- Getting Started Configuration Accessing Your Cluster
- Accessing your Cluster
- CephFS
- MDS Notifications
- CephFS Guard
- RBD
- S3
- S3 Multisite
- NVMe over Fabrics
- SMB
- NFS
- Ceph Keys
- HA Groups
- Updating croit
- Replacing OSDs
- Replacing NIC
- PG numbers
- Updating Ceph
- Migrating or restoring croit container
On this Page:
Introduction
croit will help you deploy a scale-out Ceph cluster within minutes. Nodes run an immutable, croit-served OS image, which is most commonly delivered by PXE booting them over the network.
PXE is not the only option: you can also install croit directly to a node's local disk and have it boot from there, with or without ever PXE booting it. See Local Disk Boot for the details.
Design Principles
- Immutable: The OS images are immutable and vetted by croit.
- Sensible defaults: Ships with known sensible and stable defaults.
- No vendor lock-in: Runs on commodity hardware. Possible to migrate to a self-managed Ceph cluster, without losing or moving any data.
- User Experience: If it's harder to use than alternatives, it's a bug.
Drawbacks to PXE Booting
- Initial network setup might be time-consuming.
- In case of a power outage, you might have to wait longer for your cluster nodes to come back.
- Immutability is a double-edged sword. While making management easier, you can't install packages and expect them to persist. We provide Hook Scripts to work around this.
The first two drawbacks only apply while a node boots over PXE. Nodes that boot from a local disk come back up on their own after a power outage and don't depend on the boot network. Immutability applies either way. See Local Disk Boot.
For the installation instructions see Getting Started.
