Go to main content
Oracle® ZFS Storage Appliance Administration Guide, Release OS8.7.0

Exit Print View

Updated: July 2017
 
 

Creating a Storage Pool (CLI)

Storage pools store data and can be created during or after initial configuration. Pools can contain data drives, and log, read cache, and meta devices. The following task assumes that initial configuration has been completed. Creating and configuring a storage pool is a two-step process. First, the storage devices are verified for presence and minimum functionality, and you assign drives or even entire disk shelves to the pool. Second, you select a profile for the drives based on your storage needs. If for some reason a pool is unconfigured, you can import it as described in Importing an Existing Storage Pool (CLI).

To reduce redundant data, which can be especially prevalent in replication workloads, consider the benefits of using deduplication. Allocate meta devices if deduplication will be enabled for projects or shares in this pool. For more information, see Data Deduplication. There is also an all-flash storage pool, which utilizes SSDs as data devices and optional log devices, but does not contain read cache or meta devices. See Configuring an All-Flash Storage Pool (CLI).

Before You Begin

  1. Go to configuration storage.
  2. Enter config and a name for the new storage pool.
    hostname: configuration storage> config pool0
    hostname: configuration storage (pool0) verify>
  3. Enter show to see the device information for the pool:
    hostname:configuration storage (pool0) verify> show
    ID STATUS  ALLOCATION  DATA   LOG      CACHE   META     RPM
    0      ok     custom      0   0          0/4    0/4    1.86T
    1      ok     custom      0   0/2  34G     0      0    15000
    2      ok     custom      0   0/2  34G     0      0    15000
  4. Enter set and the disk shelf or controller ID, and the number of data drives to use. You can also select available cache, meta, and log devices.

    For more information on log, cache, and meta devices, see Data Profiles for Storage Pools.


    Caution

    Caution  -  Once a data disk has been added to a pool, it cannot be removed without destroying the pool entirely and losing all data.


    ID "0" is the controller, and the remaining IDs are the disk shelves. In the following example, 1-data=8 allocates eight data drives from the first disk shelf.

    hostname:configuration storage (pool1) verify> set 1-data=8
                                1-data = 8

    This example allocates one cache device from the controller:

    hostname:configuration storage (pool1) verify> set 0-cache=1
                                0-cache = 1

    This example allocates one meta device from the controller:

    hostname:configuration storage (pool1) verify> set 0-meta=1
                                0-meta = 1
  5. Enter done.
    hostname:configuration storage (pool1) verify> done
  6. Enter show to display the profile.
    hostname:configuration storage (pool1) config> show
    PROFILE                 CAPCITY  NSPF  DESCRIPTION
    log_profile = log_stripe   17G   no    Striped log

    Note -  If you allocated cache devices to the pool, the profile is always striped.
  7. If you allocated log devices to the pool, enter set log_profile= and set the log profile to either log_mirror or log_stripe. Use log_mirror if the pool contains an even number of log devices.

    Caution

    Caution  -  A double failure can cause loss of data from a log in a striped configuration. It is highly recommended to configure a mirrored log profile for added redundancy. For more information, see Data Profiles for Storage Pools.


    hostname:configuration storage (pool1)> set log_profile=log_mirror
  8. If you allocated meta devices to the pool, enter set meta_profile= and set the meta profile to either meta_mirror or meta_stripe.
    hostname:configuration storage (pool1)> set meta_profile=meta_mirror
  9. Enter done to complete the task.
    hostname:configuration storage (pool1)> done

Related Topics