Ceph Squid : Add or Remove OSDs2024/05/03 |
|
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 # 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 --cluster ceph --name client.bootstrap-osd --keyring /var/lib/ceph/bootstrap-osd/ceph.keyring -i - osd new 7229113d-beb3-40ce-bcef-ed266dfef5e0 Running command: vgcreate --force --yes ceph-8ea4d664-ff40-43e1-b7ca-70d26978d73f /dev/sdb1 stdout: Physical volume "/dev/sdb1" successfully created. stdout: Volume group "ceph-8ea4d664-ff40-43e1-b7ca-70d26978d73f" successfully created Running command: lvcreate --yes -l 40959 -n osd-block-7229113d-beb3-40ce-bcef-ed266dfef5e0 ceph-8ea4d664-ff40-43e1-b7ca-70d26978d73f ..... ..... 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: 3666a474-14e0-4c5f-ad1e-daf2e30aed8f health: HEALTH_OK services: mon: 1 daemons, quorum node01 (age 51m) mgr: node01(active, since 50m) mds: 1/1 daemons up osd: 4 osds: 4 up (since 4m), 4 in (since 4m) rgw: 1 daemon active (1 hosts, 1 zones) data: volumes: 1/1 healthy pools: 7 pools, 289 pgs objects: 248 objects, 456 KiB usage: 211 MiB used, 640 GiB / 640 GiB avail pgs: 289 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: 3666a474-14e0-4c5f-ad1e-daf2e30aed8f
health: HEALTH_OK
services:
mon: 1 daemons, quorum node01 (age 65m)
mgr: node01(active, since 64m)
mds: 1/1 daemons up
osd: 4 osds: 4 up (since 18m), 4 in (since 18m)
rgw: 1 daemon active (1 hosts, 1 zones)
data:
volumes: 1/1 healthy
pools: 7 pools, 289 pgs
objects: 248 objects, 456 KiB
usage: 215 MiB used, 640 GiB / 640 GiB avail
pgs: 289 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: 3666a474-14e0-4c5f-ad1e-daf2e30aed8f
health: HEALTH_WARN
too many PGs per OSD (289 > max 250)
services:
mon: 1 daemons, quorum node01 (age 66m)
mgr: node01(active, since 65m)
mds: 1/1 daemons up
osd: 4 osds: 4 up (since 19m), 3 in (since 7s)
rgw: 1 daemon active (1 hosts, 1 zones)
data:
volumes: 1/1 healthy
pools: 7 pools, 289 pgs
objects: 249 objects, 456 KiB
usage: 170 MiB used, 480 GiB / 480 GiB avail
pgs: 289 active+clean
io:
client: 0 B/s wr, 0 op/s rd, 0 op/s wr
recovery: 1003 B/s, 1 keys/s, 36 objects/s
2024-05-03T04:01:27.510202+0000 mon.node01 [INF] Health check cleared: PG_AVAILABILITY (was: Reduced data availability: 7 pgs peering)
.....
.....
# 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 ceph -s
cluster:
id: 3666a474-14e0-4c5f-ad1e-daf2e30aed8f
health: HEALTH_OK
services:
mon: 1 daemons, quorum node01 (age 70m)
mgr: node01(active, since 69m)
osd: 3 osds: 3 up (since 2m), 3 in (since 4m)
rgw: 1 daemon active (1 hosts, 1 zones)
data:
pools: 5 pools, 129 pgs
objects: 226 objects, 454 KiB
usage: 175 MiB used, 480 GiB / 480 GiB avail
pgs: 129 active+clean
|
| Sponsored Link |