Skip to content

Ceph

Ceph is a distributed storage system that provides highly scalable and reliable storage for large amounts of data. It is designed to be self-managing, self-healing, and self-optimizing, making it ideal for cloud storage environments and data centers.

Ceph Logo

Main Features

  • Scalability: Ceph can scale from a few nodes to thousands of nodes, allowing for seamless growth.
  • Reliability: It uses replication and erasure coding to ensure data integrity.
  • Self-management: Ceph self-repairs and self-optimizes, reducing the need for manual intervention.
  • Flexibility: Supports multiple storage interfaces, including block, object, and file systems.

Ceph Architecture

Ceph consists of several key components:

  • Ceph Monitors (MON): Maintain a map of the cluster and ensure data consistency.
  • Ceph OSD Daemons (OSD): Store data and handle replication and recovery operations.
  • Ceph Manager Daemons (MGR): Provide additional functionalities such as monitoring and cluster management.
  • Ceph Metadata Servers (MDS): Manage metadata for the CephFS file system.

Ceph Architecture

Use Cases

  • Cloud Storage: Ceph is ideal for cloud service providers needing scalable and reliable storage.
  • Big Data: Ceph can handle large volumes of data, making it suitable for Big Data applications.
  • Backup and Recovery: Ceph's replication and erasure coding ensure that data is always available and protected.

Basic Installation with cephadm (Reef Version)

To install Ceph Reef version using cephadm, you can follow these basic steps:

  1. Prepare the nodes: Ensure all nodes have the necessary dependencies installed and have internet access.
  2. Install cephadm: Download and install cephadm on the initial node. 
    curl --silent --remote-name https://raw.githubusercontent.com/ceph/ceph/reef/src/cephadm/cephadm
chmod +x cephadm
sudo ./cephadm install
  3. Deploy the cluster: Use cephadm to deploy the cluster. 
    sudo cephadm bootstrap --mon-ip <INITIAL_NODE_IP>
  4. Add additional nodes: Add more nodes to the cluster. 
    sudo ceph orch host add <NODE_NAME> <NODE_IP>
  5. Configure the cluster: Configure the monitors, OSDs, and other necessary components using cephadm. 
    sudo ceph orch apply osd --all-available-devices
  6. Verify the installation: Ensure the cluster is functioning correctly. 
    ceph -s

For more details, you can refer to the official Ceph documentation.