Cannot retrieve config for rgw

[simonpie]

If I try to retrieve the rgw configuration ( including defaults ) using the commande ceph config show-with-defaults for the rgw service, it always end in error.

I start by running ceph auth list to retrieve the name of the daemon.

[root@lablmcp01a ~]# ceph auth list | grep -i rados -C 2
	key: AQCUBpJosfBkGhAASUSdiDyhp5acLCcLziDAKQ==
	caps: [mon] allow profile bootstrap-rgw
client.radosgw.gateway
	key: AQCrBpJoAV+3OhAALohJxHzpvjFudI3RwbsnCA==
	caps: [mon] allow rw
[root@lablmcp01a ~]# 

Then

[root@lablmcp01a ~]# ceph config show-with-defaults client.radosgw.gateway
Error ENOENT: no config state for daemon client.radosgw.gateway
[root@lablmcp01a ~]# 

Note that it works with daemons like osd.1 or the manager.

[root@lablmcp01a ~]# ceph auth list | grep -i "mgr."
	caps: [mgr] allow profile mds
	caps: [mgr] allow profile osd
	caps: [mgr] allow profile osd
	caps: [mgr] allow profile osd
	caps: [mgr] allow *
mgr.lablmcp01a
[root@lablmcp01a ~]# ceph config show-with-defaults mgr.lablmcp01a | grep -i api
restapi_base_url                                                                                                                                                                                                                                                                                                                                                                                                                                     default                                  
restapi_log_level                                                                                                                                                                                                                                                                                                                                                                                                                                    default                                  
rgw_enable_apis                                             s3, s3website, swift, swift_auth, admin, sts, iam, notifications                                                                                                                                                                                                                                                                                                                         default                                  
rgw_keystone_api_version                                    2                                                                                                                                                                                                                                                                                                                                                                                        default                                  
[root@lablmcp01a ~]# 

And the same procedure for rgw works for our two physical IBM ceph clusters we have. All tree are in quincy.

For exemple, from the management VM for the IBM ceph lab we have :

[root@lablcpmgmt01a ~]# ceph auth list | grep -i "rgw."
client.rgw.lablcpobj01a.lablcpobj01a.lwvfuk
	caps: [osd] allow rwx tag rgw *=*
client.rgw.lablcpobj01a.rgw0
client.rgw.lablcpobj01b.lablcpobj01b.hqniml
	caps: [osd] allow rwx tag rgw *=*
client.rgw.lablcpobj01b.rgw0
[root@lablcpmgmt01a ~]# ceph config show-with-defaults ceph-client.rgw.lablcpobj01a.lablcpobj01a.lwvfuk | grep -i api
restapi_base_url                                                                                                                                                                                                                                                                                                                                                                                                                                     default                     
restapi_log_level                                                                                                                                                                                                                                                                                                                                                                                                                                    default                     
rgw_enable_apis                                             s3, s3website, swift, swift_auth, admin, sts, iam, notifications                                                                                                                                                                                                                                                                                                                         default                     
rgw_keystone_api_version                                    2                                                                                                                                                                                                                                                                                                                                                                                        default                     
[root@lablcpmgmt01a ~]# 

[syncronize-issues-to-jira[bot]]

Thank you for reporting your feedback to us!

The internal ticket has been created: https://warthogs.atlassian.net/browse/CEPH-1418.

This message was autogenerated

[UtkarshBhatthere]

@sabaini can this be caused due to the use of gateway key rather than using per host key ?