Recovery of guest virtual machines after failure of a host real machine

Claim 1

In a data processing system which includes storage, a channel subsystem having I/O devices attached thereto, at least one real machine executing a host operating system and a start interpretive execution instruction that initiates the operation of at least one virtual machine having a virtual processor executing a guest operating system, a method for providing for virtual machine survival following a system incident such as a machine check including a simulated machine check, such as instruction processing damage, which occurs when said host operating system does not successfully simulate an instruction of said guest operating system, and timer damage, or a channel check or a checkstop or a malfunction detected as part of an I/O operation such as an invalid checking block code detected in storage or in a real storage key, comprising the steps of: assigning a first and second nonvolatile area of said storage to said virtual machine by said host operating system in which said first area is in said virtual machine main storage area and said second area, which is an extension of said first area, contains control blocks that describe said virtual machine to said host operating system which controls said virtual machine in said first area of said storage, accessing said second area of said storage for requests made for said virtual machine prior to detection of said system incident, detecting said system incident, determining that said virtual machine is eligible for survival by detecting a bit set in a field in one of said control blocks in said second area of storage prior to said detection of said system incident, copying information about I/O devices dedicated to said virtual machine from said first area of said storage to said second area of said storage, recording the cureent status of said virtual machine in said second area of storage and verifying that said dedicated I/O devices will permit survival of said virtual machine, all for the purpose of maintaining the integrity of said virtual machine following said survival, terminating said real machine and suspending said virtual machine in response to the detection of said system incident, refreshing said host operating system in said real machine following the detection of said system incident and restoring said virtual machine in said first area of said storage with data in said second area of said storage, restoring said virtual machine using said I/O devices dedicated to said virtual machine in said first area of storage from said second area of said storage and restarting said virtual machine, and presenting the error that caused said system incident to the guest operating system of said virtual processor affected by said error once said virtual machine is restarted following said refreshing of said host operating system in said real machine in order to prevent said virtual machine from damaging its operating system.