Ceph clusters provide distributed storage with high availability and scalability. This page allows you to create, configure, bootstrap, and manage your Ceph storage clusters.Documentation Index
Fetch the complete documentation index at: https://docs.shiftlabs.dev/llms.txt
Use this file to discover all available pages before exploring further.
Key Concepts
Cluster
A Ceph cluster is a collection of nodes running Ceph daemons that work together to provide storage services.
Bootstrap
The process of initializing a cluster by deploying Ceph components to nodes and creating the initial configuration.
Public Network
The network used for client-to-cluster communication and inter-daemon traffic.
Cluster Network
Optional dedicated network for OSD replication traffic to improve performance.
Required Permissions
| Action | Permission |
|---|---|
| View Clusters | iam:project:infrastructure:ceph:read |
| Create Cluster | iam:project:infrastructure:ceph:write |
| Edit Cluster | iam:project:infrastructure:ceph:write |
| Bootstrap Cluster | iam:project:infrastructure:ceph:write |
| Reset Cluster | iam:project:infrastructure:ceph:write |
| Delete Cluster | iam:project:infrastructure:ceph:delete |
Cluster Status
| Status | Description |
|---|---|
| Ready | Cluster is bootstrapped and operational |
| Not Ready | Cluster is registered but not yet bootstrapped |
| Active | Cluster is enabled for operations |
| Disabled | Cluster is disabled (no operations allowed) |
How to View Clusters
View Cluster List
The list shows all registered Ceph clusters with their status, version, and network configuration.
Filter and Search
Use the search box to find clusters by name, version, or network. Filter by status (Active/Disabled).
How to Create a Cluster
Configure Basic Settings
Enter the cluster configuration:
- Cluster Name: Unique identifier for the cluster
- Ceph Version: Select the Ceph release version
- Status: Enable or disable the cluster
Configure Networks
Set up the network configuration:
- Public Network: CIDR for client and monitor communication (required)
- Cluster Network: CIDR for OSD replication (optional, improves performance)
Configure Dashboard (Optional)
Enable and configure the Ceph Dashboard:
- Enable Dashboard: Toggle to enable the web UI
- Dashboard User: Initial admin username
- Dashboard Password: Initial admin password
Creating a cluster only registers it in the system. You must bootstrap the cluster to deploy Ceph components and make it operational.
How to View Cluster Details
Review Dashboard
The detail page shows a comprehensive dashboard with:
- Cluster Status: Overall health and status
- Health Alerts: Active warnings and errors
- Operations: Bootstrap and reset actions
- Upgrade Status: Version and upgrade information
- Storage Capacity: Used vs available storage
- Cluster Config: Network and version settings
- OSD Status: Object Storage Daemon health
- Monitor Status: Monitor daemon quorum
- PG Status: Placement Group health
- Node Overview: All nodes in the cluster
How to Bootstrap a Cluster
Bootstrapping initializes a new cluster by deploying Ceph components to the configured nodes.Configure Bootstrap Options
Review and configure bootstrap settings in the wizard:
- Select nodes to include
- Configure initial OSD deployment
- Set replication settings
Start Bootstrap
Click Start Bootstrap to begin the process. Progress is shown in the operation logs panel.
How to Reset a Cluster
Resetting removes all Ceph data and configuration from the cluster nodes, returning them to a clean state.How to Edit a Cluster
Modify Settings
Update the cluster configuration as needed:
- Cluster name
- Ceph version
- Network settings
- Dashboard configuration
- Active/disabled status
Some settings may require a cluster reset and re-bootstrap to take effect, such as network changes.
How to Delete a Cluster
Cluster Configuration Fields
| Field | Description |
|---|---|
| Cluster Name | Unique identifier for the cluster |
| Ceph Version | Release version (e.g., Reef, Quincy) |
| Public Network | CIDR for client/monitor traffic (e.g., 10.0.0.0/24) |
| Cluster Network | Optional CIDR for OSD replication |
| Dashboard | Enable Ceph Dashboard web UI |
| Dashboard User | Initial admin username |
| Dashboard Password | Initial admin password |
| Cluster FSID | Unique cluster identifier (auto-generated) |
| Dashboard URL | URL to access the Ceph Dashboard |
Dashboard Cards
Cluster Status Card
Shows the overall cluster health with:- Health status (HEALTH_OK, HEALTH_WARN, HEALTH_ERR)
- OSD summary (up/in counts)
- Version information
Health Alerts Card
Displays active health checks and warnings from the cluster, such as:- Degraded PGs
- OSD nearfull warnings
- Clock skew
- Slow OSD operations
Storage Capacity Card
Shows storage utilization:- Total capacity
- Used storage
- Available storage
- Usage percentage with visual indicator
OSD Status Card
Displays OSD (Object Storage Daemon) health:- Total OSD count
- Up/Down status
- In/Out status
Monitor Status Card
Shows monitor quorum status:- Total monitors
- Quorum members
- Leader information
PG Status Card
Displays Placement Group health:- Total PGs
- Active+clean count
- Degraded/misplaced counts
Node Overview Card
Shows all nodes in the cluster with:- Hostname
- Role
- Status
- Quick actions
Troubleshooting
Cluster shows 'Not Ready' after creation
Cluster shows 'Not Ready' after creation
- This is expected. New clusters must be bootstrapped to become ready
- Add nodes to the cluster first
- Then bootstrap the cluster to deploy Ceph components
Bootstrap fails
Bootstrap fails
- Check that all nodes are reachable via SSH
- Verify network configuration matches node interfaces
- Ensure sufficient disk space on nodes
- Check operation logs for specific errors
- Verify Ceph version compatibility with nodes
Cluster health shows warnings
Cluster health shows warnings
- Check the Health Alerts card for specific issues
- Common warnings: nearfull OSDs, slow requests, clock skew
- Navigate to specific component pages (OSDs, Monitors) for details
- Some warnings resolve automatically after rebalancing
Cannot connect to Dashboard
Cannot connect to Dashboard
- Verify Dashboard is enabled in cluster settings
- Check Dashboard URL is accessible from your network
- Verify credentials are correct
- Check that the mgr daemon is running
Reset operation stuck
Reset operation stuck
- Some cleanup operations may take time
- Check operation logs for progress
- If stuck, nodes may need manual cleanup
- SSH to nodes and verify Ceph processes are stopped
Network configuration errors
Network configuration errors
- Ensure CIDR format is correct (e.g.,
10.0.0.0/24) - Verify the network is routable between all nodes
- Public network must be accessible by clients
- Cluster network (if used) must be accessible by all OSDs
FAQ
What is the difference between Public and Cluster networks?
What is the difference between Public and Cluster networks?
Public Network carries all client-to-cluster traffic and inter-daemon communication. It’s required for cluster operation.Cluster Network is an optional dedicated network for OSD-to-OSD replication traffic. Using a separate cluster network improves performance by isolating replication traffic from client traffic.
When should I use a separate Cluster Network?
When should I use a separate Cluster Network?
Use a separate cluster network when:
- You have high I/O workloads
- Client traffic is competing with replication
- You want to isolate replication for security
- You have 10GbE+ network infrastructure
Can I change the Ceph version after bootstrap?
Can I change the Ceph version after bootstrap?
Yes, but this requires an upgrade process. Navigate to the cluster detail page and use the Upgrade feature to change versions. Upgrades must follow Ceph’s supported upgrade path.
What happens if I disable a cluster?
What happens if I disable a cluster?
Disabling a cluster (setting isActive to false) prevents new operations from being performed through this interface. The actual Ceph cluster continues running, but management operations are blocked.
How long does bootstrap take?
How long does bootstrap take?
Bootstrap duration depends on:
- Number of nodes
- Network speed between nodes
- Number of OSDs to deploy
- Ceph version being installed
What is the FSID?
What is the FSID?
The FSID (Filesystem ID) is a unique identifier for your Ceph cluster, generated during bootstrap. It’s used internally by Ceph to identify the cluster and should not be changed.
Do I need the Ceph Dashboard?
Do I need the Ceph Dashboard?
The Ceph Dashboard is optional but recommended. It provides:
- Web-based cluster monitoring
- Visual storage management
- Performance graphs
- Native Ceph administration
Can I have multiple clusters?
Can I have multiple clusters?
Yes. You can register and manage multiple Ceph clusters. Each cluster operates independently with its own nodes, storage, and configuration.