2

I'm trying to set up a rook Ceph cluster on my kubernetes cluster.

Topography:

  • 3 kubernetes nodes (all are master/worker pods)
  • Each node has /dev/vdX on it for ceph
  • Each node is intended to work as part of the ceph cluster

I deployed Rook operator with the helm chart with the following values:

image:
  tag: v1.5.8
  pullPolicy: IfNotPresent

resources:
  limits:
    cpu: 500m
    memory: 256Mi
  requests:
    cpu: 100m
    memory: 256Mi
enableFlexDriver: false
enableDiscoveryDaemon: true

After that was fully initialized, I tried creating a cluster with the following manifest.

apiVersion: ceph.rook.io/v1
kind: CephCluster
metadata:
  name: maincluster
  namespace: rook
spec:
  cephVersion:
    image: ceph/ceph:v15.2.9
  dataDirHostPath: /var/lib/rook
  mon:
    count: 3
    allowMultiplePerNode: false
  mgr:
    modules:
    - name: pg_autoscaler
      enabled: false
  dashboard:
    enabled: true
    ssl: true
  monitoring:
    enabled: false
    rulesNamespace: rook
  cleanupPolicy:
    sanitizeDisks:
      method: quick
  storage: # cluster level storage configuration and selection
    useAllNodes: true
    useAllDevices: true

Most things seem to work after running just those. One pod show errors, but it doesn't seem to be related to the current issue I'm having; seems to rather be a symptom (it's the csi-rbdplugin pod).

I can load the dashboard just fine with an ingress I created. However, when, on the dashboard, I try to create an OSD or do much of anything, I get a popup saying

Orchestrator is not available. Please consult the documentation on how to configure and enable the functionality.

The documenation link it has there brings me to this 404 page, so that isn't much help.

I've tried searching elsewhere but there doesn't seem to be much info on that specific error message.

Anyone know what I'm doing wrong that the Orchestrator is not available and I cannot do much of anything?

cclloyd
  • 583
  • 1
  • 13
  • 24
  • The orchestrator module is part of the MGR service. Does your ceph cluster have at least one MGR up and running? I'm not familiar with rook so I wouldn't know how to make sure MGR is started successfully. – eblock Mar 17 '21 at 09:05
  • I have the same problem, what was your solution? – dfsg76 Apr 11 '21 at 20:06

1 Answers1

2

I was able to get rid of this warning and enable orchestrator.

First of all install Rook Toolbox by following this howto

Then exec into toolbox: kubectl -n rook-ceph exec -it rook-ceph-tools-78cdfd976c-sclh9 -- /bin/bash

And once there, run following commands:

[root@rook-ceph-tools-78cdfd976c-sclh9 /]# ceph mgr module enable rook
[root@rook-ceph-tools-78cdfd976c-sclh9 /]# ceph orch set backend rook
[root@rook-ceph-tools-78cdfd976c-sclh9 /]# ceph orch status
Backend: rook
Available: True

ceph orch status should show the output as in above example.

After that I have no more warnings in Rook Dashboard.

Sebastian Vaisov
  • 21
  • 2
  • Nice. May I ask how did you got the dashboard up. I am facing some issue. https://github.com/rook/rook/issues/10532 – Sanjeev Jun 30 '22 at 15:11