Cloud

AWS storage cluster

IBM Storage Scale container native with Red Hat OpenShift Service on AWS (ROSA) provides a highly scalable and performant solution for containerized applications. You can use the IBM Storage Scale Storage cluster on public clouds (such as AWS) to store its persistent data. IBM Storage Scale storage clusters on public clouds can be deployed and managed with the cloudkit CLI. For more information, see Cloudkit in IBM Storage Scale documentation.

cloudkit is included in the IBM Storage Scale Self-Extracting (SE) package and enables users to deploy IBM Storage Scale clusters across multiple cloud providers (currently limited to AWS), providing deployment topology flexibility. cloudkit helps:

• Reduce the complexity and time that is needed to set up the storage cluster.
• Offers the ability to configure various cluster parameters, such as disk type, instance type, and network configurations.
• Offers the ability to integrate with other IBM Storage Scale features, such as data replication, data tiering, and data encryption.

Considerations

Consider the list:

• Understand the requirements in terms of performance, scalability, data availability, and data protection.
• Consider the structure of your storage, including the number of storage nodes, the file system size, and the number of replicas you want to maintain.
• ROSA and cloudkit both use Multi-Availability Zone (AZ) to ensure resiliency, but in Multi-AZ mode, the inter-AZ network limit constrains the throughput.
• Review the AWS quota limits to ensure that your existing Virtual Private Cloud (VPC) has enough free IP addresses to accommodate both ROSA and cloudkit. If you are planning to create a new VPC, you should plan the network to accommodate the IP requirements for both services.
• Users have the flexibility to start with either ROSA in new VPC mode and cloudkit in existing VPC mode (or) cloudkit in new VPC mode and ROSA in existing VPC mode, depending on the requirements.
• For optimal performance, choose a storage cluster deployment within the same subnet that is in use by the ROSA, providing low latency and higher throughput.

The following tasks assume that the installation of IBM Storage Scale container native on AWS ROSA is complete.

Deploy storage cluster

Use the cloudkit command to deploy the storage cluster.

cloudkit create cluster

cloudkit provides various deployment profiles, including Throughput-Performance-Scratch-Storage, Throughput-Performance-Persistent-Storage, Throughput-Advance-Persistent-Storage, and Balanced.

If you are deploying to a Single AZ, it is advisable to use the Throughput-Performance-Persistent-Storage profile. For multi-AZ, use the Balanced profile.

Do not use the Throughput-Performance-Scratch-Storage profile for IBM Storage Scale container native workloads. This profile uses instance storage and results in data loss when storage nodes are shut down.

Details of the cluster can be viewed by using the describe option:

cloudkit describe cluster

Perform grant filesystem operation

cloudkit performs setting up security rules for data access and exchange between the ROSA worker security group and storage cluster. It performs the prerequisite configurations such as creating users in the ContainerOperator and CsiAdmin groups on the storage cluster.

Use the cloudkit command to simplify the data management across multiple IBM Storage Scale container native clusters.

cloudkit grant filesystem

Details of the file system access granted by cloudkit can be viewed by using the describe option:

cloudkit describe grant

Perform revoke filesystem operation

If the IBM Storage Scale container native is uninstalled, revoke the access to the file system by using the following command:

cloudkit revoke filesystem

Delete storage cluster

The following action cannot be undone and will destroy all scale instances and wipe the disks that are used for the file system. Take the appropriate data backups before proceeding, failure to do so will result in permanent loss of data.

Use the following cloudkit command to delete the storage cluster:

cloudkit delete cluster