[솔라리스]디스크 미러링

0070. [솔라리스] 디스크 미러링

RAID Information

The Sun Fire V215 and V245 servers support RAID 0 and RAID 1 configurations. You should use the Solaris OS raidctl(1M) command to configure the volumes.

With Solaris 10 8/07 OS release, the raidctl(1M) command has been updated. The usage of the command has changed dramatically. Two potential areas of concern are:

• The previous version of the command required entering only raidctl to get volume and disk status. The command now requires the -l option and the volume-name argument. The change was prompted by the concern here a disk or volume in a nonoptimal or failed state might go unnoticed.
• There are options and features listed that are not supported on many platforms (for example, hot-spares) that might cause you to think they are available.

Before you use the command, read the updated raidctl(1M) man page.

For more information about RAID, the use of the raidctl(1M) command, and managing volumes, refer to the following resources:

• BigAdmin site: http://www.sun.com/bigadmin/home/

There are numerous articles on RAID that can be found by searching on RAID.

• Sun BluePrints site: http://www.sun.com/blueprints/
• Solaris Volume Manager Administration Guide (819-2789), which can be found at the following site:

http://docs.sun.com/

• raidctl(1M) man page, which can be found in the Solaris 10 Reference Manual Collection at:

http://docs.sun.com/

• VERITAS Volume Manager Administration Guide, Solaris (875-3529-10), which can be found at:

http://www.sun.com/products-n-solutions/hardware/docs/

 

Caution - If you remove the system board, you must run service-only OpenBoot PROM commands to recover the volumes. Contact your service representative for more information about these commands.

 

Note - The raidctl(1M) command does not support non-hot-pluggable devices. To service non-hot-pluggable components, you must shut down the server.

 

RAID Technology

There are several levels of RAID configurations that provide varying degrees of data availability with corresponding trade-offs in performance and cost. This section describes the RAID configurations that are supported on the Sun Fire V215 and V245 servers.

RAID 0 - Disk Striping

Disk striping (RAID 0) is a technique for increasing system throughput by using several disk drives in parallel. In nonstriped disks the operating system writes a single block to a single disk. In a striped arrangement, each block is divided and portions of the data are written to different disks simultaneously. No data is shared. Data is written across all disks (concatenated) in exactly one location. Write operations are committed in a round-robin fashion across all member disks, so performance is increased in some workload scenarios. Read operations are satisfied at one location (performs as a single disk).

When you create a RAID volume, all data contained on the drive is lost. If a single disk fails, data is lost. If a RAID 0 volume is disabled (broken), data is lost.

System performance using RAID 0 will be better than using RAID 1, but the possibility of data loss is greater because there is no way to retrieve or reconstruct data stored on a failed disk drive.

RAID 1 - Disk Mirroring

Disk mirroring (RAID 1) is a technique that uses data redundancy (two complete copies of all data stored on two separate disks) to protect against loss of data due to disk failure. One logical volume is duplicated on two separate disks (primary and secondary configuration). You can configure hardware disk mirroring using the SAS controller. This configuration provides higher performance than conventional software mirroring using volume management software. If you mirror the boot drive, you must create the RAID volume before you install the operating system.

Whenever the operating system needs to write to a mirrored volume, both disks are updated. The disks are maintained at all times with exactly the same information. When the operating system needs to read from the mirrored volume, the operating system reads from whichever disk is more readily accessible at the moment, which can result in enhanced performance for read operations.

If a single disk fails, the data is still available, but the volume will be in the degraded mode. You can replace and resynchronize the failed disk. If the volume is disabled (that is, the mirror is broken), you must relabel the drives, which causes all of the data to be lost.

On the Sun Fire V215 and V245 servers, the SAS controller supports internal hardware disk mirroring using the Solaris OS raidctl utility. A hardware disk mirror created under the raidctl utility behaves slightly differently than a disk mirror created using volume management software. Under a software mirror, each device has its own entry in the virtual device tree, and read-write operations are performed to both virtual devices. Under hardware disk mirroring, only one device (the master) appears in the device tree. The mirrored device (the slave) is invisible to the OS and is accessed only by the SAS controller.

 

Caution - Creating or restoring a disk mirror destroys all data previously stored on the disk drive.

 

RAID Volume and Drive Status

After you create a RAID volume or during the normal operation of the system, the volume can have one of the following status conditions:

• OK - The volume is optimal.
• DEGRADED - The volume has a failed drive (mirror) or is otherwise functionally minimized.
• RESYNCING - The volume is synchronizing (mirror).
• FAILED - The volume is offline.

At any time during the normal operation of the system, any one or more of the drives can have one of the following status conditions:

• OK - Disk is operating normally.
• MISSING - Disk is removed or otherwise offline.
• FAILED - Disk is in a failed state and needs service.

Physical Disk Slot Numbers, Physical Device Names, and Logical Device Names

To perform a RAID procedure, you must know the physical or logical device name for the drive that you want to use. You can use TABLE 2-1 to associate internal disk slot numbers with the logical and physical device names for each hard disk drive.

1. Verify that a hardware disk path:

 

# iostat -nE c0t1d0 Soft Errors: 0 Hard Errors: 0 Transport Errors: 0 Vendor: SEAGATE Product: ST914602SSUN146G Revision: 0603 Serial No: 081795NJZE Size: 146.56GB <146561286144 bytes> Media Error: 0 Device Not Ready: 0 No Device: 0 Recoverable: 0 Illegal Request: 0 Predictive Failure Analysis: 0 c0t0d0 Soft Errors: 0 Hard Errors: 0 Transport Errors: 0 Vendor: SEAGATE Product: ST914602SSUN146G Revision: 0603 Serial No: 081995W462 Size: 146.56GB <146561286144 bytes> Media Error: 0 Device Not Ready: 0 No Device: 0 Recoverable: 0 Illegal Request: 0 Predictive Failure Analysis: 0

 

The example indicates that no RAID volume exists. The logical device names might appear differently on your system, depending on the number and type of add-on disk controllers installed.

TABLE 2-1 Disk Slot Numbers, Logical Device Names, and Physical Device Names
Disk Slot Number	Logical Device Name[1]	Physical Device Name
Slot 0	c0t0d0	/pci@1e,600000/pci@0/pci@a/pci@0/pci@8/scsi@1/sd@0,0
Slot 1	c0t1d0	/pci@1e,600000/pci@0/pci@a/pci@0/pci@8/scsi@1/sd@1,0
Slot 2	c0t2d0	/pci@1e,600000/pci@0/pci@a/pci@0/pci@8/scsi@1/sd@2,0
Slot 3	c0t3d0	/pci@1e,600000/pci@0/pci@a/pci@0/pci@8/scsi@1/sd@3,0

 

RAID Operations With the Solaris OS Commands

This section contains instructions on how to create RAID volumes using the Solaris OS commands.

Creating a Disk Stripe With the Solaris OS Commands

When you create a disk stripe on two or more drives, the data on all of the drives will be lost. In addition, no file systems can be mounted on any of these drives.

The RAID volume has the ability to use the capacity of all the disks that are combined. However, the smallest common size is used.

 

To Create a Disk Stripe With the Solaris OS Commands

1. Verify that a hardware disk stripe does not already exist by typing:

 

# raidctl -l No RAID volumes found.

 

The example indicates that no RAID volume exists. The logical device names might appear differently on your system, depending on the number and type of add-on disk controllers installed.

2. Type the following command:

 

# raidctl -c -r 0 disk1disk2 [disk3]

 

For example:

 

# raidctl -c -r 0 c0t1d0 c0t2d0 c0t3d0

 

The RAID volume c0t1d0 is created.

Before you can use the volume, you must label it in the Solaris OS by using the format(1M) command. For instructions on labeling the volume, see Labeling a Volume.

Creating a Disk Mirror With the Solaris OS Commands

When you create a disk mirror, all of the data on all of the drives is lost. The RAID volume will have the capacity of one drive. No file systems can be mounted on either drive. The secondary drive must be as large as the primary drive.

 

To Create a Disk Mirror With the Solaris OS Commands

1. Verify that a hardware disk mirror does not already exist by typing:

 

# raidctl -l No RAID volumes found.

 

The example indicates that no RAID volume exists. The logical device names might appear differently on your system, depending on the number and type of add-on disk controllers installed.

2. Type the following command:

 

# raidctl -c master slave

 

For example:

 

# raidctl -c c0t0d0 c0t1d0

 

RAID volume c0t0d0 is created. When you create a RAID mirror, the slave drive (in this case, c0t1d0) disappears from the Solaris OS device tree.

3. To check the status of a RAID mirror, type the following command:

 

# raidctl -l RAID RAID RAID Disk Volume Status Disk Status -------------------------------------------------------- c0t0d0 RESYNCING c0t0d0 OK c0t1d0 OK

 

The example indicates that the RAID mirror is still resynchronizing with the backup drive.

The following example shows that the RAID mirror is completely restored and online.

 

# raidctl -l RAID RAID RAID Disk Volume Status Disk Status ------------------------------------ c0t0d0 OK c0t0d0 OK c0t1d0 OK

 

Under RAID 1 (disk mirroring), all data is duplicated on both drives. If a drive fails, replace that drive with a working drive and restore the mirror.

For more information about the raidctl utility, see the raidctl(1M) man page.

Before you can use the volume, you must label it in the Solaris OS by using the format(1M) command. For instructions on labeling the volume, see Labeling a Volume.

Deleting a Disk Mirror With the Solaris OS Commands

Before you perform the procedure in this section, verify which disk drive corresponds with which logical device name and physical device name. For more information, see Physical Disk Slot Numbers, Physical Device Names, and Logical Device Names.

 

To Delete a Disk Mirror With the Solaris OS Commands

1. Determine the name of the mirrored volume:

 

# raidctl -l RAID RAID RAID Disk Volume Status Disk Status ------------------------------------ c0t0d0 OK c0t0d0 OK c0t1d0 OK

 

In this example, the mirrored volume is c0t0d0.

 

Note - The logical device names might appear differently on your system, depending on the number and type of add-on disk controllers installed.

 

2. To delete the volume, type the following command:

 

# raidctl -d mirrored-volume

 

For example:

 

# raidctl -d c0t0d0 RAID Volume 'c0t0d0' deleted

 

3. To confirm that you have deleted the RAID array, type the following command:

 

# raidctl -l No RAID volumes found

 

For more information, see the raidctl(1M) man page.

After you delete the volume, you must relabel the drive in the Solaris OS by using the format(1M) command.

Updating the RAID Controller Firmware

The RAID controller firmware is part of the OpenBoot PROM software patch. If you need to update the firmware for the RAID controller, download and install the latest OpenBoot PROM software patch.

RAID Operations With the OpenBoot PROM Commands

If you have existing RAID volumes and have to replace a component, such as the system board, you must perform the procedure in this section so that the Solaris OS can see the volume.

 

Caution - The OpenBoot command, activate-volumes, should be run only by a qualified service representative.

 

In the following procedure, the show-disks command is used to determine the controller path. You can also use the probe-scsi-all command to find the path.

 

To Recover a RAID Volume

1. Set the auto-boot? variable to false.

 

ok setenv auto-boot? false auto-boot? = false

 

2. Set the fcode-debug? variable to true.

 

ok setenv fcode-debug? true fcode-debug? = true

 

3. Reset the system.

 

ok reset-all

 

4. Type the show-disks command to determine the path of the controller.

 

ok show-disks a) /pci@1e,600000/pci@0/pci@a/pci@0/pci@8/scsi@1/disk b) /pci@1e,600000/pci@0/pci@1/pci@0/ide@1f/cdrom c) /pci@1e,600000/pci@0/pci@1/pci@0/ide@1f/disk q) NO SELECTION

 

5. Enter the appropriate selection, type Control-Y to insert the path into the command line, then press Enter.

 

Enter Selection, q to quit: a /pci@1e,600000/pci@0/pci@a/pci@0/pci@8/scsi@1/disk has been selected. Type ^Y ( Control-Y ) to insert it in the command line. e.g. ok nvalias mydev ^Y for creating devalias mydev for /pci@1e,600000/pci@0/pci@a/pci@0/pci@8/scsi@1/disk     ok select /pci@1e,600000/pci@0/pci@a/pci@0/pci@8/scsi@1

 

 

Note - The /disk portion of the path was left off the path because it is the controller you want to select.

 

6. Type the show-volumes command to determine the inactive volume:

 

ok show-volumes

 

7. Activate the volume using the volume number:

 

ok volume-number activate-volume

 

8. Unselect the device:

 

ok unselect-dev

 

9. Reset the OpenBoot PROM variables, and reset the system after you have finished managing the RAID volumes.

 

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

 

If the auto-boot? variable was set to true, you might see the following message:

 

Base SAS World Wide ID is 0! This must be fixed immediately using set-sas-wwid

 

If you see this message, set the auto-boot? variable to false, and issue the reset-all command.

Labeling a Volume

After you create a RAID volume by using the raidctl(1M) command or by using the OpenBoot PROM commands, you must label the volume correctly before the volume can be used.

 

To Label a Volume

The Solaris OS must be running to perform this procedure. If you have created the volume by using the OpenBoot PROM commands, start the Solaris OS and log in as the superuser.

1. Run the format(1M) command, and choose the newly created volume from the list of drives.

You might see messages about the volume having a corrupt label or the block count being incorrect when booting or running the format command. This command needs a correct label.

2. Chose type from the Format menu.

This action displays a list of available drive types.

3. Chose Auto configure from the list.

This action defines the volume with the correct type.

4. Write the new label to the volume using the label command from the format> prompt.

The system asks you if you are ready to label the disk.

5. Type yes.

After you perform these steps, the drive is ready for use. You can now partition the drive as needed and use the volume within the Solaris OS.