Project

General

Profile

Actions
Copy link

Bug #58514

open

ceph nfs export commands broken with rook orchestrator

Added by Joshua Hoblitt over 1 year ago. Updated 11 months ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
% Done:

0%

Source:
Community (user)
Tags:
Backport:
Regression:
Yes
Severity:
1 - critical
Reviewed:
Affected Versions:
Ceph - v17.2.5
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

The following shell script fragment was working under ceph 16.2.10:

ceph mgr module enable rook
ceph mgr module enable nfs
ceph orch set backend rook

ceph nfs export rm jhome /jhome
ceph nfs export create cephfs jhome /jhome jhome
ceph nfs export rm lsstdata /lsstdata
ceph nfs export create cephfs lsstdata /lsstdata lsstdata
ceph nfs export rm project /project
ceph nfs export create cephfs project /project project
ceph nfs export rm scratch /scratch
ceph nfs export create cephfs scratch /scratch scratch

All ceph nfs export commands seem to be broken under 17.2345 (I did not test .[012]). This result has been replicated on two different k8s clusters.

$ ceph version
ceph version 17.2.3 (dff484dfc9e19a9819f375586300b3b79d80034d) quincy (stable)
$ ceph nfs export create cephfs jhome /jhome jhome
Error ENOENT: Cluster does not exist
command terminated with exit code 2
$ ceph nfs export rm jhome /jhome
Error ENOENT: Cluster does not exist
command terminated with exit code 2

Additionally, since the update to 17.2.x, the dashboard tab for "NFS" now shows that there are no NFS exports (it was populated under 16.2.10).

Actions
Copy link
 #2

Updated by Deepika Upadhyay 11 months ago 

bash-4.4$ ceph mgr module enable nfs
module 'nfs' is already enabled
bash-4.4$ ceph orch set backend rook
bash-4.4$ ceph mgr module enable rook
module 'rook' is already enabled
bash-4.4$ ceph mgr module enable nfs
module 'nfs' is already enabled
bash-4.4$ ceph orch set backend rook^C
bash-4.4$ ceph nfs cluster create nfs-cephfs
NFS Cluster Created Successfully
bash-4.4$ ceph fs subvolume create filesystem primary-subvolume --size 1024
Error ENOENT: FS 'filesystem' not found
bash-4.4$ ceph fs subvolume create my-fs primary-subvolume --size 1024
Error ENOENT: FS 'my-fs' not found
bash-4.4$ ceph fs ls
name: myfs, metadata pool: myfs-metadata, data pools: [myfs-replicated ]
bash-4.4$ ceph fs subvolume create myfs primary-subvolume --size 1024
bash-4.4$ ceph fs subvolume getpath myfs primary-subvolume
/volumes/_nogroup/primary-subvolume/2dafd538-fe5e-4618-a51f-496ae1c5da0e
bash-4.4$ ceph nfs export create cephfs --cluster-id nfs-cephfs --pseudo-path /ceph/secondary --fsname myfs --path=/volumes/_nogroup/primary-subvolume/2dafd538-fe5e-4618-a51f-496ae1c5da0e
{
    "bind": "/ceph/secondary",
    "fs": "myfs",
    "path": "/volumes/_nogroup/primary-subvolume/2dafd538-fe5e-4618-a51f-496ae1c5da0e",
    "cluster": "nfs-cephfs",
    "mode": "RW" 
}
bash-4.4$ ceph nfs export info nfs-cephfs /ceph/secondary
{
  "export_id": 1,
  "path": "/volumes/_nogroup/primary-subvolume/2dafd538-fe5e-4618-a51f-496ae1c5da0e",
  "cluster_id": "nfs-cephfs",
  "pseudo": "/ceph/secondary",
  "access_type": "RW",
  "squash": "none",
  "security_label": true,
  "protocols": [
    4
  ],
  "transports": [
    "TCP" 
  ],
  "fsal": {
    "name": "CEPH",
    "user_id": "nfs.nfs-cephfs.1",
    "fs_name": "myfs" 
  },
  "clients": []
}

I am not sure, but this alternative way works, is this how it should be working any PR that did syntax change?

Actions
Copy link

Also available in: Atom PDF