International Business Machines Corporation

Manish Misra

Ravi K Kulkarni

Preetha R Kondajeri

John Daniel Upton

Joel Gerard Goodwin

Suparna Sinha Roy

Each of a plurality of partitions within a logical partitioned data processing system is configured for an inter partition communication area (IPCA) allocated from partition's own system memory. Each partition's IPCA combined together forms a non-contiguous block of memory which is treated as a virtual shared resource (VSR). Access to VSR is controlled by hypervisor to maintain data security and coherency of the non-shared resources of a partition. Messages are written to and read from VSR under a specific partition's IPCA for inter partition communication. No physical shared or non-shared resources are involved during inter partition communication, hence no extra overhead on those resources, thus achieving optimized performance during inter partition communication.

Inter partition communication within a logical partitioned data processing system

A computer system for performing error monitoring of partitions. A partition status buffer (PSB) denotes a status (GOOD, BAD, NOCARE) of each partition of at least two partitions. The BAD status denotes that the partition has encountered at least one error that is currently unrepaired. A global supervisor mapping (GSM) associates each partition (designated as a supervised partition) with a supervisor partition in a one-to-one mapping. The supervisor partition determines its supervised partition from the GSM and ascertains the status of its supervised partition from the PSB. If the status of the supervised partition is BAD then the supervisor partition performs a recovery procedure. The recovery procedure: obtains a grant of access to physical and logical resources of the supervised partition which contains error data of the supervised partition; gathers the error data; sets the status of the supervised partition to the NOCARE status.

Computer system for performing error monitoring of partitions

A computer program product and computer system for error monitoring partitions in a computer system. Provided to each partition is a partition status indicator (PSI) denoting a RUNNING or FAIL status of the partition, and an error log area (ELA) for storing partition error entries. The ELA includes a partition identifier, an entry status indicator (ESI) indicating READ/UNREAD status for the error entry, and an error identifier. An error procedure performed for each first partition whose partition status indicator indicates the FAIL status includes: copying each error entry in the ELA of the first partition whose ESI indicates the UNREAD status into the ELA of a second (running) partition; setting the ESI to the READ status for each copied error entry in the ELA of the first partition; and having the ESI set to the UNREAD status for each copied error entry in the ELA of the second partition.

Error monitoring of partitions in a computer system using partition status indicators

A computer program product for error monitoring partitions in a computer system. A global supervisor mapping (GSM) associates each supervised partition with a supervisor partition that monitors the supervised partition. A partition status buffer (PSB) denotes a status (GOOD, BAD, NOCARE) of the partition. The BAD status denotes that the partition has encountered at least one error that is currently unrepaired. The supervisor partition determines its supervised partition from the GSM and ascertains the status of its supervised partition from the PSB. If the status of the supervised partition is BAD then a recovery procedure is performed by the supervisor partition. The recovery procedure: obtains a grant of access to physical and logical resources of the supervised partition which contains error data of the supervised partition; gathers the error data; and sets the status of the supervised partition to the NOCARE status.

Error monitoring of partitions in a computer system using supervisor partitions

A method and computer program product for error monitoring partitions in a computer system. A partition status buffer (PSB) denotes a status (GOOD, BAD, NOCARE) of each partition of at least two partitions. The BAD status denotes that the partition has encountered at least one error that is currently unrepaired. A global supervisor mapping (GSM) associates each partition (designated as a supervised partition) with a supervisor partition in a one-to-one mapping. The supervisor partition determines its supervised partition from the GSM and ascertains the status of its supervised partition from the PSB. If the status of the supervised partition is BAD then the supervisor partition performs a recovery procedure. The recovery procedure: obtains a grant of access to physical and logical resources of the supervised partition which contains error data of the supervised partition; gathers the error data; sets the status of the supervised partition to the NOCARE status.

Method and computer program product for error monitoring of partitions in a computer system using supervisor partitions

A method, computer system, and computer program product for error monitoring partitions in the computer system. Provided to each partition is a partition status indicator (PSI) denoting a RUNNING or FAIL status of the partition. During booting, it is ascertained whether each partition has the RUNNING or FAIL status, and the PSI of each partition is accordingly set to the RUNNING status or the FAIL status, respectively. An error procedure performed for each first partition whose partition status indicator indicates the FAIL status includes: copying each error entry in the ELA of the first partition whose ESI indicates the UNREAD status into the ELA of a second (running) partition; setting the ESI to the READ status for each copied error entry in the ELA of the first partition; and having the ESI set to the UNREAD status for each copied error entry in the ELA of the second partition.

Method, system, computer program product, and computer system for error monitoring of partitions in the computer system using partition status indicators

A method for inter partition communication within a logical partitioned data processing system where each partition is configured for an inter partition communication area (IPCA) allocated from partition's own system memory. Each partition's IPCA combined together forms a non-contiguous block of memory which is treated as a virtual shared resource (VSR). Access to VSR is controlled by hypervisor to maintain data security and coherency of the non-shared resources of a partition. Messages are written to and read from VSR under a specific partition's IPCA for inter partition communication. No physical shared or non-shared resources are involved during inter partition communication, hence no extra overhead on those resources, thus achieving optimized performance during inter partition communication.

Method for inter partition communication within a logical partitioned data processing system

A method and computer program product for error monitoring partitions in a computer system. A global supervisor mapping (GSM) associates each supervised partition with a supervisor partition that monitors the supervised partition. A partition status buffer (PSB) denotes a status (GOOD, BAD, NOCARE) of the partition. The BAD status denotes that the partition has encountered at least one error that is currently unrepaired. The supervisor partition determines its supervised partition from the GSM and ascertains the status of its supervised partition from the PSB. If the status of the supervised partition is BAD then a recovery procedure is performed by the supervisor partition. The recovery procedure: obtains a grant of access to physical and logical resources of the supervised partition which contains error data of the supervised partition; gathers the error data; and sets the status of the supervised partition to the NOCARE status.

A method and computer program product for error monitoring partitions in a computer system. Provided to each partition is a partition status indicator (PSI) denoting a RUNNING or FAIL status of the partition, and an error log area (ELA) for storing partition error entries. The ELA includes a partition identifier, an entry status indicator (ESI) indicating READ/UNREAD status for the error entry, and an error identifier. An error procedure performed for each first partition whose partition status indicator indicates the FAIL status includes: copying each error entry in the ELA of the first partition whose ESI indicates the UNREAD status into the ELA of a second (running) partition; setting the ESI to the READ status for each copied error entry in the ELA of the first partition; and having the ESI set to the UNREAD status for each copied error entry in the ELA of the second partition.

A method, system, and product are disclosed for isolating a defect in a memory system by determining in which particular component of the memory system the defect exists. The memory system includes multiple components. The components include a physical memory module, a memory card to which the physical memory module is attached, and a memory controller for controlling the memory card. The memory card includes one or more electrical buffers for driving or detecting the memory signals. The buffers may be used as virtual memory elements. Each component is tested separately in order to identify the defective component with the help of virtual memory system elements. The components are tested by first testing the physical memory module. If the physical memory module passes the test, the memory card is then tested. If the memory card passes its test, the memory controller is tested.

Method, system, and product for isolating memory system defects to a particular memory system component

A method, apparatus, and computer instructions for backing up data in a nonvolatile random access memory. Selected data is stored in the nonvolatile random access memory in available space in at least one electrically erasable programmable read only memory in the data processing system to form stored data. The stored data is used to restore the nonvolatile random access memory if nonvolatile random access memory becomes corrupted.

Growing community of inventors

Bengaluru, India

Manish Misra