Description of problem ====================== When one tries to use ceph toolbox pod on an external cluster, the ceph tools won't connect to the ceph cluster since cephx keys haven't been configured. See also: BZ 2066906 - ceph command failed with error: RADOS permission error on consumer cluster It seems reasonable to me that ODF should handle setup of authentication (cephx key pairs) for ceph toolbox to work. This could include extensions in cluster importer script, and in ocs operator deploying the toolbox. Version-Release number of selected component ============================================ ODF 4.10 How reproducible ================ 100% Steps to Reproduce ================== 1. Install OCP cluster 2. Install ODF operator and create storage system importing already existing ceph cluster (via OCP Console web ui) 3. Deploy ceph toolbox pod: ``` $ oc patch ocsinitialization ocsinit -n openshift-storage --type json --patch '[{ "op": "replace", "path": "/spec/enableCephTools", "value": true }]' ``` 4. Try to use the toolbox: ``` $ oc get pods -n openshift-storage | grep ceph-tools $ oc rsh -n openshift-storage rook-ceph-tools-foo-bar bash [root@compute-0 /]# ceph -s ``` Actual results ============== ``` [root@compute-0 /]# ceph -s [errno 1] RADOS permission error (error connecting to the cluster) ``` Expected results ================ Ceph tools can authenticate the the cluster out of the box: ``` [root@compute-0 /]# ceph -s cluster: id: 6f73f37c-863f-4851-94ca-ea7fbd5eb044 health: HEALTH_OK services: mon: 3 daemons, quorum a,b,c (age 3d) mgr: a(active, since 3d) mds: rook-ceph-cephfilesystem:1 {0=rook-ceph-cephfilesystem-a=up:active} 1 up:standby-replay osd: 3 osds: 3 up (since 3d), 3 in (since 3d) data: pools: 4 pools, 128 pgs objects: 22 objects, 2.2 KiB usage: 12 GiB used, 285 GiB / 297 GiB avail pgs: 128 active+clean io: client: 767 B/s rd, 1 op/s rd, 0 op/s wr ``` Workaround ========== To use ceph toolbox with external ceph cluster, one have to configure cephx authentication between ceph toolbox pod and the ceph cluster manually. See: - https://bugzilla.redhat.com/show_bug.cgi?id=2066906#c3 - https://docs.ceph.com/en/latest/rados/operations/user-management/
QE team has a workaround in ocs-ci: https://github.com/red-hat-storage/ocs-ci/blob/35853ce0bb2a7efd9469267b3b1b2407eedc82f6/ocs_ci/ocs/utils.py#L775
No RFE in 4.11, will create a Jira for this. Putting NI on myself.
Based on discussion here https://issues.redhat.com/browse/RHSTOR-4512 Closing this BZ