Preparing the installer node

The installer node refers to the node where the cloudkit is being run.

The installer node can be of one of the following types:

On-premise host: A physical host or virtual machine that is located on the premises. A reliable and secure network connection must exist between the installer node and the cloud environment. The cloudkit uses this connection to communicate with the cloud infrastructure to provision and configure the required resources.
Cloud VM: A virtual machine that is located on the cloud.

The cloudkit execution is currently only supported on Red Hat Enterprise Linux (RHEL). Therefore, the installer node must run any one of these operating systems. For information about the supported RHEL versions, see Table 2.

It is essential that this machine has access to the base Red Hat repositories.

Note: The installer node stores critical information regarding the IBM Storage Scale cluster on the cloud, therefore, it is crucial to ensure its security. In addition, this critical information that is stored in the user’s home directory (

/<home
directory>/scale-cloudkit/data

) must be backed up to ensure that it is not lost.

Extracting the cloudkit binary

The cloudkit rpm binary is bundled as a part of IBM Storage Scale self-extraction (SE) package. For more information, see Extracting the IBM Storage Scale software on Linux nodes.

By default, the cloudkit rpm is extracted from the SE package into the /usr/lpp/mmfs/<release_version>/cloudkit_rpms directory.

Installation of cloudkit package for Red Hat Enterprise Linux (RHEL)

To install IBM Storage Scale cloudkit packages, change the directory to where the installation package is extracted. For example, run this command:

cd /usr/lpp/mmfs/<x.x.x.x>/cloudkit_rpms

Then, run the following command:

yum -y install gpfs.cloud-deployer*.rpm

Now, the cloudkit is extracted into the /usr/lpp/mmfs/<release_version>/cloudkit directory.

Note: The cloudkit rpm was introduced with IBM Storage Scale 6.0.0 release. For versions earlier than IBM Storage Scale 6.0.0, the cloudkit binary is directly extracted from the SE package and placed in the /usr/lpp/mmfs/<release_version>/cloudkit directory.

Preparing the cloudkit environment file

The cloudkit encrypts all data that is stored inside it by using a key that is stored in a passphrase file (cloudkit_config.ini), which is provided by the user; the owner of the passphrase file is granted read and write permission. The cloudkit_config.ini file must be created in the installer node and follow the next format:

Note: Ensure that this passphrase file is backed-up to a secure location.

[default]
cloudkit_key = <PassPhrase>

During the cloudkit initialization, by default, feature_gate = Stable gets set. This is recommended for production clusters and lists only the generally available features. To enable technical preview features on test clusters, you need to change this setting to: feature_gate = TechPreviewNoUpgrade, and then run the cloudkit init command.

[default]
cloudkit_key = <PassPhrase>
feature_gate = TechPreviewNoUpgrade

Warning: Enabling the TechPreviewNoUpgrade feature set on your cluster prevents future upgrades and it cannot be undone. You should not enable this feature set on production clusters.

Example:

In this example, the passphrase file is created at /usr/lpp/mmfs/<release_version>/cloudkit/cloudkit_config.ini with the following content:

# cat .env
[default]
cloudkit_key = XXXXXXXX

This passphrase file path is required during cloudkit init:

To export, use export env file, with the syntax export CLOUDKIT_PASSPHRASE_FILE_PATH<ENV_FILE_PATH>:
```
export CLOUDKIT_PASSPHRASE_FILE_PATH=/usr/lpp/mmfs/<release_version>/cloudkit/cloudkit_config.ini
```

To configure, run the cloudkit init command:

# ./cloudkit init
I: Logging at /root/scale-cloudkit/logs/cloudkit-06-12-2024_0-2-45.log
? Passphrase file path for encrypting DB contents:  /usr/lpp/mmfs/6.0.1.0/cloudkit/cloudkit_config.ini