Manager status and modules

Intermediate⏱ ~15 minceph

What you'll learn

Read manager status output
Manage manager modules
Diagnose a failed module
Verify manager services are reachable

Prerequisites

None — start here.

Verified against Ceph Tentacle 20.2.x · Ceph Squid 19.2.x (supported previous) · cephadm matches the verified Ceph release · podman 4.x · csi-rbd and csi-cephfs current · RBD / CephFS / RGW current (matches Ceph release) · Linux kernel 5.15+ (5.10 minimum) · Ubuntu 24.04 LTS (Ceph host baseline) · Debian 12 (Bookworm) (Ceph host baseline) · Rocky Linux / RHEL / AlmaLinux 9.x (Ceph host baseline) · Proxmox VE 9.x (cross-course integration) · Kubernetes 1.31+ (cross-course integration) · 2026-08-18

Why this matters in production

Manager modules provide the dashboard, the metrics, the orchestrator, and the balancer. When one fails, the symptom is the loss of that capability rather than anything Ceph reports as a storage problem.

Status

ceph mgr stat
ceph mgr dump --format json | jq '{active_name, active_addr, standbys: [.standbys[].name]}'

{
  "active_name": "ceph-mon-01.abcdef",
  "active_addr": "10.20.0.10:6800/1234",
  "standbys": ["ceph-mon-02.ghijkl"]
}

Modules

ceph mgr module ls --format json | \
  jq -r '{enabled: .enabled_modules, always: .always_on_modules}'

ceph mgr module enable prometheus
ceph mgr module disable dashboard
ceph mgr services

{
  "dashboard": "https://10.20.0.10:8443/",
  "prometheus": "http://10.20.0.10:9283/"
}

ceph mgr services lists the endpoints modules are serving, which is what to check when a dashboard or metrics endpoint is unreachable.

Common modules

Module	Provides
`cephadm`	the orchestrator
`dashboard`	the web interface
`prometheus`	the metrics endpoint
`balancer`	automatic PG distribution
`pg_autoscaler`	automatic pg_num management
`devicehealth`	SMART collection and failure prediction
`crash`	crash report collection
`restful`	the REST API
`telemetry`	optional upstream reporting

Diagnosing a failed module

ID=12
ceph health detail | grep -A5 MGR_MODULE_ERROR
ceph crash ls
ceph crash info ${ID}

ceph config set mgr mgr/dashboard/log_level debug
ceph tell mgr.${ID} mgr_status
journalctl -u ceph-mgr@${ID} --since '30 min ago'

A module raising an exception is reported through MGR_MODULE_ERROR and recorded in the crash log, so ceph crash ls frequently identifies it directly.

Failover

ceph mgr fail
ceph mgr stat
ceph mgr services

Failing over the manager restarts every module on the new active daemon, which resolves a module in a bad state without restarting the process.

Quiz

Knowledge check · 4 questions

Q1. The dashboard is unreachable while the orchestrator and metrics continue working. What has happened?

The dashboard module has failed while the others continue — module exceptions are isolated The active manager has failed and the standby took over without the dashboard module enabled The monitors have lost quorum, so the dashboard cannot authenticate while cached metrics continue The manager host is down and the dashboard endpoint was bound to that host's address only
Q2. `ceph mgr fail` is a reasonable first response to a misbehaving manager module.

True False
Q3. Investigate missing metrics.

Prometheus scraping of the Ceph metrics endpoint has been failing for an hour. The cluster reports HEALTH_WARN with MGR_MODULE_ERROR. The dashboard and orchestrator work normally.
Q4. Why are manager module exceptions caught rather than allowed to terminate the process?

Passing score: 75%. Answers are checked in this browser.

Production discipline

Try ceph mgr fail before investigating a misbehaving module; it restarts every module on the standby and resolves transient module state without a daemon restart. Check ceph mgr services when an endpoint is unreachable — it reports what the modules are actually serving.

Cross-course references

Kubernetes: a failing controller within the controller-manager degrades one capability similarly
Linux: plugin isolation within a host process is a recurring design and a recurring diagnostic