Autonomically suspending and resuming logical partitions when i/o reconfiguration is required

Abstract

A partition manager includes an I/O reconfiguration mechanism and a logical partition suspend/resume mechanism that work together to perform autonomic I/O reconfiguration in a logically partitioned computer system. When I/O reconfiguration is required, the affected logical partitions are suspended, the I/O is reconfigured, and the affected logical partitions are resumed. Because the logical partitions are suspended during I/O reconfiguration, any ghost packet that may occur when the I/O is reconfigured is ignored.

1 . An apparatus comprising:
 at least one processor;   a memory coupled to the at least one processor;   a plurality of logical partitions defined on the apparatus, the plurality of logical partitions comprising at least one logical partition that owns an identified Input/Output (I/O) resource and at least one logical partition that does not own the identified I/O resource; and   a partition manager residing in the memory and executed by the at least one processor and executing separately from the plurality of logical partitions, the partition manager performing the steps of:
 (1) detecting a hardware state in the computer system that requires reconfiguration of the identified I/O resource; 
 (2) suspending all of the plurality of logical partitions by inhibiting dispatch of tasks to all of the plurality of logical partitions and waiting until all pending tasks in all of the plurality of logical partitions are complete; 
 (3) reconfiguring the identified I/O resource; and 
 (4) resuming all of the plurality of logical partitions by enabling dispatch of tasks to all of the plurality of logical partitions. 
   
   
   
       2 . A computer readable program product comprising:
 (A) a partition manager executing separately from a plurality of logical partitions, the plurality of logical partitions comprising at least one logical partition that owns an identified Input/Output (I/O) resource and at least one logical partition that does not own the identified I/O resource, the partition manager performing the steps of:
 (1) detecting a hardware state in the computer system that requires reconfiguration of the identified I/O resource; 
 (2) suspending all of the plurality of logical partitions by inhibiting dispatch of tasks to all of the plurality of logical partitions and waiting until all pending tasks in all of the plurality of logical partitions are complete; 
 (3) reconfiguring the identified I/O resource; and 
 (4) resuming all of the plurality of logical partitions by enabling dispatch of tasks to all of the plurality of logical partitions; and 
   (B) recordable media bearing the partition manager.   
   
   
       3 . An apparatus comprising:
 at least one processor;   a memory coupled to the at least one processor;   a plurality of Input/Output (I/O) towers coupled to the apparatus via a plurality of I/O loops;   a plurality of logical partitions defined on the apparatus, the plurality of logical partitions comprising a first logical partition that owns an identified I/O resource in a first I/O loop and a second logical partition that does not own the identified I/O resource in the first I/O loop;   a partition manager residing in the memory and executed by the at least one processor, the partition manager managing the plurality of logical partitions and executing separately from the plurality of logical partitions, the partition manager performing the steps of:
 (1) detecting when the first I/O loop is unbalanced; 
 (2) quiescing I/O resources in the first loop; 
 (3) determining which of the plurality of logical partitions own I/O resources in the first loop; 
 (4) suspending the logical partitions determined in step ( 3 ); 
 (5) rebalancing the first I/O loop by allocating at least one I/O resource in the first loop from the first logical partition to the second logical partition; 
 (6) enabling the I/O resources in the first loop after rebalancing in step ( 5 ); and 
 (7) resuming the logical partitions suspended in step ( 4 ). 
   
   
   
       4 . A computer readable program product comprising:
 (A) a partition manager executing separately from a plurality of logical partitions, the partition manager performing the steps of:
 (1) detecting when at least one Input/Output (I/O) loop is unbalanced; 
 (2) quiescing I/O resources in the at least one I/O loop; 
 (3) determining which of the plurality of logical partitions own I/O resources in the at least one I/O loop; 
 (4) suspending the logical partitions determined in step ( 3 ); 
 (5) rebalancing the at least one I/O loop by allocating at least one I/O resource in the at least one I/O loop from a first logical partition to a second logical partition; 
 (6) enabling the I/O resources in the at least one I/O loop after rebalancing in step ( 5 ); 
 (7) resuming the logical partitions suspended in step ( 4 ); and 
   (B) recordable media bearing the partition manager.