Adding disks
About adding disks
When you add disks to a filer, you need to insert them in a disk shelf according to the disk shelf manufacturer instructions. They then become hot spare disks. You also add the disks to a volume with the vol add command.
You can use disks of various sizes in a volume according to the following rules:
- If an added disk replaces a failed data disk, its capacity is limited to that of the failed disk.
- If an added disk is not replacing a failed data disk and it is not larger than the parity disk, its full capacity (subject to rounding) is available as a data disk.
When to add disks
You add disks when
- You want to add storage capacity to the filer
- You are running out of hot spare disks
Considerations when adding disks to the filer
Original disk configuration
Network Appliance recommends initial RAID group sizes of 8 to 14 disks. The default RAID group size for Data ONTAP 6.x is 8 disks. Network Appliance recommends adding RAID groups and/or disks before the file system is 80 to 90% full.
The number of drives and RAID groups at initial filer configuration affects read and write performance. A greater number of drives means a greater number of independently seeking disk-drive heads reading data. Write performance can also benefit from more drives; however, the difference can be masked by the effect of NVRAM and the manner in which WAFL manages write operations.
Larger RAID groups increase the probability of a failed disk and increase disk reconstruction time. For best performance, you need to determine the tradeoff between improved performance when all disks are working and performance degradation while a disk is being reconstructed.
Adding more disks to the filer
Additional RAID groups in multiples of 8-14 are recommended for overall improved performance. However, as more disks are configured, the performance increase levels off. Performance is affected more with the fewer disks you add until the striping across all the disks levels out. The more disks you add, the less the performance is affected until the striping across all the disks levels out. When the striping levels out, there is an increase in operations per second and a reduced response time.
Adding more drives will not significantly improve performance if the applied load is not placing stress on the drives in a filer's array. However, this is the time to add more drives to meet future storage requirements.
Adding hot spare disks to the filer
You should periodically check the number of hot spares you have on your filer and add disks if you have fewer than one per shelf. If there are no hot spare disks, you should add disks to disk shelves so they become available as hot spares.
For more information, see Multiple Volumes and Multiple RAID Groups on NetApp Filers, http://www.netapp.com/tech_library/3027.html.
Determining the number of hot spare disks
To determine how many hot spare disks you have on your filer, complete the following step.
Step
|
Action
|
|
1 |
Enter the following command:
vol status -s
Result
If there are hot spare disks, a display like the following appears with a line for each spare disk, grouped by checksum type:
Spare disks for block or zoned checksum volumes
RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys(MB/blks)
--------- ----- ------------ ---- -------------- --------------
spare 7.17 7 2 1 FC:A 0 17046/34910224
spare 7.26 7 3 2 FC:A 0 34209/70061520
...
Spare disks for zoned checksum volumes only
RAID Disk HA.ID HA SHELF BAY CHAN Used (MB/blks) Phys(MB/blks)
--------- ----- ------------ ---- -------------- --------------
spare 7.21 7 2 5 FC:A 0 17366/35566480
spare 7.19 7 2 3 FC:A 0 17366/35566480
|
Rules for adding disks to a volume
You add disks to a volume according to the following rules:
- You can add only hot spare disks to a volume.
- You must specify the volume to which you are adding disks.
- You can add block checksum disks to zoned checksum volumes; you can add only block checksum disks to block checksum volumes.
- Disks must come from the same spare disk pool (if using mirrored volumes).
Parity disk determination
If an added disk is larger than the existing parity disk, the added disk becomes the parity disk, and the former parity disk becomes a data disk. This enforces a Data ONTAP rule that the parity disk must be at least as large as the largest data disk in a RAID group.
Adding new disks to the filer
To add new disks to the filer, complete the following steps.
Step
|
Action
|
|
1 |
If the disks are...
|
Then...
|
|
Fibre channel disks
|
Go to Step 2.
|
|
SCSI disks
|
Enter the following command, then go to Step 2:
disk swap
|
|
2 |
Install one or more disks according to the hardware guide for your disk shelf.
Result
The system displays a message confirming that one or more disks were installed, then waits 15 seconds as the disks are recognized. The system recognizes the disks as hot spare disks.
Note If you add multiple disks, they might require 25 to 40 seconds to come up to speed as the system checks the device addresses on each adapter. |
|
3 |
Verify that the disks were added by entering the following command:
vol status -s
Result
The number of hot spare disks in the RAID Disk column under Spare Disks increases by the number of disks you installed.
|
Adding new disks to a volume
To add new disks to a volume, complete the following steps.
Step
|
Action
|
|
1 |
Verify that there are hot spare disks available for you to add by entering the following command:
vol status -s
|
|
2 |
Add the disks to a volume by entering the following command:
vol add vol-name ndisks[@disk-size] | -d disk-name...
vol-name is the name of the volume to which you are adding disks.
ndisks is the number of disks to use.
disk-size is the disk size, in number of gigabytes, to use. You must have at least ndisks available disks of the size you specify.
disk-name is the disk ID of one or more spare disks; separate disk IDs with a space. The disk ID is under the HA.ID column in the vol status -s display.
Note If you want to use block checksum disks in a zoned checksum volume and there are a still zoned checksum hot spare disks, use the -d option to select the disks. Examples
The following command adds four 36-GB disks to the thisvol volume:
vol add thisvol 4@36
The following command adds the disks 7.17 and 7.26 to the thisvol volume:
vol add thisvol -d 7.17 7.26
|
Note
When adding disks to a volume, Data ONTAP attempts to distribute I/O loads, such as alternating host adapters, to improve reconstruction performance. You can override this behavior by specifying the disks you want.
Adding new disks to a RAID group
If a volume has more than one RAID group, you can specify the RAID group to which you are adding disks. To add new disks to a specific RAID group of a volume, complete the following step.
Step
|
Action
|
|
1 |
Enter the following command:
vol add vol-name ndisks[@disk-size] | -d disk-name... -g raidgroup
raidgroup is a RAID group in the volume specified by vol-name.
Example
The following command adds two disks to RAID group 0 of the vol0 volume:
vol add vol0 -g rg0 2
|
Note
The number of disks you can add to a specific RAID group is limited by the RAID group size. See Chapter 4, RAID Group Management, for more information.
Forcibly adding disks to volumes
The following are situations that would fail if you tried to add disks to a volume:
- The disks specified in the vol add command would cause the disks on an unmirrored volume to span adapters.
- The disks specified in the vol add command add disks to a RAID group other than the last RAID group, thereby making it impossible for the file system to revert to a version prior to Data ONTAP 6.4.
To force Data ONTAP to add disks in these situations, complete the following step.
Step
|
Action
|
|
1 |
Enter the following command:
vol add vol-name -f -g raidgroup -d disk-name...
Note You must use the -g raidgroup option to specify a RAID group other than the last RAID group in the volume. |
Copyright © 1994-2003, Network Appliance, Inc. All rights
reserved.
Part No. 210-00322
File: disks7.htm
Last updated: 01/22/03 16:17:03