手动添加ceph的mds

前提条件

已经部署好ceph集群（ceph-luminous）

1、创建mds(只创建一个单节点)

[root@cluster9 ceph-cluster]# ceph-deploy mds create cluster9

2、创建data pool和metadata pool

[root@cluster9 ceph-cluster]# ceph osd pool create cephfs_data 128 128
pool 'cephfs_data' created
[root@cluster9 ceph-cluster]# ceph osd pool create cephfs_metadata 128 128
pool 'cephfs_metadata' created

3、新建cephfs

[root@cluster9 ceph-cluster]# ceph fs new cephfs cephfs_metadata cephfs_data
new fs with metadata pool 5 and data pool 4

4、查看集群状态

[root@cluster9 ceph-cluster]# ceph -s
  cluster:
    id:     d81b3ce4-bcbc-4b43-870e-430950652315
    health: HEALTH_WARN
            too many PGs per OSD (384 > max 250)
 
  services:
    mon: 1 daemons, quorum cluster9
    mgr: cluster9(active)
    mds: cephfs-1/1/1 up  {0=cluster9=up:active}
    osd: 3 osds: 3 up, 3 in
 
  data:
    pools:   3 pools, 384 pgs
    objects: 26 objects, 2.22KiB
    usage:   3.07GiB used, 10.9TiB / 10.9TiB avail
    pgs:     384 active+clean

解决上面的问题，进入ceph-deploy目录，修改ceph.conf

 

 将上在的值修改

然后执行以下命令

ceph-deploy --overwrite-conf config push cluster9

systemctl restart ceph-mgr.target

查看是否生效

[root@cluster9 ceph-cluster]# ceph --show-config | grep "mon_max_pg_per_osd"
mon_max_pg_per_osd = 500

[root@cluster9 ceph-cluster]# ceph -s
  cluster:
    id:     d81b3ce4-bcbc-4b43-870e-430950652315
    health: HEALTH_OK
 
  services:
    mon: 1 daemons, quorum cluster9
    mgr: cluster9(active)
    mds: cephfs-1/1/1 up  {0=cluster9=up:active}
    osd: 3 osds: 3 up, 3 in
 
  data:
    pools:   3 pools, 384 pgs
    objects: 26 objects, 2.22KiB
    usage:   3.07GiB used, 10.9TiB / 10.9TiB avail
    pgs:     384 active+clean

cephfs 部署完