Overview

emptyDir Volumes

Some applications require additional storage, but whether the data is still available after a restart is not important. For example, though limited by memory, cache services can move infrequently-used data to the storage with slower access than memory. As a result, the overall performance is not impacted significantly. Other applications require read-only data injected as files, such as configuration data or secrets.

Ephemeral volumes are designed for the above scenario. Because volumes follow the pods' lifetime and get created and deleted along with pods, pods can be stopped and restarted without being limited to where PVs are available.

CCI supports several types of ephemeral volumes for different purposes.

• emptyDir: It is empty at pod startup, with storage coming from the disk or memory of the environment where the pod is running.
• ConfigMap, downwardAPI, and secret: inject data of different types into pods.

By default, CCI allocates 30 GiB of free ephemeral storage space to a pod. The storage space is unavailable when the pod is not in use. If you need to use more ephemeral storage space, you can expand the capacity. For details, see Ephemeral Storage.

General Purpose File System (SFS 3.0) Volumes

CCI allows you to mount a volume that is created from a general-purpose file system to a container for persistent data storage. General Purpose File System (formerly SFS 3.0) volumes are commonly used in ReadWriteMany scenarios for large-capacity expansion and cost-sensitive services, such as media processing, content management, big data analysis, and workload analysis. For details, see General Purpose File System (SFS 3.0) Volumes.

SFS Turbo Volumes

CCI allows you to mount SFS Turbo volumes to specific container paths. SFS Turbo file systems are fast, on-demand, and scalable. They are suitable for DevOps, containerized microservices, and enterprise office applications. For details, see SFS Turbo Volumes.

OBS Parallel File System Volumes

CCI allows you to create volumes from OBS parallel file systems. Parallel file systems are optimized object-based file systems. They are designed for big data scenarios where OBS is used as the unified data lake storage. It features access latency in milliseconds, TB/s-level bandwidth, millions of IOPS, and high compatibility, performance, scalability, and reliability. For details, see OBS Parallel File System Volumes.