LINBIT Ansible NFS Cluster

Build an HA NFS Cluster using Ansible with packages from LINBIT.

System requirements:

• An account at https://my.linbit.com (contact [email protected]).
• Deployment environment must have Ansible 2.7.0+ and python-netaddr.
• All target systems must have passwordless SSH access.
• All hostnames used in inventory file are resolvable (better to use IP addresses).
• Target systems are RHEL 7, 8 or 9 (or RHEL compatible distros).

Usage

Edit hosts.ini for your environment and target nodes.

Explanation of the variables:

• cluster_vip is the virtual IP your clients will use to connect to the cluster.
• cluster_vip_cidr is the cidr netmask for your virtual IP belongs in.
• drbd_backing_disk is a raw block device or LVM that DRBD will use as backing storage.
• drbd_replication_network is the subnet your DRBD replication interface exists in.

For example:

[nodes]
192.168.7.190
192.168.7.191
192.168.7.192

[nodes:vars]
cluster_vip="192.168.7.195"
cluster_vip_cidr="24"
drbd_backing_disk="/dev/vdb"
drbd_replication_network="172.16.7.0/24"

# ansible_user needs sudo/root access
ansible_user="root"

If you do not want to be prompted for your https://my.linbit.com credentials, pass --extra-vars to ansible-playbook at execution time:

$ ansible-playbook -e lb_user="username" -e lb_pass="password" -e lb_con_id="1234" -e lb_clu_id="1234" site.yaml

You will be prompted for any variables not set with --extra-vars.

Notes

You can add and remove nodes to the cluster by adding/removing nodes in the inventory file and re-rerunning the playbook. You will lose any data created on your cluster when re-running the playbook (see wipefs task in site.yaml). When "scaling down" (removing nodes) by re-running the playbook, be sure to stop Pacemaker/Corosync/DRBD on the node(s) you're removing.

Feedback

Please pass any issues or questions you may have regarding this playbook to your LINBIT account manager or [email protected].