Planning for AFM to cloud object storage

The considerations for the planning AFM to cloud object storage relation are as follows:

The AFM to cloud object storage relation is between an AFM to cloud object storage fileset that is hosted on an IBM Storage Scale cluster and cloud object storage bucket, or a directory within an AFM to cloud object storage fileset and a cloud bucket. For more information, see AFM to cloud object storage operation modes. The important elements that need consideration in the IBM Storage Scale cluster are the host system, storage, operating system, and networking. For more information, see Planning for GPFS.

Before you set the AFM to cloud object storage relation, consider the following points:
  1. Cloud object storage such as Amazon S3Start of change, Microsoft Azure Blob,End of change and IBM Cloud® Object Storage must be in accordance with the workload requirement. Access keys and secret keys that are provided by these services need to be set up for AFM to cloud object storage filesets.

    To transfer the data between AFM to cloud object storage and cloud object storage endpoints, select an appropriate protocol, for example, HTTP or HTTPS.

  2. On an IBM Storage Scale cluster, file system space and inodes (number of inodes) need to be planned for an AFM to cloud object storage fileset first before provision. For more information, see File system creation considerations.

    To control the number of inodes that are used by the AFM to cloud object storage fileset, you can specify a limit on the number of inodes for that fileset. To set inodes limit, use the mmchfileset command after the mmafmcosconfig command is run to set the relationship. For more information, see mmchfileset command.

  3. Planning of gateway nodes

    An IBM Storage Scale cluster can have one or more gateway nodes. Multiple gateway nodes are helpful, if you want to assign different filesets to a different gateway node. The multiple nodes on a gateway node are useful to load balance the workload and also helpful for recovery and upgrade scenarios.

    Gateway nodes need to have access to cloud object storage endpoints. For more information about gateway node recommendation, see General recommendations for AFM gateway node configuration.

  4. Network considerations

    Plan necessary network bandwidth between ¸ and IBM Storage Scale clusters to replicate data and workload.

    The networking element is the major contributor in the overall performance delivery of the IBM Storage Scale cluster where a few nodes act as Network Shared Disk (NSD) servers while most of the client nodes access the file system over a TCP/IP network. Gateway nodes in this host system connect to the outside world, which in this case is Amazon S3 cloud object storageStart of change, Microsoft Azure Blob,End of change or IBM Cloud Object Storage.

    All the firewalls are set up to allow the gateway nodes to discover and communicate with these cloud storage endpoints.

  5. When you work with temporary credentials for accessing Cloud Object Storage such as Amazon S3,Start of change Microsoft Azure Blob,End of change IBM Cloud, and other services, you must obtain the security token and create fileset by specifying --user-keys option of mmafmcosconfig command.

    The mmuid2keys (/var/mmfs/etc/mmuid2keys) file must be configured to provide credentials in the following format as shown:
    Access_Key:Secret_Key:STS
  6. Licensing requirement:

    To use IBM Storage Scale multi-site replication feature, upgrade to advance/DM/ECE edition.