Documentation
Rate This Document
Findability
Accuracy
Completeness
Readability
SubmitClose

Deploying a Ceph Cluster

  1. Create a local repository.
    1
2
3
4
5
6
    yum -y install createrepo
cd /home/ceph-glz
cp /home/rpmbuild/RPMS/aarch64/*rpm ./ 
createrepo ./
cd /etc/yum.repos.d/
vi ceph-local.repo

     1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
    [local]
name=local
baseurl=file:///home/ceph-glz
enable=1
gpgcheck=0
[Ceph-noarch]
name = Ceph noarch packages
baseurl = http://download.ceph.com/rpm-nautilus/el7/noarch
enabled = 1
gpgcheck = 1
type = rpm-md
gpgkey = https://download.ceph.com/keys/release.asc
priority = 1
  2. Install the RPM package of the compression algorithm.
    1
    rpm -ivh /home/ceph-glz/glz-1.0.0-centos.aarch64.rpm
  3. Deploy the MON and MGR nodes.

    For details, see the corresponding Ceph deployment guide.

    In the deployment guide, the Ceph image source is the official Ceph image, which is a Ceph RPM package that does not contain the compression algorithm plugin. Therefore, you need to configure Ceph using the local repository.

  4. Modify the Ceph configuration file ceph.conf.
    • The configuration items provided in this document are applicable only for the HDD scenario.
    • bluestore_min_alloc_size_hdd indicates the minimum allocated size of the BlueStore object. The default value is 32 KB. You can change the value to 8 KB to reduce the size of the compressed data and ensure the compression ratio.
    • The product of osd_op_num_shards_hdd and osd_op_num_threads_per_shard_hdd is the number of threads for the OSD process to process I/O requests. The default value is 5 x 1. You can change the value to 12 x 2 to deliver the maximum performance of the compression algorithm.
    1
    vi /etc/ceph/ceph.conf
    1. Change the default value of bluestore_min_alloc_size_hdd to 8 KB.
      1
      bluestore_min_alloc_size_hdd = 8192

      The modification must be performed before OSD deployment. Otherwise, the modification does not take effect. If an OSD node has been deployed, deploy it again.

    2. Change the default number of OSD threads.
      1
2
      osd_op_num_shards_hdd = 12
osd_op_num_threads_per_shard_hdd = 2

      The modification can be dynamically adjusted after the OSD node is deployed.

  5. Deploy the OSD node.

    For details, see Deploying OSD Nodes in the Ceph Block Storage Deployment Guide (CentOS 7.6 & openEuler 20.03).