Resource allocation in high availability (ha) systems

A computer implemented method, comprising:

• detecting, by one or more processor of a computer, that a primary node of a high availability set in a high availability system has more incoming computational tasks than the primary node can reliably handle due to a lack of resources allocated for the primary node, wherein the primary node runs one or more compute process for processing the incoming computational tasks by use of resources allocated for the primary node, and wherein the high availability set includes two or more backup nodes in a queue, each backup node of the two or more backup nodes respectively includes resources allocated for the respective backup node;
• selecting, by the one or more processor of the computer, from all backup nodes of the high availability system, a backup node that has the resources lacking from the primary node and that is second most likely to be used for processing the incoming computational tasks as a helper node, according to an offloading policy of the high availability system;
• adjusting, by the one or more processor of the computer, a state of the helper node from the selecting to indicate that the helper node is being used for the incoming computational tasks but not being the primary node, and dequeuing the helper node from the queue of the backup nodes, according to a backup policy of the high availability system;
• sending, by the one or more processor of the computer, some of the incoming computational tasks for the high availability set to the helper node from the selecting instead of to the primary node.

A system comprising:

• a memory;
• one or more processor in communication with memory;
• and
• program instructions executable by the one or more processor via the memory to perform a method for a high availability system, comprising:
• a primary compute node performing one or more compute service for incoming computational tasks for a high availability set in the high availability system, comprising a virtual machine and resources;
• one or more backup compute nodes additional to the primary compute node on respective virtual machines with respective resources, being present in the high availability set as a fail-safe measure for the primary compute node, wherein the high availability system includes one or more high availability set including the high availability set, wherein all compute nodes in the high availability system are respectively associated with a respective readiness state, wherein the backup compute nodes are in a queue and a backup policy dictates that a backup compute node positioned at the head of the queue is in a state of most readiness;
• and
• a high availability coordinator in a controller node of each of high availability set in the high availability system, wherein the high availability coordinator manages one or more compute service that processes computational tasks with high availability basis, wherein operations by the high availability coordinator comprise:
• generating virtual machines and designating the virtual machines as respective compute nodes including the primary compute node and the backup computer nodes of the high availability set;
• detecting whether or not the primary compute node of one of the high availability set is about to fail;
• determining an action to avoid a failure of the primary compute node from the detecting, wherein the action is selected from borrowing and offloading;
• ascertaining whether or not the primary compute node continues the compute service subsequent to taking the action from the determining;
• and
• handling the primary compute node and the backup compute nodes according to a preconfigured backup policy or a preconfigured failure policy, responsive to the ascertaining.