Ceph broker request incomplete

Hi Team,

On every deployment (OpenStack charm base HA, Yoga version, Ubuntu 20.04) I am facing the below issue on glance and cinder. But I can able to create volumes and able to upload images to glance. How to fix this?

“Ceph broker request incomplete”

Juju status output:

Unit                                  Workload  Agent  Machine   Public address  Ports               Message
ceph-mon/0                            active    idle   0/lxd/0   10.93.244.57                        Unit is ready and clustered
ceph-mon/1                            active    idle   0/lxd/1   10.93.244.58                        Unit is ready and clustered
ceph-mon/2*                           active    idle   0/lxd/2   10.93.244.56                        Unit is ready and clustered
ceph-osd/0*                           active    idle   0         10.93.241.129                       Unit is ready (17 OSD)
cinder/0                              active    idle   1/lxd/0   10.93.242.62    8776/tcp            Unit is ready
  cinder-ceph/1                       active    idle             10.93.242.62                        Unit is ready
  cinder-hacluster/1                  active    idle             10.93.242.62                        Unit is ready and clustered
  cinder-mysql-router/1               active    idle             10.93.242.62                        Unit is ready
cinder/1                              active    idle   2/lxd/0   10.93.242.82    8776/tcp            Unit is ready
  cinder-ceph/2                       waiting   idle             10.93.242.82                        Ceph broker request incomplete
  cinder-hacluster/2                  active    idle             10.93.242.82                        Unit is ready and clustered
  cinder-mysql-router/2               active    idle             10.93.242.82                        Unit is ready
cinder/2*                             active    idle   3/lxd/0   10.93.242.60    8776/tcp            Unit is ready
  cinder-ceph/0*                      active    idle             10.93.242.60                        Unit is ready
  cinder-hacluster/0*                 active    idle             10.93.242.60                        Unit is ready and clustered
  cinder-mysql-router/0*              active    idle             10.93.242.60                        Unit is ready
glance/0                              waiting   idle   1/lxd/1   10.93.242.64    9292/tcp            Ceph broker request incomplete
  glance-hacluster/1                  active    idle             10.93.242.64                        Unit is ready and clustered
  glance-mysql-router/1               active    idle             10.93.242.64                        Unit is ready
glance/1                              active    idle   2/lxd/1   10.93.242.66    9292/tcp            Unit is ready
  glance-hacluster/0*                 active    idle             10.93.242.66                        Unit is ready and clustered
  glance-mysql-router/0*              active    idle             10.93.242.66                        Unit is ready
glance/2*                             active    idle   3/lxd/1   10.93.242.61    9292/tcp            Unit is ready
  glance-hacluster/2                  active    idle             10.93.242.61                        Unit is ready and clustered
  glance-mysql-router/2               active    idle             10.93.242.61                        Unit is ready

Hi Rajasekar

Looking at the status output you included it would appear that you only have a single unit of the ceph-osd application; Ceph uses replication to ensure data resilience and has a default of replication factor of 3 - as you only have a single OSD pools can’t be created to the requests from Cinder and Glance to the ceph-mon charm can’t be fulfilled (if I remember right the charm won’t be able to create the pools in this situation).

You need at least 3 units of the ceph-osd application.