Secure cluster pairing for business continuity and disaster recovery

An apparatus, comprising:

• a communication interface that enables network communications;
• a hardware processor;
• and
• a memory storing data and instructions executable by the hardware processor,
• wherein the hardware processor is configured to execute the instructions to:
• establish a first subnet for replication in a first cluster that includes a plurality of host devices, each of the host devices including a respective controller virtual machine, which together form a virtual local area network for replication, each of the controller virtual machines being assigned an Ethernet interface;
• assign a respective replication Internet Protocol (IP) address to the Ethernet interface of each of the controller virtual machines;
• modify route tables and firewall rules of the controller virtual machines to allow communications between nodes of the first subnet;
• designate a node in the first cluster to control replication operations with a second cluster, and assign to the node in the first cluster a replication cluster IP address;
• receive, by the node in the first cluster, via the replication cluster IP address, information related to a second subnet for replication in a second cluster;
• determine whether the second subnet and the first subnet are engaged in a self-pairing operation;
• in response to determining that the second subnet and the first subnet are not engaged in a self-pairing operation, configure the first subnet with the information related to the second subnet to pair the first cluster with the second cluster;
• generate a dedicated communication channel for replication between the first cluster and the second cluster based on the configuring;
• communicate information for disaster recovery to the second cluster through the dedicated communication channel so as to provide backup protection for the first cluster;
• determine that the second subnet is not reachable for a predetermined amount of time;
• and
• remove the dedicated communication channel for replication between the first cluster and the second cluster by un-pairing the first cluster from the second cluster, wherein the un-pairing is performed at the first cluster, but not at the second cluster, which is not reachable.

A non-transitory computer-readable storage medium encoded with software comprising computer executable instructions which, when executed by a hardware processor of a node in a first computer cluster, cause the hardware processor to:

• establish a first subnet for replication in a first cluster that includes a plurality of host devices, each of the host devices including a respective controller virtual machine, which together form a virtual local area network for replication, each of the controller virtual machines being assigned an Ethernet interface;
• assign a respective replication Internet Protocol (IP) address to the Ethernet interface of each of the controller virtual machines;
• modify route tables and firewall rules of the controller virtual machines to allow communications between nodes of the first subnet;
• designate a node in the first cluster to control replication operations with a second cluster, and assign to the node in the first cluster a replication cluster IP address;
• receive, by the node in the first cluster, via the replication cluster IP address, information related to a second subnet for replication in a second cluster;
• determine whether the second subnet and the first subnet are engaged in a self-pairing operation;
• in response to determining that the second subnet and the first subnet are not engaged in a self-pairing operation, configure the first subnet with the information related to the second subnet to pair the first subnet with the second subnet;
• generate a dedicated communication channel for replication between the first cluster and the second cluster based on the configuring;
• communicate information for disaster recovery to the second cluster through the dedicated communication channel so as to provide backup protection for the first cluster;
• determine that the second subnet is not reachable for a predetermined amount of time;
• and
• remove the dedicated communication channel for replication between the first cluster and the second cluster by un-pairing the first cluster from the second cluster, wherein the un-pairing is performed at the first cluster, but not at the second cluster, which is not reachable.