arch_single_shared
This scenario is a variation of the local storage setup. Here, the storage for virtual machines (VMs) and the image repository are provided by a NFS/NAS server. Running VMs directly from shared storage can enhance the fault tolerance of the system in the event of a host failure, although it comes with the drawback of increased I/O latency.
The NFS/NAS server is configured to export the datastore folders to the hosts in the OpenNebula cloud. In this example we assume that the following structure is created in the NFS/NAS sever:
ls -ln /srv
total 0
drwxr-xr-x 2 9869 9689 6 Jun 26 17:55 0
drwxr-xr-x 2 9869 9689 6 Jun 26 17:55 1
drwxr-xr-x 2 9869 9689 6 Jun 26 17:55 2and exported to the OpenNebula servers, for example:
$ cat /etc/exports
# /etc/exports
#
# See exports(5) for more information.
#
# Use exportfs -r to reread
# /export 192.168.1.10(rw,no_root_squash)
/srv 172.20.0.0/24(rw,soft,intr,async,rsize=32768, wsize=32768)Virtual disk images are stored in local storage, with the front-end hosting an image repository (image datastore). These images are subsequently transferred from the front-end to the hypervisors to initiate the virtual machines (VMs). Both the front-end and hypervisors utilize the directory /var/lib/one/datastores to store these images. It is possible to either utilize the root file system (FS) for this directory or symlink from any other location.
The following snippet shows the configuration required if no mount points are used:
ds:
mode: ssh/var/lib/one/datastores/.
If you want to use a dedicated volume mounted in a custom location (for example /mnt/one_datastores/), then you need to pre-create directories (owned by oneadmin) for each datastore and use the following snippet which will cause symlinks to be created automatically:
ds:
mode: ssh
mounts:
- type: system
path: /mnt/one_datastores/system/
- type: image
path: /mnt/one_datastores/default/
- type: files
path: /mnt/one_datastores/files/The final setting on the hosts will be:
$ tree /mnt/one_datastores/
/mnt/one_datastores/
├── system
├── default
└── files
$ tree /var/lib/one/datastore/
/var/lib/one/datastores/
├── 0 -> /mnt/one_datastores/system/
├── 1 -> /mnt/one_datastores/default/
└── 2 -> /mnt/one_datastores/files/The most basic network configuration is a flat network (bridged). We will use the main interface of the Host to connect the VMs to the Network. The interfaces used in this mode are depicted in the following picture:
To create the virtual network for the VMs you need to pick up some IP. These IP addresses need to be reachable through the Network used by the main interface of the host, as the VM traffic will be forwarded through it.
The following snippet shows how to define a virtual network using some IPs in the Admin Network (the one used by the hosts):
vn:
bridge:
admin_net:
managed: true
template:
VN_MAD: bridge
PHYDEV: eth0
BRIDGE: br0
AR:
TYPE: IP4
IP: 10.0.0.50
SIZE: 48
NETWORK_ADDRESS: 10.0.0.0
NETWORK_MASK: 255.255.255.0
GATEWAY: 10.0.0.1
DNS: 1.1.1.1If there is any other interface in the hosts you can use them. For example to define a dedicated VM network using bon0 and vxlan networking:
vxlan:
vms_net:
managed: true
template:
VN_MAD: vxlan
PHYDEV: bond0
BRIDGE: br1
VLAN_ID: 123
FILTER_IP_SPOOFING: 'NO'
FILTER_MAC_SPOOFING: 'YES'
GUEST_MTU: 1450
AR:
TYPE: IP4
IP: 192.168.0.10
SIZE: 100
NETWORK_ADDRESS: 192.168.0.0
NETWORK_MASK: 255.255.255.0
GATEWAY: 192.168.0.1
DNS: 192.168.0.1The Ansible playbook installs a complete suite of OpenNebula services including the base daemons (oned and scheduler), the OpenNebula Flow and Gate services and Sunstone Web-UI. You can just need to select the OpenNebula version to install and a pick a password for oneadmin
one_pass: opennebula
one_version: '6.6'You can use your enterprise distribution with the Ansible playbooks. Simply add your token to the var file. Also you can enable the Prometheus and Grafana integration part of the Enterprise Edition:
one_token: example:example
features:
prometheus: trueThe following file show the complete settings to install a single front-end with two hosts using local storage:
---
all:
vars:
ansible_user: root
one_version: '6.6'
one_pass: opennebulapass
vn:
admin_net:
service:
managed: true
template:
VN_MAD: bridge
PHYDEV: eth0
BRIDGE: br0
AR:
TYPE: IP4
IP: 172.20.0.100
SIZE: 48
NETWORK_ADDRESS: 172.20.0.0
NETWORK_MASK: 255.255.255.0
GATEWAY: 172.20.0.1
DNS: 1.1.1.1
frontend:
hosts:
fe1: { ansible_host: 172.20.0.7 }
node:
hosts:
node1: { ansible_host: 172.20.0.8 }
node2: { ansible_host: 172.20.0.9 }-
1. Prepare the inventory file: Update the
local.ymlfile in the inventory file to match your infrastructure settings. Please be sure to update or review the following variables:-
ansible_user, update it if different from root. -
one_pass, change it to the password for the oneadmin account -
one_version, be sure to use the latest stable version here
-
-
2. Check the connection: Verify the network connection, ssh and sudo configuration run the following command:
ansible -i inventory/local.yml all -m ping -b- 3. Site installation: Now we can run the site playbook that install and configure OpenNebula services
ansible-playbook -i inventory/local.yml opennebula.deploy.mainNow that the OpenNebula cloud is installed and ready to use let's review your installation. Let's first check the hosts, ssh into the frontend and check the hosts registered in OpenNebula:
# sudo -i -u oneadmin
$ onehost list
In the same way you can check the datastores and virtual networks
$ onedatastore list
$ onevnet list
Finally let's create a simple VM. Let's download an alpine image from the OpenNebula MarketPlace:
$ onemarketapp export
And instantiate the template attached to the admin_net network:
$ onetemplate instantiate