Ceph Level 1 Foundations Training Course

Ceph Level 1 – Foundations

Ceph is an open-source distributed storage platform created to deliver scalable, software-defined block, object and file storage on commodity hardware. Originally developed by Sage Weil and now stewarded through the wider upstream Ceph community with strong enterprise backing, it was designed around goals such as resilience, horizontal scale, self-healing behaviour and removal of traditional storage bottlenecks through intelligent data distribution.

This module introduces Ceph as a distributed storage platform and explains the architectural concepts learners need before they can operate or troubleshoot it properly, giving them the right foundation for understanding how data moves through the cluster and why Ceph is widely used in scale-out infrastructure and Proxmox-backed storage environments.

Course purpose

Give learners a solid grounding in how Ceph works as distributed storage, so they can understand the platform architecture, follow the path of data through the cluster, and build the right mental model before moving into operations or troubleshooting.

Suggested duration

1 day
or 1.5 days with labs

Target audience

storage administrators
platform engineers
Proxmox administrators using Ceph-backed storage
engineers new to distributed storage

Prerequisites

Linux command-line basics
basic storage concepts
general networking knowledge

Learning outcomes

explain what Ceph is and where it fits
describe how Ceph stores and distributes data
understand the relationship between RADOS, pools, PGs and CRUSH
identify the main Ceph daemon roles
explain the basic flow of data through the cluster

Detailed module structure

Unit 1: What Ceph is and where it fits

Topics:

Ceph as distributed, software-defined storage
block, object and file interfaces
common deployment scenarios
why Ceph is attractive for scale-out infrastructure
trade-offs: complexity, consistency, recovery overhead, operational discipline

Unit 2: Core architecture and data flow

Topics:

how clients interact with the cluster
object-based internal design
introduction to RADOS
how data is broken up, placed and replicated
why Ceph avoids traditional central metadata bottlenecks for object placement

Unit 3: RADOS and object storage concepts

Topics:

what RADOS provides
objects and object collections
logical vs physical data placement
how higher-level services build on top of RADOS

Unit 4: Pools and placement groups

Topics:

what pools are for
logical separation of data
how placement groups act as distribution units
why PG sizing matters
operational impact of too few or too many PGs

Lab ideas:

inspect cluster pool configuration
view PG distribution at a high level

Unit 5: CRUSH and data placement

Topics:

what CRUSH is
deterministic data placement
failure domains
topology-aware placement
why CRUSH removes the need for lookup tables in the traditional sense

Lab ideas:

inspect a CRUSH hierarchy
discuss how host/rack failure domains affect resilience

Unit 6: Ceph daemons and their roles

Topics:

MON
MGR
OSD
MDS
optionally RGW if relevant to your environment

Standard naming: MON = monitor, MGR = manager, OSD = object storage daemon, MDS = metadata server.

Lab ideas:

identify daemon roles in a live or lab cluster
map services to their operational responsibilities

Assessment

Short concept check

explain pool vs PG
explain CRUSH at a high level
identify daemon roles

Architecture walkthrough

“What happens when a client writes data to Ceph?”