It may seem a bit confusing, but where radosgw runs via systemd rather than upstart - the client id is actually rgw.<hostname>. The charm provides a convenience of mapping client.radosgw.gateway (legacy client id) into the new client id mapping transparently. Thus, when specifying client.radosgw.gateway it will be mapped the the client.rgw.<hostname> section itself.
As such, this setting applies properly to the ceph.conf. You can audit this by using the admin socket like so:
# The current values of the rgw service
root@juju-504a16-zaza-fc8300987aa2-7:/etc/ceph# ceph daemon /var/run/ceph/ceph-client.rgw.juju-504a16-zaza-fc8300987aa2-7.376436.93842140922864.asok config get rgw_relaxed_s3_bucket_names
{
"rgw_relaxed_s3_bucket_names": "true"
}
Where the admin socket name will actually vary, but will be in the /var/run/ceph directory.
It may seem a bit confusing, but where radosgw runs via systemd rather than upstart - the client id is actually rgw.<hostname>. The charm provides a convenience of mapping client. radosgw. gateway (legacy client id) into the new client id mapping transparently. Thus, when specifying client. radosgw. gateway it will be mapped the the client. rgw.<hostname> section itself.
As such, this setting applies properly to the ceph.conf. You can audit this by using the admin socket like so:
# The current values of the rgw service 504a16- zaza-fc8300987a a2-7:/etc/ ceph# ceph daemon /var/run/ ceph/ceph- client. rgw.juju- 504a16- zaza-fc8300987a a2-7.376436. 93842140922864. asok config get rgw_relaxed_ s3_bucket_ names relaxed_ s3_bucket_ names": "true"
root@juju-
{
"rgw_
}
Where the admin socket name will actually vary, but will be in the /var/run/ceph directory.