DEFINE STGPOOL (Define an active-data pool assigned to sequential-access devices)

Privilege class

To issue this command, you must have system privilege.

Parameters

pool_name (Required)

Specifies the name of the storage pool to be defined. The name must be unique, and the maximum length is 30 characters.

device_class_name (Required)

Specifies the name of the sequential access device class to which this active-data pool is assigned. You can specify any device class except DISK.

POoltype=ACTIVEdata (Required)

Specifies that you want to define an active-data pool.

DESCription

Specifies a description of the active-data pool. This parameter is optional. The maximum length of the description is 255 characters. Enclose the description in quotation marks if it contains any blank characters.

ACCess

Specifies how client nodes and server processes (such as reclamation) can access files in the active-data pool. This parameter is optional. The default value is READWRITE. Possible values are:

READWrite

Specifies that files can be read from and written to the volumes in the active-data pool.

READOnly

Specifies that client nodes can only read files that are stored on the volumes in the active-data pool.

Server processes can move files within the volumes in the storage pool. The server can use files in the active-data pool to restore files to primary storage pools. However, no new writes are permitted to volumes in the active-data pool from volumes outside the storage pool. A storage pool cannot be copied to the active-data pool.

UNAVailable

Specifies that client nodes cannot access files that are stored on volumes in the active-data pool.

Server processes can move files within the volumes in the storage pool. The server can use files in the active-data pool to restore files to primary storage pools. However, no new writes are permitted to volumes in the active-data pool from volumes outside the storage pool. A storage pool cannot be copied to the active-data pool.

COLlocate

Specifies whether the server attempts to keep data, stored on as few volumes as possible, that belong to one of the following candidates:

• A single client node
• A group of file spaces
• A group of client nodes
• A client file space

This parameter is optional. The default value is NO.

Collocation reduces the number of sequential access media mounts for restore, retrieve, and recall operations. However, collocation increases both the amount of server time that is needed to collocate files for storing and the number of volumes required. Collocation can also impact the number of processes migrating disks to sequential pool.

You can specify one of the following options:

No

Specifies that collocation is disabled. During migration from disk, processes are created at a file space level.

GRoup

Specifies that collocation is enabled at the group level for client nodes or file spaces. For collocation groups, the server attempts to put data for nodes or file spaces that belong to the same collocation group on as few volumes as possible.

If you specify COLLOCATE=GROUP but do not define any collocation groups, or if you do not add nodes or file spaces to a collocation group, data is collocated by node. Consider tape usage when you organize client nodes or file spaces into collocation groups.

For example, if a tape-based storage pool consists of data from nodes and you specify COLLOCATE=GROUP, the server completes the following actions:

• Collocates the data by group for grouped nodes. Whenever possible, the server collocates data that belongs to a group of nodes on a single tape or on as few tapes as possible. Data for a single node can also be spread across several tapes that are associated with a group.
• Collocates the data by node for ungrouped nodes. Whenever possible, the server stores the data for a single node on a single tape. All available tapes that already have data for the node are used before available space on any other tape is used.
• During migration from disk, the server creates migration processes at the collocation group level for grouped nodes, and at the node level for ungrouped nodes.

If a tape-based storage pool consists of data from grouped file spaces and you specify COLLOCATE=GROUP, the server completes the following actions:

• Collocates by group, the data for grouped file spaces only. Whenever possible, the server collocates data that belongs to a group of file spaces on a single tape or on as few tapes as possible. Data for a single file space can also be spread across several tapes that are associated with a group.
• Collocates the data by node (for file spaces that are not explicitly defined to a file space collocation group). For example, node1 has file spaces named A, B, C, D, and E. File spaces A and B belong to a filespace collocation group but C, D, and E do not. File spaces A and B are collocated by filespace collocation group, while C, D, and E are collocated by node.
• During migration from disk, the server creates migration processes at the collocation group level for grouped file spaces.

Data is collocated on the least amount of sequential access volumes.

NODe

Specifies that collocation is enabled at the client node level. For collocation groups, the server attempts to put data for one node on as few volumes as possible. If the node has multiple file spaces, the server does not try to collocate those file spaces. For compatibility with an earlier version, COLLOCATE=YES is still accepted by the server to specify collocation at the client node level.

If a storage pool contains data for a node that is a member of a collocation group and you specify COLLOCATE=NODE, the data is collocated by node.

For COLLOCATE=NODE, the server creates processes at the node level when you migrate data from disk.

FIlespace

Specifies that collocation is enabled at the file space level for client nodes. The server attempts to place data for one node and file space on as few volumes as possible. If a node has multiple file spaces, the server attempts to place data for different file spaces on different volumes.

For COLLOCATE=FILESPACE, the server creates processes at the file space level when you migrate data from disk.

REClaim

Specifies when the server reclaims a volume, which is based on the percentage of reclaimable space on a volume. Reclaimable space is the amount of space that is occupied by files that are expired or deleted from the Tivoli® Storage Manager database.

Reclamation makes the fragmented space and space occupied by inactive backup files on volumes usable again by moving any remaining unexpired files and active backup files from one volume to another volume. This makes the original volume available for reuse. This parameter is optional. You can specify an integer from 1 to 100. The default value is 60.

When determining which volumes in a storage pool to reclaim, the Tivoli Storage Manager server first determines the reclamation threshold indicated by the RECLAIM. The server then examines the percentage of reclaimable space for each volume in the storage pool. If the percentage of reclaimable space on a volume is greater that the reclamation threshold of the storage pool, the volume is a candidate for reclamation.

For example, suppose storage pool FILEPOOL has a reclamation threshold of 70 percent. This value indicates that the server can reclaim any volume in the storage pool that has a percentage of reclaimable space that is greater that 70 percent. The storage pool has three volumes:

• FILEVOL1 with 65 percent reclaimable space
• FILEVOL2 with 80 percent reclaimable space
• FILEVOL3 with 95 percent reclaimable space

When reclamation begins, the server compares the percent of reclaimable space for each volume with the reclamation threshold of 70 percent. In this example, FILEVOL2 and FILEVOL3 are candidates for reclamation because their percentages of reclaimable space are greater than 70. To determine the percentage of reclaimable space for a volume, issue the QUERY VOLUME command and specify FORMAT=DETAILED. The value in the field Pct. Reclaimable Space is the percentage of reclaimable space for the volume.

If you change the value from the default, specify a value of 50 percent or greater so that files stored on two volumes can be combined onto a single output volume.

When an active-data pool volume that is offsite becomes eligible for reclamation, the reclamation process attempts to obtain the unexpired files on the reclaimable volume from a primary or active-data pool that is onsite. The process then writes these files to an available volume in the original active-data pool. Effectively, these files are moved back to the onsite location. However, the files can be obtained from the offsite volume after a disaster if a database backup is used that references the files on the offsite volume. Because of the way reclamation works with offsite volumes, use it carefully with active-data pools.

RECLAIMPRocess

Specifies the number of parallel processes to use for reclaiming the volumes in this storage pool. This parameter is optional. Enter a value from 1 to 999. The default value is 1.

When calculating the value for this parameter, consider the number of sequential storage pools that will be involved with the reclamation and the number of logical and physical drives that can be dedicated to the operation. To access a sequential access volume, IBM® Tivoli Storage Manager uses a mount point and, if the device type is not FILE, a physical drive. The number of available mount points and drives depends on other Tivoli Storage Manager and system activity and on the mount limits of the device classes for the sequential access storage pools that are involved in the reclamation.

For example, suppose that you want to reclaim the volumes from two sequential storage pools simultaneously and that you want to specify four processes for each of the storage pools. The storage pools have the same device class. Each process requires two mount points and, if the device type is not FILE, two drives. (One of the drives is for the input volume, and the other drive is for the output volume.) To run eight reclamation processes simultaneously, you need a total of at least 16 mount points and 16 drives. The device class for the storage pools must have a mount limit of at least 16.

If the number of reclamation processes you specify is more than the number of available mount points or drives, the processes that do not obtain mount points or drives will wait for mount points or drives to become available. If mount points or drives do not become available within the MOUNTWAIT time, the reclamation processes will end. For information about specifying the MOUNTWAIT time, see DEFINE DEVCLASS (Define a device class).

The Tivoli Storage Manager server will start the specified number of reclamation processes regardless of the number of volumes that are eligible for reclamation. For example, if you specify ten reclamation processes and only six volumes are eligible for reclamation, the server will start ten processes and four of them will complete without processing a volume.

RECLAMATIONType

Specifies the method by which volumes are reclaimed and managed. This parameter is optional. The default value is THRESHOLD. Possible values are the following:

THRESHold

Specifies that volumes belonging to this storage pool are reclaimed based on the threshold value in the RECLAIM attribute for this storage pool.

SNAPlock

Specifies that FILE volumes belonging to this storage pool will be managed for retention using NetApp Data ONTAP software and NetApp SnapLock volumes. This parameter is only valid for storage pools being defined to a server that has data retention protection enabled and that is assigned to a FILE device class. Volumes in this storage pool are not reclaimed based on threshold; the RECLAIM value for the storage pool is ignored.

All volumes in this storage pool are created as FILE volumes. A retention date, derived from the retention attributes in the archive copy group for the storage pool, is set in the metadata for the FILE volume using the SnapLock feature of the NetApp Data ONTAP operating system. Until the retention date has expired, the FILE volume and any data on it cannot be deleted from the physical SnapLock volume on which it is stored.

The RECLAMATIONTYPE parameter for all storage pools being defined must be the same when defined to the same device class name. The DEFINE command fails if the RECLAMATIONTYPE parameter specified is different from what is currently defined for storage pools that are already defined to the device class name.

OFFSITERECLAIMLimit

Specifies the number of offsite volumes to have their space reclaimed during reclamation for this storage pool. This parameter is optional. The default value is NOLIMIT. Possible values are:

NOLimit

Specifies that you want to have the space reclaimed in all of your offsite volumes.

number

Specifies the number of offsite volumes to have their space reclaimed. You can specify an integer from 0 to 99999. A value of zero means that none of the offsite volumes will be reclaimed.

Important: When determining the value for the OFFSITERECLAIMLIMIT, consider using the statistical information in the message issued at the end of the offsite volume reclamation operation. Alternatively, you can use the following Tivoli Storage Manager SQL select command to obtain the statistical information from the SUMMARY table for the offsite volume reclamation operation:

select * from summary where activity='OFFSITE RECLAMATION'

The statistical information includes the following items:

• The number of offsite volumes that were processed
• The number of parallel processes that were used
• The total amount of time required for the processing

The order in which offsite volumes are reclaimed is based on the amount of unused space in a volume. (Unused space includes both space that has never been used on the volume and space that has become empty because of file deletion.) Volumes with the largest amount of unused space are reclaimed first.

For example, suppose an active-data pool contains three volumes: VOL1, VOL2, and VOL3. VOL1 has the largest amount of unused space, and VOL3 has the least amount of unused space. Suppose further that the percentage of unused space in each of the three volumes is greater than the value of the RECLAIM parameter. If you do not specify a value for the OFFSITERECLAIMLIMIT parameter, all three volumes are reclaimed when the reclamation runs. If you specify a value of 2, only VOL1 and VOL2 are reclaimed when the reclamation runs. If you specify a value of 1, only VOL1 is reclaimed.

MAXSCRatch (Required)

Specifies the maximum number of scratch volumes that the server can request for this storage pool. You can specify an integer from 0 to 100000000. By allowing the server to request scratch volumes as needed, you avoid having to define each volume to be used.

The value specified for this parameter is used to estimate the total number of volumes available in the active-data pool and the corresponding estimated capacity for the active-data pool.

Scratch volumes are automatically deleted from the storage pool when they become empty. However, if the access mode for a scratch volume is OFFSITE, the volume is not deleted from the active-data pool until the access mode is changed. This allows an administrator to query the server for empty, offsite scratch volumes and return these to the onsite location.

When scratch volumes with the device type of FILE become empty and are deleted, the space that the volumes occupied is freed by the server and returned to the file system.

Tip: For server-to-server operations that use virtual volumes and that store a small amount of data, consider specifying a value for the MAXSCRATCH parameter that is higher than the value you typically specify for write operations to other types of volumes. After a write operation to a virtual volume, Tivoli Storage Manager marks the volume as FULL, even if the value of the MAXCAPACITY parameter on the device-class definition has not been reached. The Tivoli Storage Manager server does not keep virtual volumes in FILLING status and does not append to them. If the value of the MAXSCRATCH parameter is too low, server-to-server operations can fail.

REUsedelay

Specifies the number of days that must elapse after all files are deleted from a volume before the volume can be rewritten or returned to the scratch pool. This parameter is optional. You can specify an integer from 0 to 9999. The default value is 0, which means that a volume can be rewritten or returned to the scratch pool as soon as all the files are deleted from the volume.

Important: Use this parameter to help ensure that when you restore the database to an earlier level, database references to files in the active-data pool are still valid. You must set this parameter to a value greater than the number of days you plan to retain the oldest database backup. The number of days that are specified for this parameter should be the same as the number specified for the SET DRMDBBACKUPEXPIREDAYS command.

OVFLOcation

Specifies the overflow location for the storage pool. The server assigns this location name to a volume that is ejected from the library by the command. This parameter is optional. The location name can be a maximum length of 255 characters. Enclose the location name in quotation marks if the location name contains any blank characters.

DATAFormat

Specifies the data format to use to copy files to this storage pool and restore files from this storage pool. The default format is the NATIVE server format. Possible values are:

NATive

Specifies the data format is the native IBM Tivoli Storage Manager server format and includes block headers.

NONblock

Specifies the data format is the native IBM Tivoli Storage Manager server format and does not include block headers.

Important: The default minimum block size on a volume that is associated with a FILE device class is 256 KB, regardless how much data is being written to the volume. For certain tasks (for example, using content-management products, using the DIRMC client option to store directory information, or migrating very small files using Tivoli Storage Manager for Space Management or Tivoli Storage Manager HSM for Windows), you can minimize wasted space on storage volumes by specifying the NONBLOCK data format. In most situations, however, the NATIVE format is preferred.

CRCData

Specifies whether a cyclic redundancy check (CRC) validates storage pool data when audit volume processing occurs on the server. This parameter is only valid for NATIVE data format storage pools. This parameter is optional. The default value is NO. By setting CRCDATA to YES and scheduling an AUDIT VOLUME command you can continually ensure the integrity of data stored in your storage hierarchy. Possible values are:

Yes

Specifies that data is stored containing CRC information, allowing for audit volume processing to validate storage pool data. This mode impacts performance because additional overhead is required to calculate and compare CRC values between the storage pool and the server.

No

Specifies that data is stored without CRC information.

Tip: For storage pools that are associated with the 3592, LTO, or ECARTRIDGE device type, logical block protection provides better protection against data corruption than CRC validation for a storage pool. If you specify CRC validation for a storage pool, data is validated only during volume auditing operations. Errors are identified after data is written to tape.

To enable logical block protection, specify a value of READWRITE for the LBPROTECT parameter on the DEFINE DEVCLASS and UPDATE DEVCLASS commands for the 3592, LTO, or ECARTRIDGE device types. Logical block protection is supported only on the following types of drives and media:

• IBM LTO5 and later.
• IBM 3592 Generation 3 drives and later with 3592 Generation 2 media and later.
• Oracle StorageTek T10000C drives.

DEDUPlicate

Specifies whether the data that is stored in this storage pool will be deduplicated. This parameter is optional and is valid only for storage pools that are defined with a FILE device class. The default value is NO.

IDENTIFYPRocess

Specifies the number of parallel processes to use for server-side duplicate identification. This parameter is optional and is valid only for storage pools that are defined with a FILE device class. Enter a value from 0 to 50.

The default value for this parameter is 0. Duplicate-identification processes for a copy storage pool are not necessary if you specify duplicate-identification processes for the primary storage pool. When Tivoli Storage Manager analyzes a file in a storage pool, Tivoli Storage Manager also analyzes the file in all other storage pools.

When calculating the value for this parameter, consider the workload on the server and the amount of data requiring data deduplication. Server-side duplicate identification requires disk I/O and processor resources, so the more processes you allocate to data deduplication, the heavier the workload that you place on your system. In addition, consider the number of volumes that require processing. Server-side duplicate-identification processes work on volumes containing data that requires deduplication. If you update a storage pool, specifying that the data in the storage pool is to be deduplicated, all the volumes in the pool require processing. For this reason, you might have to define a high number of duplicate-identification processes initially. Over time, however, as existing volumes are processed, only the volumes containing new data have to be processed. When that happens, you can reduce the number of duplicate-identification processes.

Remember: Duplicate-identification processes can be either active or idle. Processes that are working on files are active. Processes that are waiting for files to work on are idle. Processes remain idle until volumes with data to be deduplicated become available. The output of the QUERY PROCESS command for a duplicate-identification process includes the total number of bytes and files that have been processed since the process first started. For example, if a duplicate-identification process processes four files, becomes idle, and then processes five more files, then the total number of files processed is nine. Processes end only when canceled or when the number of duplicate-identification processes for the storage pool is changed to a value less than the number currently specified.