Resource management for parent child workload

1. An apparatus for resource management for a parent child workload comprising:

a processor;

a computer-readable storage media storing code executable by the processor to:

organize a plurality of processes into a plurality of process groups, wherein each process group comprises a given parent process and all child processes of the given parent process, and wherein each process group has a process level;

calculate a process cost for each process group by querying each child process for a process history, determining a run time for each process history, and calculating the process cost as a product of the run time and a child task count, wherein the child task count estimates a number of child processes that are executed to complete the process group and the process cost approximates computing resources required to finish a process group;

determine a process level for each process group, wherein a subprocess of a process group has a lower process level in a process hierarchy;

assign a process priority to each process group based on computing resources required to finish the process group as a function of a rank of the process cost for the process group, wherein a process group with a highest process cost is assigned a highest process priority;

iteratively assign computing resources to subgroups with lower process levels of a given process group with a highest process priority at the given process level;

detect a shortage of computing resources for assignment to subgroups of the given process group with the highest process priority at the given process level;

in response to detecting the shortage of computing resources, terminate a lower process priority process group;

reassign freed computing resources to a higher process priority process group; and

in response to the computing resources being available for a reassignment time interval, automatically restart the terminated lower process priority process group.

2. The apparatus of claim 1 , wherein the process priority is an inverse function of the process cost.

3. The apparatus of claim 1 , wherein the process cost is calculated as a product of a square root of the run time and the child task count.

4. A method for resource management for a parent child workload comprising:

organizing, by use of a processor, a plurality of processes into a plurality of process groups, wherein each process group comprises a given parent process and all child processes of the given parent process, and wherein each process group has a process level;

calculating a process cost for each process group by querying each child process for a process history, determining a run time for each process history, and calculating the process cost as a product of the run time and a child task count, wherein the child task count estimates a number of child processes that are executed to complete the process group and the process cost approximates computing resources required to finish a process group;

determining a process level for each process group, wherein a subprocess of a process group has a lower process level in a process hierarchy;

assigning a process priority to each process group based on computing resources required to finish the process group as a function of a rank of the process cost for the process group, wherein a process group with a highest process cost is assigned a highest process priority;

iteratively assigning computing resources to subgroups with lower process levels of a given process group with the highest process priority at a given process level;

detecting a shortage of computing resources for assignment to subgroups of the given process group with the highest process priority at the given process level;

in response to detecting the shortage of computing resources, terminating a lower process priority process group;

reassigning freed computing resources to a highest process priority process group; and

in response to the computing resources being available for a reassignment time interval, automatically restarting the terminated lower process priority process group.

5. The method of claim 4 , wherein the process priority is an inverse function of the process cost.

6. The method of claim 4 , wherein the process cost is calculated as a product of a square root of the run time and the child task count.

7. A computer program product comprising a computer-readable storage medium having program instructions embodied therewith, the program instructions readable/executable by a processor to cause the processor to:

organize a plurality of processes into a plurality of process groups, wherein each process group comprises a given parent process and all child processes of the given parent process, and wherein each process group has a process level; calculate a process cost for each process group by querying each child process for a process history, determining a run time for each process history, and calculating the process cost as a product of the run time and a child task count, wherein the child task count estimates a number of child processes that are executed to complete the process group and the process cost approximates computing resources required to finish a process group;

determine a process level for each process group, wherein a subprocess of a process group has a lower process level in a process hierarchy;

assign a process priority to each process group based on computing resources required to finish the process group as a function of a rank of the process cost for the process group, wherein a process group with a highest process cost is assigned a highest process priority;

iteratively assign computing resources to subgroups with lower process levels of a given process group with a highest process priority at the given process level;

detect a shortage of computing resources for assignment to subgroups of the given process group with the highest process priority at the given process level;

in response to detecting the shortage of computing resources, terminate a lower process priority process group;

reassign freed computing resources to a higher process priory process group; and

in response to the computing resources being available for a reassignment time interval, automatically restart the terminated lower process priority process group.

8. The computer program product of claim 7 , wherein the process priority is an inverse function of the process cost.

9. The computer program product of claim 7 , wherein the process cost is calculated as a product of a square root of the run time and the child task count.