Administering AFM

The following topics assist you in Administering AFM.

Migrating data by using active file management
When you migrate data at the fileset level, an AFM fileset is created in the read-only (RO) mode to pull the data. The RO-mode fileset is converted into an LU-mode AFM fileset after the migration for local read-write. AFM LU-enabled applications perform the read-write operations unlike the AFM RO mode where applications can perform only read operations.
Creating an AFM relationship by using the NFS protocol
Creating an AFM relationship by using GPFS protocol
The following topics describe how to set up the home and cache cluster.
Checking the synchronization status of an AFM fileset
During the data synchronization, check the status of an AFM fileset by using the mmafmctl command.
Pre-populating metadata by using the out of band prefetch
Sometimes, prefetching of data and metadata from a home target might take more time. During prefetch, AFM queues a lookup operation to a gateway node for each file. This operation might take several minutes or hours based on the amount of data and network latency. If metadata on the home is modified, AFM must re-populate new metadata on an AFM cache fileset.
AFM to cloud object storage policy-based deletion for the manual updates mode
AFM to cloud object storage supports policy-based deletion for a manual updates (MU) mode by using mmafmcosctl delete command. A policy can be defined by system administrators and run by using the mmafmcosctl delete command that helps automatic selection and deletion of the files or objects from cloud object storage buckets. A policy rule is an SQL-like statement that directs the mmafmcosctl delete command to delete files to the cloud object storage based on criteria defined in the policy file.
Improving write and remove operations efficiency in the manual updates mode
The manual updates (MU) mode fileset provides the flexibility to upload and download files or objects to and from cloud object storage after you finalize the set of objects to upload or download. Unlike other AFM to cloud object storage objectfs fileset modes, MU mode depends on manual intervention from administrators to upload and download the data to be in-sync. Administrators can also automate upload and download by using ILM policies to search specific files or objects to upload or download or remove.
Evicting metadata or inode automatically from AFM filesets
IBM storage scale is distributed storage system designed for large-scale data management, to ensure efficient space utilization, it employs a metadata eviction mechanism that automatically removes least recently used metadata or inode entries from AFM cache filesets. Eviction is used to make space for new objects and files on an AFM filesets. Eviction releases metadata (inode) in the fileset, if the fileset inode usage exceeds the fileset quota. This feature is useful when the storage that is provided by IBM Storage Scale is less than the targets or cloud object storage. The quotas can be defined when you set the AFM relationship by using the mmsetquota command.

Parent topic: Administering