SPARC T3-4
 
SPARC T4-2
 
SPARC T3-2
 
SPARC T4-1
 
SPARC T5-2
 
SPARC T5-4
 
SPARC T3-1
 
SPARC T5-8
 
SPARC T4-4
 

PLA-Support>Sun Systems>SPARC>CMT>SN-SPARC: T5
 

In this Document

Purpose

Troubleshooting Steps

References

Applies to:

Purpose

 This document provides guidance about what data to collect and how to analyze disk failures when an explorer cannot be provided.

Troubleshooting Steps

Please gather and analyze the output of the commands listed below.

Be informed that an Explorer gathers all this data but it's important to know the commands that reveal disk information, what to check for, and the interpretation of each output.

I. If you are not able to reach Solaris, provide the output of the following commands from OBP (ok> prompt):

ok> printenv                         Determine boot device (internal or external!!!). If external and an Oracle storage array, then open an SR into the Storage group.

ok> setenv auto-boot? false   Change the PROM auto-boot? parameter to false.

ok> reset-all                         Reset the system before executing probe-scsi-all command (to prevent the system from hanging).

ok> probe-scsi-all                 The probe-scsi-all command transmits an inquiry command to all SCSI devices connected to the system SCSI host adapters, including any host adapters installed in PCI slots.
                                           The first identifier listed in the display is the SCSI host adapter address in the system device tree followed by the SCSI device identification data.

ok> devalias                         Lists the device aliases and the associated paths of devices that might be connected to the system.

ok> show-disks                     Lists disks and has options to add to dev alias.

II. If you are able to reach Solaris, provide the output of the following commands:

1. #prtdiag -v

Please check the LED status, environmental status.

2. #cfgadm -alv

This command displays information about a system's SCSI controllers and their attached devices.

3. #format

The command displays a list of recognized disks under AVAILABLE DISK SELECTIONS.

Sometimes when an disk has failed it's no longer available in the format output.

4. #uptime

This command shows how long the system has been up.

5. #iostat -En

You can search for hardware errors but please correlate the output you see here with the uptime.

Iostat counters are reset at system reboot.

If a failed disk has been replaced previously and the system was not rebooted, the errors will still be reported in Iostat for the new disk.

If you can't reboot the server,  check Document 1012731.1 Reset the iostat -E hard/soft/tran errors counters without rebooting.

6. #metastat -t 

Use this command for disks configured under Solaris Volume Manager.

You need to check if any submirros that are in status "Need maintenance".

Also check Document 1002410.1 How to recover mirrored metadevice with both submirrors in a "Needs Maintenance" state caused by a bad disk block.

7. #zpool status

Use this command for disks configured under ZFS.

You need to check if any pools are in state degraded.

8. Get the output from dmesg and messages 

#dmesg

# more /var/adm/messages

Look for error blocks that are reported and for messages mentioning that the disk has been taken offline.

If after proceeding with steps 1-8 you only notice error blocks on a certain disk, run the command format > analyze > read to repair the defective blocks on a disk.
format> analyze > read - command reads each sector on the current disk. Repairs defective blocks as a default (it doesn't harm SunOS).

9. If Internal LSI Hardware RAID Volumes are configured using one the following 3 utilities: Sas2ircu, Raidconfig and  Fcode check Document 1483874.1 - How to Identify a Failed Disk when Internal LSI Hardware RAID Volumes are configured (T3-x and T4-x).

If the internal disks are configured using a RAID HBA highlighted by the path to disks with "LSI,mrsas", check Document 1471121.1 Some or all internal disks in SPARC T3 / T4 Servers may not show up in "# format" or "ok> probe-scsi-all".

10.Displaying Disk Physical Locations:

a. Please get the output of the command #prtconf -v.

And then identify the value of the 'phy-num'.

Further details can be found in Document 1365089.1 - How to locate a failed internal disk drive on a T3/T4/T5/T7/T8 series system.

b. For servers that use Solaris 11, you can use the #diskinfo command.

Details can be found at https://docs.oracle.com/cd/E53394_01/html/E54782/disksprep-3.html.

c. If the disk is under MPXIO /scsi_vhci control check Document 1542744.1 How to Identify the Target ID for an Internal System Disk under MPXIO/scsi_vhci Control.

11.  raidconfig list all -v  - raidconfig is a command available when Hardware Management Pack (HMP) is installed.  HMP is automatically installed on Solaris 11.2 and later.

 Sample output from a T3-4:

12. If after taking steps 1-9 you have gathered sufficient information probing the fact that a disk has failed, please contact Oracle support.

References

This solution has no attachment