Skip to content

xinxinsh/ceph-deployment

Repository files navigation

Simple instruction
=========================================

ansible-playbook prepilot.yml

ansible-playbook provision.yml

ansible-playbook install.yml

ansible-playbook initialize.yml

Some hints for group_vars/all.yml
=========================================
1. Each ceph node need two nics
One is for data and the other is for management

2. Each ceph node need another three hard disks at least
For example, sdb, sdc, sdd 


How to add osd to cluster

1. Edit inventory/host
   (a) if the new osd is on a new host, add a line as follows
       tt-slave2 ansible_connection=ssh ansible_user=root ansible_ssh_pass=huayun ansible_host=10.21.1.99 ansible_port=22 ceph_network_ip=10.21.1.99 vm_network_ip=10.21.1.99
   (b) then add this host to [EXT] section
       [EXT]
       tt-slave2

2. Edit group_vars/all.yml
   (a) Edit ext_osd_layouts section
       if the journal of new osd is on seperate disk, go to (b), otherwise go to (c),
       the follow is an example of new osd on vda without seperate journal

      - device: vda
        use_cache: no
        cache: vdb1
        use_type: yes
        type: hdd
       
   (b) Edit ext_disk_partition_layouts section
       the following example shows vdb as journal partition

       - device: vdb
         partitions:
         - percent: [4096s, 100%]
           name: journal01
           filesystem: xfs

   (c) Edit ext_osd_layout_map section 
       if the host is new to cluster, edit this section, the following example shows host
       tt-slave2 is new host with osd layout (ceph_osd_layout_2) and partition layout 
       (disk_partition_layout2)
  
       tt-slave2:
         partition: disk_partition_layout2
         osd: ceph_osd_layout_2

3. ansible-playbook extend.yml

About

ceph ansible deployment

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages