Sun Fire T2000 Server
 
Sun SPARC Enterprise T1000 Server
 
Sun SPARC Enterprise T5140 Server
 
Sun SPARC Enterprise T2000 Server
 
Sun Fire T1000 Server
 

PLA-Support>Sun Systems>SPARC>CMT>SN-SPARC: Tx000
 

raidctl output may show inconsistent configuration information if there are two raid volumes configured, and Solaris version is Solaris10_U4 or older and raidctl patches are not applied

In this Document

Symptoms

Cause

Solution

References

Created from <SR 3-7109455966>

Applies to:

Symptoms

• System is running Solaris10_U4 or earlier;

# cat /etc/release
                          Solaris 10 8/07 s10s_u4wos_12b SPARC
              Copyright 2007 Sun Microsystems, Inc.  All Rights Reserved.
                           Use is subject to license terms.
                               Assembled 16 August 2007

•  And Solaris is missing kernel patches associated with raidctl: 127127-02 and/or 127127-07;

# showrev -p | egrep 127127
#

Notice that the output is empty, meaning patches are not installed.

• And there are two raid volumes created with inlernal LSI controller.

# raidctl -l
Controller: 0
       Volume:c0t0d0
       Volume:c0t2d0
       Disk: 0.2.0
       Disk: 0.2.0
       Disk: 0.2.0
       Disk: 0.2.0
       Disk: 0.2.0
       Disk: 0.2.0
       Disk: 0.2.0
       Disk: 0.2.0
       Disk: 0.2.0
       Disk: 0.2.0
       Disk: 0.2.0
#

Notice there are two volumes created;

and " Disk: 0.2.0 " is displayed many times; there should be 4 different devices, not only one !!

• And raidctl output shows duplicate information, meaning that both raid volumes show the same output.

# raidctl -l c0t0d0
Volume                  Size    Stripe  Status   Cache  RAID
       Sub                     Size                    Level
               Disk
----------------------------------------------------------------
c0t0d0                  68.3G   N/A     DEGRADED N/A    RAID1
               0.3.0   68.3G           FAILED
               0.2.0   68.3G           GOOD
#
#
# raidctl -l c0t2d0
Volume                  Size    Stripe  Status   Cache  RAID
       Sub                     Size                    Level
               Disk
----------------------------------------------------------------
c0t2d0                  136.6G  N/A     DEGRADED N/A    RAID1
               0.3.0   136.6G          FAILED
               0.2.0   136.6G          GOOD

Notice from above that both c0t0d0 and c0t2d0 are shown as they were configured from the same pair of disks (0.3.0 and 0.2.0). This is not a supported configuration; output is inconsistent !

Cause

System may have hit known bugs; e.g:
  BUG 15441261 - SUNBT6622823-SOLARIS_10U5 raidctl gets incorrect 2nd volume disk information
  BUG 15415824 - SUNBT6510056-SOLARIS_10U5 raidctl should display correct disk infomation of arra

Solution

Workaround (requires a system outage):

Sysadmins may rely on OBP commands in order to view the correct raid configuration and raid volume components statuses:

1. Backup your system (all file-systems that rely on internal disks). Not mandatory, but recommended as a precaution.
2. Connect to console and capture full output from terminal to a session log file.
3. Bring the system down to OBP:
   

   # shutdown -g0 -y -i0

    
4. Set the auto-boot parameter to false, set the fcode-debug parameter to true and perform a reset-all:
   

   ok setenv fcode-debug? true
   ok setenv auto-boot? false
   ok reset-all

    
5. Determine the current device path for LSI controller's virtual volumes ("show-disks" command is used only to confirm the controller path. From "show-disks" menu, simply choose "q" in order to quit):
   

   ok show-disks

    
6. From previous output, match the system model and raid controller model in order to run OBP command "select" with the appropriate controller device path.
   
   List of device paths:
     T2000 with part number 501-6843 using the PCI-X SAS card: /pci@7c0/pci@0/pci@1/pci@0,2/LSILogic,sas@2
     T2000 with part number 501-7501 using the onboard 1064 controller: /pci@780/pci@0/pci@9/scsi@0
     T1000: /pci@7c0/pci@0/pci@8/scsi@2
    
   
   Follows an example for Sun Fire T2000 using the on-board LSI1064E controller. Select the following path:
   

   ok select /pci@780/pci@0/pci@9/scsi@0

    
7. After selecting the controller path, use 'show-volumes' command to display current volumes and their component statuses:
   

   ok show-volumes
   Volume 0 Target 2  Type IM (Integrated Mirroring)
    Optimal  Enabled
    2 Members                                         286607360 Blocks, 146 GB
    Disk 1
      Primary  Online
      Target 4        FUJITSU MBB2147RCSUN146G0505
    Disk 0
      Secondary  Online
      Target 3        SEAGATE ST914602SSUN146G0703
   Volume 1 Target 0  Type IM (Integrated Mirroring)
    Degraded  Enabled
    2 Members                                         143243264 Blocks, 73 GB
    Disk 3
      Primary  Online
      Target 5        SEAGATE ST973402SSUN72G 0400
    Disk 2
      Secondary  Missing  Out Of Sync
      Target 1
   
   

    
8. From previous command, you can see current and reliable information regarding volumes configuration and components statuses. Notice each volume has its unique pair of devices.
9. Unselect the raid controller
   

   ok unselect-dev

    
10. With consistent status information from the output of these commands, sysadmin can move forward in order to troubleshoot the raid volumes. In the example above, Disk 2 is missing due to a hard disk failure.
11. Restore the NVRAM settings for normal operation and reboot your system:
    

    ok setenv auto-boot? true
    ok setenv fcode-debug? false
    ok reset-all
    ok boot
    

     

Solution:

Either

•        Download and apply the most up-to-date Solaris 10 patch bundle
  or
•        Upgrade your system to SOLARIS_10U5 or later, which include the fix for these bugs

Both approaches fix the raidctl misbehavior described in this document.

After fix is implemented, you can confirm raidctl output is correct:

# raidctl -l c0t0d0
Volume                  Size    Stripe  Status   Cache  RAID
       Sub                     Size                    Level
               Disk
----------------------------------------------------------------
c0t0d0                  68.3G   N/A     DEGRADED OFF    RAID1
               0.0.0   68.3G           GOOD
               0.1.0   68.3G           FAILED
#
#
# raidctl -l c0t2d0
Volume                  Size    Stripe  Status   Cache  RAID
       Sub                     Size                    Level
               Disk
----------------------------------------------------------------
c0t2d0                  136.6G  N/A     OPTIMAL  OFF    RAID1
               0.2.0   136.6G          GOOD
               0.3.0   136.6G          GOOD

References

This solution has no attachment