Sun Microsystems, Inc.  Oracle System Handbook - ISO 7.0 May 2018 Internal/Partner Edition
   Home | Current Systems | Former STK Products | EOL Systems | Components | General Info | Search | Feedback

Asset ID: 1-79-2332709.1
Update Date:2018-02-12
Keywords:

Solution Type  Predictive Self-Healing Sure

Solution  2332709.1 :   Sparc Servers : Replacing a Smaller Older Disk with a Newer Larger Capacity Disk  


Related Items
  • Sun Enterprise 4500 Server
  •  
  • Sun SPARC Enterprise M9000-64 Server
  •  
  • Sun Fire E25K Server
  •  
  • Sun Fire V440 Server
  •  
  • Sun SPARC Enterprise M5000 Server
  •  
  • Sun Fire V480 Server
  •  
  • Sun Fire 6800 Server
  •  
  • Sun SPARC Enterprise M3000 Server
  •  
  • Sun Fire V445 Server
  •  
  • Sun Fire 12K Server
  •  
  • Sun SPARC Enterprise M9000-32 Server
  •  
  • Sun Fire V890 Server
  •  
  • Sun Fire 4800 Server
  •  
  • Sun Fire V880 Server
  •  
  • Sun Fire 15K Server
  •  
  • Sun SPARC Enterprise M4000 Server
  •  
  • Sun SPARC Enterprise M8000 Server
  •  
  • Sun Fire V490 Server
  •  
Related Categories
  • PLA-Support>Sun Systems>SPARC>Enterprise>SN-SPARC: Mx000
  •  
  • Tools>Primary Use>Configuration
  •  




In this Document
Purpose
Scope
Details
References


Applies to:

Sun Fire 12K Server
Sun Fire 15K Server
Sun Fire E25K Server
Sun Fire 4800 Server
Sun Fire 6800 Server
SPARC

Purpose

Many older, smaller, and  low capacity drives have become discontinued by their manufacturer.  When parts supply or mean time between failure becomes unacceptable to Oracle, we may send a newer supported drive as a replacement.  It requires some pre-setup to replace the failed disk but will increase the stability of SUN Sparc Servers for our customers.

Scope

This Document will describe how to successfully mirror a new replacement disk which has a larger capacity then original failed mirror disk.

Details

Assumptions for this Document

You have run the appropriate utilities and have determined that C1t1d0 has failed and that c1t0d0 is the remaining working member of the servers boot pool.

Determining the failing disk slot position

  • Consult your SR / Task to determine which disk number is requiring replacement. 
  • To find the disk number to slot position of the failing drive, consult the Oracle Systems Handbook

Determine if the replacement system is part of a working mirror

  • If your records indicate that the internal disk you are replacing is part of an existing mirror you must follow this doc to completion.
  • If you are replacing an un-mirrored data disk, you could just proceed with the same procedures you would normally do utilizing backup and restore and it will work just fine. 
  • If there is a question please ask your TSE engineer to help you with the right answer.

Determining the size and file system label of working mirror disk

It is possible that you may have one of three file systems on the disk.

1:  M3000/M8000/M9000 and v440/V445 disks may have a LSI hardware raid file system on them.  If you are not sure, reference the following doc and if hardware raid is present follow the directions in that DOC. 

1009043.1 - How to Replace a Failed Disk Mirrored with the raidctl Command

2: All systems without hardware raid, (SVM,ZFS)  will have either SMI or EFI labels on them and mirrored with software. 

It is important to match the label when re-mirroring the disk.

 

When you have replaced the disk it is critical to match the label.  You will utilize this label info when replacing the failed disk.


Labeling The new disk to match the existing Disk

Before adding the replacement disk to the mirror we need to insure that the labels match.  Therefore it is important to run the format command with the -e qualifier on the new disk to properly type and  label it to make it ready to use.  Even if the disk is correctly labeled out of the box it is important to complete this step because it puts a "default" label with a "default" partition table on the replacement disk.  This will ensure that the next steps are successful.

EXAMPLE:

# format -e

 Searching for disks...done

AVAILABLE DISK SELECTIONS:

AVAILABLE DISK SELECTIONS:
0. c0t0d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
/pci@1e,600000/SUNW,qlc@4/fp@0,0/ssd@w21000004cf98a81b,0  <<< Disk Size is <SUN72GB so it is a 72 Giga Byte Disk  This is the "working" side of the mirror
1. c0t1d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
/pci@1e,600000/SUNW,qlc@4/fp@0,0/ssd@w21000004cf9b5e30,0 
2. c0t2d0 <SEAGATE-ST373307FSUN72G-0407-68.37GB>
/pci@1e,600000/SUNW,qlc@4/fp@0,0/ssd@w21000004cf58dd52,0
3. c0t3d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
/pci@1e,600000/SUNW,qlc@4/fp@0,0/ssd@w2100000c50f1cdee,0
4. c0t4d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
/pci@1e,600000/SUNW,qlc@4/fp@0,0/ssd@w2100000c50f1a896,0
5. c0t5d0 <SUN72G cyl 14087 alt 2 hd 24 sec 424>
/pci@1e,600000/SUNW,qlc@4/fp@0,0/ssd@w2100000c50247cdc,0
Specify disk (enter its number): 1

C0t1d0    << Chose this disk because according to SR, it was the failing mirror diskAlso note the type of disk is <SUN72G>  A 72 GB disk

[disk formatted]

The format command returned disk to match the size of the working mirror in this case the working mirror member is 72 GB  So we want to make it be seen as a "72GB" disk drive

Format> type

AVAILABLE DRIVE TYPES:
0. Auto configure
...... snip .....

19. SUN72G     <<<< 19 is the correct type
20. FUJITSU-MAW3073NC-0103
21. other

Specify disk type (enter its number)[20]: 19

Next find out how the working mirror is labeled.  It could be either "SMI or EFI".

Oracle Solaris utilizes 2 disk labels.

SMI this is the legacy label which Solaris has always used on SPARC.
EFI This is a newer label system which allows a partition for grub on X86 systems.

It really only matters that the label utilized  label is utilized on the working mirror is duplicated on the replacement disk.  It is important to verify the label on the working side of the mirror and match it to the new replacement disk. In the example below c0t1d0s0 is the remaining working volume in the mirror so it is queried with the prtvtoc command.  If the label has a partition 8 it is a EFI label

Example:

"SMI Label"
~# prtvtoc /dev/rdsk/c0t1d0s0
* /dev/rdsk/c0t1d0s0 partition map
*
* Dimensions:
* 512 bytes/sector
* 848 sectors/track
* 24 tracks/cylinder
* 20352 sectors/cylinder
* 14089 cylinders
* 14087 accessible cylinders
*
* Flags:
* 1: unmountable
* 10: read-only
*
*                            First      Sector         Last
* Partition Tag Flags Sector   Count         Sector    Mount Directory
       0        2    00     0        286698624
       2        5    01     0        286698624 286698623

-----------------------------------------------------------------------------------------

EFI Label
# prtvtoc /dev/rdsk/c0t1d0s0
* /dev/rdsk/c0t1d0s0 partition map
*
* Dimensions:
* 512 bytes/sector
* 286739329 sectors
* 286739262 accessible sectors
*
* Flags:
*    1: unmountable
*   10: read-only
*
* Unallocated space:
*           First      Sector      Last
*           Sector   Count      Sector
*              34      222          255*
*                            First      Sector         Last
* Partition Tag Flags Sector   Count         Sector    Mount Directory
        0       4     00   256      286722656 286722911
        8      11    00   286722912 16384   286739295

-----------------------------------------------------------------------------------------

If you look at partition 0 you can see that sector 0 does not start at block 0 this is a EFI label.

-----------------------------------------------------------------------------------------

If you have a SAS/SATA/SCSI disk you can utilize "cfgadm" to deconfigure your device.  

 "Solaris 10 Replacing SCSI/SATA/SAS Disk"

OR

If you have a fiberdisk, (V480,V880,V490,V890, you will not be able to use cfgadm to unconfigure the disk.  Utilize luxadm utility to replace it.

How to Identify and Remove a Failed Internal Disk in Sun Fire V480, V490 Under MPxIO Control (Doc ID 2017748.1)

Once disk is unconfigured it is ok to remove it from the system. Insert the new disk and reconfigure it into the system.


 Now  label the disk.  It is important to complete this step because it puts a default partition table on it so we can create the  new partition table in the next command

format> label
[0] SMI Label
[1] EFI Label
Specify Label type[1]: 0 <<< If working disk is EFI labeled choose "1" if working disk is SMI choose "0"
Auto configuration via format.dat[no]? y

Exit format

# prtvtoc /dev/dsk/c0t1d0s2 |fmthard -s - /dev/rdsk/c0t0d0s2
fmthard: New volume table of contents now in place.

 

Now follow the procedure to rebuild partitions and add the disk back to the correct root pool.

In this example c0t0d0s0 which was the replacement disk was changed to an SMI label to match c0t1d0s0.

If it was found that the working mirror disk was EFI then we would have changed it to a EFI label with the label command under format -e.

Once the labels match you are ready to replace the defective mirror


 Exception Information

If issues arise and you cannot get disk to mirror the following doc will provide more reference for the "# format -e > type" command and use the type menu item on the replacement.  See Utilizing the type command to assist mirroring.

References

<NOTE:2049270.1> - Solaris 10 Fails To Replace a SCSI Disk under SVM - Application(s) accessing previously attached SATA device have to release it before newly inserted device can be made accessible.
<NOTE:2287066.1> - Using a 146GB FC-AL disk to replace a 72GB FC-AL disk which is part of a mirrored volume.
<NOTE:1010753.1> - Solaris Volume Manager (SVM) How to Replace Internal System FC-AL Disks in 280R, V480, V490, V880, V890, and E3500 Servers
<NOTE:2328288.1> - How to Replace a Disk in a ZFS rpool for a Sparc Solaris System

Attachments
This solution has no attachment
  Copyright © 2018 Oracle, Inc.  All rights reserved.
 Feedback