Ceph » RADOS

OSD created using ceph-deploy "ceph-deploy osd create --filestore", metadata after starting up is:

{
    "id": 9,
    "arch": "x86_64",
    "back_addr": "192.168.19.6:6802/2738",
    "backend_filestore_dev_node": "unknown",
    "backend_filestore_partition_path": "unknown",
    "ceph_version": "ceph version 12.1.2 (b661348f156f148d764b998b65b90451f096cb27) luminous (rc)",
    "cpu": "Intel(R) Celeron(R) CPU G1610T @ 2.30GHz",
    "default_device_class": "hdd",
    "distro": "centos",
    "distro_description": "CentOS Linux 7 (Core)",
    "distro_version": "7",
    "filestore_backend": "xfs",
    "filestore_f_type": "0x58465342",
    "front_addr": "192.168.18.6:6803/2738",
    "hb_back_addr": "192.168.19.6:6803/2738",
    "hb_front_addr": "192.168.18.6:6804/2738",
    "hostname": "gravel6.rockery",
    "kernel_description": "#1 SMP Tue Jul 4 15:04:05 UTC 2017",
    "kernel_version": "3.10.0-514.26.2.el7.x86_64",
    "mem_swap_kb": "4063228",
    "mem_total_kb": "7965928",
    "os": "Linux",
    "osd_data": "/var/lib/ceph/osd/ceph-9",
    "osd_journal": "/var/lib/ceph/osd/ceph-9/journal",
    "osd_objectstore": "filestore",
    "rotational": "1" 
}

While our code can't be expected to work out block devices in absolutely all cases, this seems like pretty much the simplest case we have for creating filestore OSDs, on a centos 7 node using bare SATA drives, so I'm thinking this is a bug if we're not finding the device node even in this case.