authority)
 
mod-name = eft
mod-version = 1.16
(end detector)
 
ena = 0x3ddfe7a2c3f07401
msg = eft's allocation of 20 bytes exceeds module memory limit (10485756)
__ttl = 0x1
__tod = 0x4ce2e339 0x31a6631

Workaround:

If the fmdump -eV output indicates that you have encountered CR 7000649, contact your authorized service provider to see if a fix is available.

Schedule a reboot of the server with diagnostics enabled to allow POST to attempt to find the suspect DIMM.

After halting all of the Solaris sessions on your system, complete the following:

1. Power off the system:

   -> stop /SYS

2. Set keyswitch_state to Diag:

   -> show /SYS keyswitch_state 
   /SYS
   Properties:
   keyswitch_state = Normal
   -> set /SYS keyswitch_state=diag
   Set 'keyswitch_state' to 'diag'
   -> show /SYS keyswitch_state 
   Properties:/SYSkeyswitch_state = Diag
    
    
    
    
    
    
    

   ---

   Note - This option overrides previously set values of diagnostic properties.

   ---

3. Power on the server:

   -> start /SYS
   Are you sure you want to start /SYS (y/n)? yStarting /SYS

4. After the power-on sequence is complete, enter the ILOM Fault Management shell and use the ILOM show faulty command to look for any faulty DIMMs.

   1. Start the fault management shell:

      -> start /SP/faultmgmt/shell 
      Are you sure you want to start /SP/faultmgmt/shell (y/n)? y

   2. Display the list of faulty components:

      faultmgmtsp> fmadm faulty

5. Replace any faulty DIMMs that are reported. If no faulty DIMMs are reported, contact your authorized service provider for further assistance.