 | ||
| Storage Checkpoint allocation policies | ||
|---|---|---|
| Prev | Using Storage Checkpoints and Storage Rollback | Next |
The Veritas File System provides Multi-Volume File Systems (MVS) when used in conjunction with the Volumes Set feature in Veritas Volume Manager. A volume set is a container for multiple different volumes. MVS enables creation of a single file system over multiple volumes, each volume with properties of its own. This helps administrators specify which data goes on which volume types. For more details about MVS, see Veritas Volume Manager Administrator's Guide. Setting up a storage configuration for MVS operations is a system administrator's responsibility and requires superuser (root) privileges.
Multi-Volume File Systems provide a database administrator, through the checkpoint administration interface, the ability to create Storage Checkpoint Allocation Policies.
A Storage Checkpoint Allocation policy specifies a list of volumes and the order in which to attempt allocations. Once defined, a database administrator can use these policies to:
Control where the storage checkpoint should be created, enabling separation of metadata and data of a storage checkpoint to different volumes.
Separate storage checkpoints so that data allocated to a storage checkpoint is isolated from the primary file system. This helps control the space used by the checkpoint and prevents the checkpoint from fragmenting the space in the primary fileset.
When policies are assigned to a storage checkpoint, the database administrator must specify the mapping to both metadata and file data. If no policies are specified for the storage checkpoint, the data is placed randomly within the primary file system. Data and metadata of storage checkpoints can have different policies assigned to them or use the same policy to be applied to data and metadata. Multiple checkpoints can be assigned the same checkpoint allocation policy. A partial policy is also allowed; a partial policy means that the policy does not exist on all file systems used by the database.
Once the policy is assigned to checkpoints, the allocation mechanism attempts to satisfy the request from each device in the policy in the order the devices are defined. If the request cannot be satisfied from any of the devices in the policy, the request will fail, even if other devices exist in the file system which have space. Only those devices can provide allocation that are listed in the policy. This implementation is the mechanism for preventing allocation requests from using space in other devices which are not specified in the policy. It is recommended that you allocate sufficient space for the volumes defined in the Storage Checkpoint policy or update the policy to include additional volumes. This also helps in retaining the old Storage Checkpoints.
Once the assigned policy is deleted, the allocation for metadata and file data for subsequent requests of storage checkpoint will return to the no policy assigned state.
For VxFS file systems disk layout Version 7, the volumes on the VxFS Multi-Volume File System cvan be either one of these types: dataonly and metadataok. Only metadataok volumes can be used to store checkpoint metadata. By default, only the first volume that is being added to the VxVM volume set is a metadataok volume. This means only the first volume that is being added to the VxVM volume set can be specified in the ckpt_metadata_policy by default. Use the following file system command to change the default setting. To check the flags of each volume in a VxFS Multi-Volume File System, execute the following file system command as root:
/opt/VRTS/bin/fsvoladm queryflags mountpoint
To change a dataonly volume to a metadataok volume, execute the following file system command as root:
/opt/VRTS/bin/fsvoladm clearflags dataonly mountpoint
vol-name
The following are usage notes for Storage Checkpoint allocation policies: