Ubuntu 26.04

Ceph Tentacle : Add or Remove OSDs2026/04/27

 

This is how to add or remove OSDs from existing Cluster.

                                         |
        +--------------------+           |           +----------------------+
        |   [dlp.srv.world]  |10.0.0.30  |  10.0.0.31|    [www.srv.world]   |
        |     Ceph Client    +-----------+-----------+        RADOSGW       |
        |                    |           |           |                      |
        +--------------------+           |           +----------------------+
            +----------------------------+----------------------------+
            |                            |                            |
            |10.0.0.51                   |10.0.0.52                   |10.0.0.53 
+-----------+-----------+    +-----------+-----------+    +-----------+-----------+
|   [node01.srv.world]  |    |   [node02.srv.world]  |    |   [node03.srv.world]  |
|     Object Storage    +----+     Object Storage    +----+     Object Storage    |
|     Monitor Daemon    |    |                       |    |                       |
|     Manager Daemon    |    |                       |    |                       |
+-----------------------+    +-----------------------+    +-----------------------+

[1] For example, Add a [node04] node to OSDs on Admin Node.
For Block device on new [node04] Node, use [/dev/sdb] on this example.
# transfer public key

root@node01:~#
ssh-copy-id node04

# if UFW is enabled, allow service ports

root@node01:~#
ssh node04 "ufw allow 6800:7568/tcp"

# install required packages

root@node01:~#
ssh node04 "apt update; apt -y install ceph python3-packaging"
# transfer required files

root@node01:~#
scp /etc/ceph/ceph.conf node04:/etc/ceph/ceph.conf

root@node01:~#
scp /etc/ceph/ceph.client.admin.keyring node04:/etc/ceph

root@node01:~#
scp /var/lib/ceph/bootstrap-osd/ceph.keyring node04:/var/lib/ceph/bootstrap-osd
# configure OSD

root@node01:~# ssh node04 \
"chown ceph:ceph /etc/ceph/ceph.* /var/lib/ceph/bootstrap-osd/*; \
parted --script /dev/sdb 'mklabel gpt'; \
parted --script /dev/sdb "mkpart primary 0% 100%"; \
ceph-volume lvm create --data /dev/sdb1" 
Running command: /usr/bin/ceph-authtool --gen-print-key
Running command: /usr/bin/ceph-authtool --gen-print-key
Running command: /usr/bin/ceph --cluster ceph --name client.bootstrap-osd --keyring /var/lib/ceph/bootstrap-osd/ceph.keyring -i - osd new d946f596-ca7e-4fc9-b355-8bb0cfb152c5
Running command: vgcreate --force --yes ceph-24b660d9-371f-407a-b45c-1ca2e55563bc /dev/sdb1
 stdout: Physical volume "/dev/sdb1" successfully created.
 stdout: Volume group "ceph-24b660d9-371f-407a-b45c-1ca2e55563bc" successfully created
Running command: lvcreate --yes -l 40959 -n osd-block-d946f596-ca7e-4fc9-b355-8bb0cfb152c5 ceph-24b660d9-371f-407a-b45c-1ca2e55563bc
 stdout: Logical volume "osd-block-d946f596-ca7e-4fc9-b355-8bb0cfb152c5" created.
Running command: /usr/bin/mount -t tmpfs tmpfs /var/lib/ceph/osd/ceph-3
.....
.....
Running command: /usr/bin/systemctl start ceph-osd@3
--> ceph-volume lvm activate successful for osd ID: 3
--> ceph-volume lvm create successful for: /dev/sdb1

# after few minutes, it's OK if HEALTH_OK
root@node01:~# ceph -s 
  cluster:
    id:     c77dd493-52c5-415c-af28-c10993ee4f6c
    health: HEALTH_OK

  services:
    mon: 1 daemons, quorum node01 (age 2h) [leader: node01]
    mgr: node01(active, since 84m)
    mds: 1/1 daemons up
    osd: 4 osds: 4 up (since 100s), 4 in (since 106s)
    rgw: 1 daemon active (1 hosts, 1 zones)

  data:
    volumes: 1/1 healthy
    pools:   8 pools, 321 pgs
    objects: 285 objects, 468 KiB
    usage:   184 MiB used, 640 GiB / 640 GiB avail
    pgs:     321 active+clean
[2] To remove an OSD Node from existing Cluster, run commands like follows.
For example, Remove [node04] node.
root@node01:~#
ceph -s

  cluster:
    id:     c77dd493-52c5-415c-af28-c10993ee4f6c
    health: HEALTH_OK

  services:
    mon: 1 daemons, quorum node01 (age 2h) [leader: node01]
    mgr: node01(active, since 84m)
    mds: 1/1 daemons up
    osd: 4 osds: 4 up (since 100s), 4 in (since 106s)
    rgw: 1 daemon active (1 hosts, 1 zones)

  data:
    volumes: 1/1 healthy
    pools:   8 pools, 321 pgs
    objects: 285 objects, 468 KiB
    usage:   184 MiB used, 640 GiB / 640 GiB avail
    pgs:     321 active+clean

root@node01:~#
ceph osd tree

ID  CLASS  WEIGHT   TYPE NAME        STATUS  REWEIGHT  PRI-AFF
-1         0.62476  root default
-3         0.15619      host node01
 0    hdd  0.15619          osd.0        up   1.00000  1.00000
-5         0.15619      host node02
 1    hdd  0.15619          osd.1        up   1.00000  1.00000
-7         0.15619      host node03
 2    hdd  0.15619          osd.2        up   1.00000  1.00000
-9         0.15619      host node04
 3    hdd  0.15619          osd.3        up   1.00000  1.00000

# specify OSD ID of a node you'd like to remove

root@node01:~#
ceph osd out 3

marked out osd.3.
# live watch cluster status
# after running [ceph osd out ***], rebalancing is executed automatically
# to quit live watch, push [Ctrl + c]

root@node01:~#
ceph -w

  cluster:
    id:     c77dd493-52c5-415c-af28-c10993ee4f6c
    health: HEALTH_WARN
            too many PGs per OSD (321 > max 250)

  services:
    mon: 1 daemons, quorum node01 (age 2h) [leader: node01]
    mgr: node01(active, since 85m)
    mds: 1/1 daemons up
    osd: 4 osds: 4 up (since 2m), 3 in (since 12s)
    rgw: 1 daemon active (1 hosts, 1 zones)

  data:
    volumes: 1/1 healthy
    pools:   8 pools, 321 pgs
    objects: 287 objects, 468 KiB
    usage:   192 MiB used, 480 GiB / 480 GiB avail
    pgs:     321 active+clean

  io:
    client:   2.0 KiB/s rd, 127 B/s wr, 1 op/s rd, 0 op/s wr
    recovery: 2.0 KiB/s, 1 keys/s, 30 objects/s
.....
.....

# after status turns to [HEALTH_OK], disable OSD service on the target node

root@node01:~#
ssh node04 "systemctl disable --now ceph-osd@3.service"

# remove the node to specify target OSD ID

root@node01:~#
ceph osd purge 3 --yes-i-really-mean-it

purged osd.3
root@node01:~#
ceph -s

  cluster:
    id:     c77dd493-52c5-415c-af28-c10993ee4f6c
    health: HEALTH_OK

  services:
    mon: 1 daemons, quorum node01 (age 2h) [leader: node01]
    mgr: node01(active, since 88m)
    osd: 3 osds: 3 up (since 70s), 3 in (since 2m)
    rgw: 1 daemon active (1 hosts, 1 zones)

  data:
    pools:   6 pools, 161 pgs
    objects: 263 objects, 459 KiB
    usage:   192 MiB used, 480 GiB / 480 GiB avail
    pgs:     161 active+clean
Matched Content