About Storage
This page explains how storage works with your resources and ACTIVATE.
Note
We make a distinction between the storage you create on the ACTIVATE platform and the storage you create outside the ACTIVATE platform. The information on this page and our pages for creating, configuring, attaching, and sharing storage apply to storage resources made on the platform.
If you would rather create storage outside ACTIVATE, or if you have existing storage resources that you'd like to work with, please see the following:
Our section Transferring Data contains instructions for AWS, Azure, and Google.
Storage Types
When creating storage resources, you’ll choose whether your storage is ephemeral or persistent.
Ephemeral storage is created and destroyed with a cluster. This type of storage is provisioned when you start the cluster it’s attached to, and it’s destroyed when you stop the cluster. Ephemeral storage resources do not have a power button, and they won't appear in the Storage Resources module on the the Compute page.
Persistent storage is created and destroyed independently. If you want to destroy a persistent storage resource, you can do so from the Storage page by clicking the power button.
Persistent storage can be attached to multiple clusters.
At this time, only Lustre storage options can be ephemeral.
Storage Provisioning Logs
For persistent storage resources, you can see the creation and deletion logs on the resource's Events tab. If a persistent storage resource is active, you’ll see all the clusters that the storage is attached to on its Details tab. For more information, please see Details in Configuring Storage.
Ephemeral storage resources do not have a page of their own. Instead, you can see their creation logs on the cluster they're attached to by clicking the cluster's status when it's provisioning.
Buckets
Typically, users would have to use a cloud service provider’s CLI to upload or retrieve data from a bucket. With our new filesystem feature, buckets are instead mounted as a filesystem; this way, you can interact with a bucket as if it’s a directory rather than using CLI commands.