Configuring Cluster

Prerequisites

  • A configured cluster witness system. See Witness for more information.

  • All cluster members must be powered on.

  • All cluster members must be using the same management network.

  • Two BrickStor SP storage nodes using the same software version.

  • Must have a common storage pool accessible by all cluster nodes.

  • Physical heartbeat network connection between the HA cluster nodes.

  • Common data network interface with the same name on both nodes. It could be a physical interface or an aggregate.

Determine Common Data Network Interface

A data network interface is used for client data access for sharing protocols such as SMB or NFS. To configure HA, there must me a common physical interface present on both cluster nodes. It is used to create virtual network interfaces (VNICs) associated with a particular HA resource group. A common network interface ensures VNICs are created on the desired network during the failover events. This physical interfaces must have identical names on both cluster nodes.

This common data interface is typically composed of two or more physical interfaces aggregated together using LACP (IEEE 802.3ad).

To determine the name of the common data network interface or aggregate, Navigate to the Network Links page. Note the Name of the desired network interface for cluster creation.

Configuring Heartbeat

To create the Heartbeat connection between the HA cluster nodes, follow the steps below on both cluster nodes:

  1. Navigate to the Network Links page.

  2. Select Add Address on the action dropdown of the desired network link.

  3. Enter hb0 for the Name.

  4. Enter the IP address and subnet mask for Address/Subnet.

  5. Click Add.

Creating a Cluster

To create a High Availability Cluster utilizing two storage nodes and a witness server, follow the steps below:

  1. Navigate to the HA page.

  2. Click Create Cluster.

  3. Under the Local Node enter the management and the heartbeat IP addresses of the currently managed BrickStor SP system.

  4. Under the Remote Node enter the management and the heartbeat IP addresses of the other BrickStor SP system joining this cluster.

  5. Enter the IP address of the witness server.

  6. Enter the name of the common data network interface. See Determine Common Data Network Interface for more information.

  7. Click Create.

Setup confd

When the cluster is created, both HA nodes are automatically joined to the confd cluster. The witness must be manually joined. This is important since confd requires three systems to have a functioning cluster.

To join a witness to the confd cluster, follow the steps below:

  1. Determine the cluster leader

    1. From either of the cluster nodes, run confadm member show all.

    2. Note the IP address of the leader.

  2. On the witness system, using a terminal, run confadm join and follow the prompts.

    1. Enter the IP of the cluster leader.

    2. Enter the admin credentials.

    3. Accept the confirmations and allow up to 90 seconds for successful joining.