I am using GlusterFS to create and mount volumes across 4 machines. Say for example, the machines are called machine1, machine2, machine3 and machine4.
My peers have already been successfully probed.
I have used the following command to create my volume:
sudo gluster volume create ssl replica 2 transport tcp machine1:/srv/gluster/ssl machine2:/srv/gluster/ssl machine3:/srv/gluster/ssl machine4:/srv/gluster/ssl force
I then start the volume with:
sudo gluster volume start ssl
I have mounted the directory /myproject/ssl using the following command:
sudo mount -t glusterfs machine1:/ssl /myproject/ssl
When mounted on each machine, everything works as expected and the /myproject/ssl directory has data shared across all machines.
The question is, how on Earth do I do this the Ansible way?
Here are my attempts at doing those two commands the Ansible way:
- name: Configure Gluster volume.
gluster_volume:
state: present
name: "{{ gluster.brick_name }}"
brick: "{{ gluster.brick_dir }}"
replicas: 2
cluster: "{{ groups.glusterssl | join(',') }}"
host: "{{ inventory_hostname }}"
force: yes
become: true
become_user: root
become_method: sudo
run_once: true
ignore_errors: true
- name: Ensure Gluster volume is mounted.
mount:
name: "{{ gluster.brick_name }}"
src: "{{ inventory_hostname }}:/{{ gluster.brick_name }}"
fstype: glusterfs
opts: "defaults,_netdev"
state: mounted
become: true
become_user: root
become_method: sudo
Despite a peer probe already returning successful on a previous task, the Configure Gluster volume task fails with:
fatal: [machine3]: FAILED! =>
{"changed": false,
"failed": true,
"invocation": {
"module_args": {
"brick": "/srv/gluster/ssl",
"bricks": "/srv/gluster/ssl",
"cluster": ["machine1", "machine2", "machine3", "machine4"],
"directory": null,
"force": true,
"host": "machine3",
"name": "ssl",
"options": {},
"quota": null,
"rebalance": false,
"replicas": 2,
"start_on_create": true,
"state": "present",
"stripes": null,
"transport": "tcp"},
"module_name": "gluster_volume"},
"msg": "failed to probe peer machine1 on machine3"}
If I replace this Ansible task with the first shell command I suggested, everything works fine, but then the Ensure Gluster volume is mounted fails with:
fatal: [machine3]: FAILED! =>
{"changed": false,
"failed": true,
"invocation": {
"module_args": {
"dump": null,
"fstab": "/etc/fstab",
"fstype": "glusterfs",
"name": "ssl", "opts":
"defaults,_netdev",
"passno": null, "src":
"machine3:/ssl",
"state": "mounted"},
"module_name": "mount"},
"msg": "Error mounting ssl: Mount failed. Please check the log file for more details.\n"}
The relevant log output is:
[2016-10-17 09:10:25.602431] E [MSGID: 114058] [client-handshake.c:1524:client_query_portmap
_cbk] 2-ssl-client-3: failed to get the port number for remote subvolume. Please run 'gluster volume status' on server to see if brick process is running.
[2016-10-17 09:10:25.602480] I [MSGID: 114018] [client.c:2042:client_rpc_notify] 2-ssl-client-3: disconnected from ssl-client-3. Client process will keep trying to connect to glusterd until brick's port is available
[2016-10-17 09:10:25.602500] E [MSGID: 108006] [afr-common.c:3880:afr_notify] 2-ssl-replicate-1: All subvolumes are down. Going offline until atleast one of them comes back up.
[2016-10-17 09:10:25.616402] I [fuse-bridge.c:5137:fuse_graph_setup] 0-fuse: switched to graph 2
So, the volume doesn't get started by the Ansible task.
My question is, essentially, how do I create, mount and start a volume in the same way I did with the 3 commands mentioned above, the Ansible way?
