Managing use of lease resources allocated on fallover in a high availability computing environment

Abstract

Responsive to a cluster manager for a particular node from among multiple nodes allocating at least one leased resource for a resource group for an application workload on the particular node, on fallover of the resource group from another node to the particular node, setting a timer thread, by the cluster manager for the particular node, to track an amount of time remaining for an initial lease period of the at least one leased resource. Responsive to the timer thread expiring while the resource group is holding the at least one leased resource and at least one other node from among the plurality of nodes with at least one active leased resource available to handle the resource group during at least one additional lease period at a second fee that is lower than an additional fee for continuing to use the at least one leased resource after the timer thread has expired, moving the resource group to the at least one other node and automatically starting a timer set to an amount of time in a shortest lease period remaining from among the at least one additional lease period for use of the at least one active leased resource.

1 . A method for managing resources, comprising:
 responsive to a cluster manager for a particular node from among a plurality of nodes communicatively connected through a network allocated at least one leased resource for a resource group for an application workload on the particular node, on fallover of the resource group from another node from among the plurality of nodes to the particular node, setting, by the cluster manager for the particular node, a timer thread to track an amount of time remaining for an initial lease period of the at least one leased resource, the initial lease period available for a first fee, the cluster manager coupled to at least one processor and memory; and   responsive to the timer thread expiring while the resource group is holding the at least one leased resource and at least one other node from among the plurality of nodes with at least one active leased resource available to handle the resource group during at least one additional lease period at a second fee that is lower than an additional fee for continuing to use the at least one leased resource after the timer thread has expired, moving, by the cluster manager for the particular node, the resource group to the at least one other node and automatically starting a timer set to an amount of time in a shortest lease period remaining from among the at least one additional lease period for use of the at least one active leased resource.   
     
     
         2 . The method according to  claim 1 , wherein responsive to the timer thread expiring while the resource group is holding the at least one leased resource and at least one other node from among the plurality of nodes with at least one active leased resource available to handle the resource group during at least one additional lease period at a second fee that is lower than an additional fee for continuing to use the at least one leased resource after the timer thread has expired, moving, by the cluster manager for the particular node, the resource group to the at least one other node and automatically starting a timer set to an amount of time in a shortest lease period remaining from among the at least one additional lease period for use of the at least one active leased resource further comprises:
 triggering fallover of the resource group to the least one other node; and   initiating a message indicating the resource group has been moved to the at least one other node.   
     
     
         3 . The method according to  claim 1 , further comprising:
 responsive to starting the timer set to the amount of time in the shortest lease period remaining, outputting, but the cluster manager for the particular node, a notification that a new fee is incurred for the use of the at least one active leased resource for the amount of time in the shortest lease period remaining.   
     
     
         4 . A computer system for managing resources, the computer system comprising:
 a cluster manager, coupled to at least one processor and memory, for a particular node from among a plurality of nodes communicatively connected through a network;   the cluster manager programmed to:   responsive to the cluster manager allocated at least one leased resource for a resource group for an application workload on the particular node, on fallover of the resource group from another node from among the plurality of nodes to the particular node, set a timer thread to track an amount of time remaining for an initial lease period of the at least one leased resource, the initial lease period available for a first fee, the cluster manager coupled to at least one processor and memory; and   responsive to the timer thread expiring while the resource group is holding the at least one leased resource and at least one other node from among the plurality of nodes with at least one active leased resource available to handle the resource group during at least one additional lease period at a second fee that is lower than an additional fee for continuing to use the at least one leased resource after the timer thread has expired, move the resource group to the at least one other node and automatically starting a timer set to an amount of time in a shortest lease period remaining from among the at least one additional lease period for use of the at least one active leased resource.   
     
     
         5 . The computer system according to  claim 4 , the cluster manager further programmed to:
 trigger fallover of the resource group to the least one other node; and   initiate a message indicating the resource group has been moved to the at least one other node.   
     
     
         6 . The computer system according to  claim 4 , the cluster manager further programmed to:
 responsive to starting the timer set to the amount of time in the shortest lease period remaining, output a notification that a new fee is incurred for the use of the at least one active leased resource for the amount of time in the shortest lease period remaining.   
     
     
         7 . A method, comprising:
 responsive to a cluster manager for a particular node from among a plurality of nodes communicatively connected through a network allocating at least one leased resource for a resource group for an application workload on the particular node on fallover of the resource group from another node from among the plurality of nodes to the particular node, setting, by the cluster manager for the particular node, a timer thread to track an amount of time remaining for an initial lease period of the at least one leased resource; and   responsive to the timer thread expiring while the resource group is holding the at least one leased resource, maintain, by the cluster manager for the particular node, the resource group comprising the at least one leased resource for an additional lease period and automatically incurring an additional fee for use of the at least one leased resource by the particular node only if the particular node has the capacity to handle the resource group at a lowest cost from among the plurality of nodes.   
     
     
         8 . The method according to  claim 7 , wherein responsive to a cluster manager for a particular node from among a plurality of nodes communicatively connected through a network allocating at least one leased resource for a resource group for an application workload on the particular node on fallover of the resource group from another node from among the plurality of nodes to the particular node, setting, by the cluster manager for the particular node, a timer thread to track an amount of time remaining for an initial lease period of the at least one leased resource further comprises:
 activating, by the cluster manager for the particular node, the at least one leased resource from an inactive resource to an active resource by acquiring a license for the at least one leased resource, wherein the cluster manager is only enabled to allocate the at least one leased resource as an active resource, wherein the license for the at least one leased resource specifies an initial fee for the initial lease period and the additional fee for the additional lease period, wherein the license for the at least one leased resource comprises an agreement to pay the additional fee for the additional lease period if the at least one leased resource is still held by the resource group at the expiration of the initial lease period; and   allocating, by the cluster manager for the particular node, the at least one leased resource to the resource group, wherein the resource group holds the at least one leased resource in the resource group until at least one of the application workload has completed or the resource group is moved to an available node from among the plurality of nodes.   
     
     
         9 . The method according to  claim 7 , wherein responsive to the timer thread expiring while the resource group is holding the at least one leased resource, maintain, by the cluster manager for the particular node, the resource group comprising the at least one leased resource for an additional lease period and automatically incurring an additional fee for use of the at least one leased resource by the particular node only if the particular node has the capacity to handle the resource group at a lowest cost from among the plurality of nodes further comprises:
 receiving, by the cluster manager for the particular node, a message indicating an error on the another node;   deciding, by the cluster manager for the particular node, to trigger a fallover of the resource group from the another node to the particular node; and   managing, by the cluster manager for the particular node, fallover of the resource group from the another node to the particular node by allocating a required minimum selection of resources to the resource group, where the required minimum selection of resources comprises the at least one leased resource, and restarting the application workload on the allocated required minimum selection of resources of the resource group.   
     
     
         10 . The method according to  claim 7 , wherein responsive to a cluster manager for a particular node from among a plurality of nodes communicatively connected through a network allocating at least one leased resource for a resource group for an application workload on the particular node on fallover of the resource group from another node from among the plurality of nodes to the particular node, setting, by the cluster manager for the particular node, a timer thread to track an amount of time remaining for an initial lease period of the at least one leased resource further comprises:
 detecting, by the cluster manager for the particular node, an insufficient number of resources available for allocation for the resource group on the particular node, wherein the resources available for allocation do not comprise the at least one leased resource;   acquiring, by the cluster manager for the particular node, a license for leasing the at least one leased resource, in exchange for a fee, wherein the license requires that if the at least one leased resource is not released prior to the end of the first lease period then an additional fee is automatically incurred for the additional lease period; and   activating, by the cluster manager for the particular node, the at least one leased resource from an inactive state to an active state using the license.   
     
     
         11 . The method according to  claim 7 , wherein responsive to a cluster manager for a particular node from among a plurality of nodes communicatively connected through a network allocating at least one leased resource for a resource group for an application workload on the particular node on fallover of the resource group from another node from among the plurality of nodes to the particular node, setting, by the cluster manager for the particular node, a timer thread to track an amount of time remaining for an initial lease period of the at least one leased resource further comprises:
 communicatively connecting, by the cluster manager for the particular node, with a plurality of other cluster managers on the other nodes from among the plurality of nodes, wherein the plurality of nodes share at least one storage resource.   
     
     
         12 . The method according to  claim 7 , further comprising:
 responsive to setting the timer thread, monitoring, by the cluster manager for the particular node, whether the resource group releases the at least one leased resource; and   responsive to detecting the resource group releasing the at least one leased resource, canceling, by the cluster manager for the particular node,] the timer thread.   
     
     
         13 . The method according to  claim 7 , further comprising:
 detecting, by the cluster manager for the particular node, the timer thread expiring while the resource group is holding the at least one leased resource;   determining, by the cluster manager for the particular node, whether there is a selection of at least one other node from among the plurality of nodes with capability to handle one or more resource group requirements for the resource group;   responsive to detecting no selection of at least one other node with capability to handle the resource group requirements, maintaining, by the cluster manager for the particular node, the resource group on the particular node comprising the at least one leased resource for the additional lease period;   responsive to detecting the selection of the at least one other node with capability to handle the resource group requirements, sending, by the cluster manager for the particular node, a capacity request to a separate cluster manager on each of the selection of the at least one other node;   receiving, by the cluster manager for the particular node, a separate capacity response to each capacity request from each separate cluster manager, wherein each separate capacity response comprises a separate current capacity from one of the selection of the at least other node of the resources available for allocation;   calculating, by the cluster manager for the particular node, a particular cost associated with the particular node continuing to handle the resource group and a separate cost associated with the selection of the at least one other node handling the resource group with each separate current capacity;   responsive to the cluster manager calculating the particular cost is the lowest cost from among the particular cost and each separate cost, maintaining, by the cluster manager for the particular node, the resource group on the particular node comprising the at least one leased resource for the additional lease period; and   responsive to the cluster manager calculating another cost from among each separate cost is the lowest cost from among the particular cost and each separate cost, triggering fallover, by the cluster manager for the particular node, of the resource group from the particular node to a next node associated with the another cost from among the selection of the at least one other node and deactivating the at least one leased resource.   
     
     
         14 . The method according to  claim 7 , further comprising:
 responsive to maintaining the resource group on the particular node comprising the at least one leased resource for the additional lease period, triggering, by the cluster manager for the particular node, a message to a client interface indicating the lease of the at least one leased resource has been automatically extended for the additional lease period for the additional fee.