Ansible playbooks to deploy Ceph, the distributed filesystem.

Go to file

Leseb be26352a2c Merge pull request #88 from Sysnove/hosts Replaced all by the list of host groups to be able to run the playbook with non-ceph servers.		2014-07-08 14:12:22 +02:00
fetch	Add Ceph Playbook	2014-03-03 19:08:51 +01:00
group_vars	Use a local key for package instead of downloading it	2014-07-03 15:53:13 +02:00
roles	Merge pull request #89 from Sysnove/partitions	2014-07-08 10:27:11 +02:00
.gitignore	Use Vagrant's Ansible provisioner	2014-05-11 19:34:37 -04:00
LICENSE	Add Ceph Playbook	2014-03-03 19:08:51 +01:00
README.md	Update README.md	2014-06-13 16:30:50 +02:00
Vagrantfile	Use Vagrant's Ansible provisioner	2014-05-11 19:34:37 -04:00
ansible.cfg	Use Vagrant's Ansible provisioner	2014-05-11 19:34:37 -04:00
maintenance.yml	Add maintenance playbook	2014-04-09 17:51:59 +02:00
purge.yml	Fix #73 by using `ansible_fqdn`	2014-04-18 10:55:31 +02:00
rolling_update.yml	Replaced all by the list of host groups to be able to run the playbook with non-ceph servers.	2014-06-26 09:52:06 +02:00
site.yml	Replaced all by the list of host groups to be able to run the playbook with non-ceph servers.	2014-06-26 09:52:06 +02:00

README.md

ceph-ansible

Ansible playbook for Ceph!

What does it do?

General support for:

Monitors
OSDs
MDSs
RGW

More details:

Authentication (cephx), this can be disabled.
Supports cluster public and private network.
Monitors deployment. You can easily start with one monitor and then progressively add new nodes. So can deploy one monitor for testing purpose. For production, I recommend to always use an odd number of monitors, 3 tends to be the standard.
Object Storage Daemons. Like the monitors you can start with a certain amount of nodes and then grow this number. The playbook either supports a dedicated device for storing the journal or both journal and OSD data on the same device (using a tiny partition at the beginning of the device).
Metadata daemons.
Collocation. The playbook supports collocating Monitors, OSDs and MDSs on the same machine.
The playbook was validated on Debian Wheezy, Ubuntu 12.04 LTS and CentOS 6.4.
Tested on Ceph Dumpling and Emperor.
A rolling upgrade playbook was written, an upgrade from Dumpling to Emperor was performed and worked.

Setup with Vagrant

Run your virtual machines:

$ vagrant up
...
...
...
 ____________
< PLAY RECAP >
 ------------
        \   ^__^
         \  (oo)\_______
            (__)\       )\/\
                ||----w |
                ||     ||


mon0                       : ok=16   changed=11   unreachable=0    failed=0
mon1                       : ok=16   changed=10   unreachable=0    failed=0
mon2                       : ok=16   changed=11   unreachable=0    failed=0
osd0                       : ok=19   changed=7    unreachable=0    failed=0
osd1                       : ok=19   changed=7    unreachable=0    failed=0
osd2                       : ok=19   changed=7    unreachable=0    failed=0
rgw                        : ok=20   changed=17   unreachable=0    failed=0

Check the status:

$ vagrant ssh mon0 -c "sudo ceph -s"
    cluster 4a158d27-f750-41d5-9e7f-26ce4c9d2d45
     health HEALTH_OK
     monmap e3: 3 mons at {ceph-mon0=192.168.0.10:6789/0,ceph-mon1=192.168.0.11:6789/0,ceph-mon2=192.168.0.12:6789/0}, election epoch 6, quorum 0,1,2 ceph-mon0,ceph-mon1,ceph-mon
     mdsmap e6: 1/1/1 up {0=ceph-osd0=up:active}, 2 up:standby
     osdmap e10: 6 osds: 6 up, 6 in
      pgmap v17: 192 pgs, 3 pools, 9470 bytes data, 21 objects
            205 MB used, 29728 MB / 29933 MB avail
                 192 active+clean

To re-run the Ansible provisioning scripts:

$ vagrant provision

Specifying fsid and secret key in production

The Vagrantfile specifies an fsid for the cluster and a secret key for the monitor. If using these playbooks in production, you must generate your own fsid in group_vars/all and monitor_secret in group_vars/mons. Those files contain information about how to generate appropriate values for these variables.