Skip to main content
RunBook Academy

← All labs in Ceph

Lab · advanced · ~90 min

Lab 10: OSD replacement

B · Nested virtualisationA · Physical hardware

Objectives

  • Simulate an OSD failure (stop the daemon)
  • Mark the OSD out
  • Wait for backfill
  • Replace the disk
  • Recreate the OSD
  • Validate recovery

Prerequisites

  • Lab 1 complete

Scenario

A disposable Ceph cluster is required. The lab is one exercise in a continuous series of ceph labs that build up to the capstone.

Topology

flowchart TD
 A[Operator workstation] --> B[cephadm]
 B --> C[Ceph cluster
+ MON/MGR/OSDs]
 C --> D[Application]
 D --> E[Backup target]

Requirements

  • Hardware: 90 minute session; 4+ nested VMs or bare-metal hosts.
  • Software: cephadm-compatible Linux (Ubuntu 24.04, Debian 12, Rocky 9).
  • Network access: bond or LACP; cluster and public networks.
  • Time / DNS: configured and verified.

Tasks

The student executes the lab procedure end to end, recording evidence.

Validation

The student runs the validation commands and confirms the expected outcome.

Expected Outcome

The student has demonstrated the competency in a disposable cluster.

Troubleshooting

Common lab failures (network, time, podman, cephx) and how to recover are documented in the upstream Ceph documentation.

Cleanup

# Tear down the cluster
cephadm bootstrap --uninstall

Production notes

Where this lab differs from a permanent install (time, hardware fidelity, capacity) is documented.

What You Learned

The lab maps to one or more of the published learning outcomes for the Ceph course.

Deliverables

  • · Replaced OSD
  • · Logs of recovery

Verification status

Last reviewed
2026-08-17
Executed end to end
2026-08-17