Note the following:

• The disk array methods support the Veritas File System (VxFS). With a few exceptions, these methods do not support software-based volume managers such as the Veritas Volume Manager (VxVM) or any native Linux or Solaris volume managers. If your client data is configured in Veritas VxVM volumes, use either the legacy array-specific methods for UNIX clients (TimeFinder, ShadowImage, BusinessCopy) or a software-based snapshot method such as VxVM or FlashSnap.

  Note:

  The following array methods support Veritas Volume Manager (VxVM) volumes: Hitachi_CopyOnWrite and Hitachi_ShadowImage. The IBM_DiskStorage_FlashCopy method (on the IBM DS6000) supports VxVM on the AIX platform.

• Use caution with Instant Recovery rollback. An Instant Recovery point-in-time rollback overwrites the entire LUN (source disk) with the contents of the snapshot or mirror disk. If you have multiple file systems or multiple partitions configured on the hardware array LUN (the source disk), one or more of those file systems or partitions sharing the snapshot disk or mirror disk may have older data that you do not want to write back to the source. When the rollback takes place, any older data on the snapshot disk or mirror disk replaces the newer data on the source.

• In a clustered environment, Instant Recovery point-in-time rollback is not supported.

• Except for required preconfiguration as explained in this chapter, do not perform manual operations on any snapshot resources that NetBackup uses. After the preconfiguration is completed as explained in this chapter, NetBackup automatically manages the required LUNs, snapshots, clones, and mirrors.

  Warning:

  If you make other changes to the snapshot resources, the NetBackup catalog may be invalidated. For instance, restores may fail from backups consisting of the snapshots that have been deleted outside the view of NetBackup.

  For example, DO NOT do the following:

  • Do not remove a snapshot resource that NetBackup created.

  • Do not create a snapshot resource in a storage group.

  • Do not change the state of a snapshot resource, such as by manually resynchronizing it.

  • You should not use array device targets in more than one policy. If you want to reuse devices after destroying an old policy, you need to manually expire the backup images, which exist for that policy.