International Business Machines Corporation

Globalfoundries Inc.

Robin Han

Michael Paul Groover

Ming Zhao

Wei Tang

Yi Zhou

Edward Lin

Yan Su

Ronald David Martens

Todd Charles Sorenson

Trung Ngoc Nguyen

Brian Anthony Rinaldi

Enrique Q Garcia

Gary William Batchelor

Jay T Kirch

Veronica S Davila

Yang Liu

Franck Excoffier

Tony Jing Zhang

Andreas B M Koster

Mario Sweeney

Mario Kiessling

Mingzhi Zhao

Cindy K Hoac

Diana Y Ong

Provided are techniques for concurrent Input/Output (I/O) enclosure firmware/Field-Programmable Gate Array (FPGA) update in a multi-node environment. First notifications are sent to each I/O enclosure management engine on each of a plurality of server nodes that code activation for a first set of I/O enclosures is starting. An update image is distributed to the first set of I/O enclosures. The update image on the first set of I/O enclosures is activated by sending an activate reset command to each of the first set of I/O enclosures, wherein a reset is not propagated to other devices within each I/O enclosure in the first set of I/O enclosures in response to determining that the reset is an activate reset. In response to the activate reset command completing, second notifications are sent to each I/O enclosure management engine that code activation for the first set of I/O enclosures has completed.

Concurrent I/O enclosure firmware/field-programmable gate array (FPGA) update in a multi-node environment

Provided are a computer program product, system, and method for applying a machine learning algorithm to problem analysis in a code load operation of a data storage system. A code load driver is provided that receives code load operation information for an event during a code load operation in the storage system. The code load operation information indicates an error in the code load operation at a time of the event. A portion of the code load inform is processed by a machine learning to obtain a label indicating whether to halt the code load operation. In response to a label indicating the code load operation is to be halted, the code load operation is halted.

Applying machine learning in storage system code load problem analysis

Provided are techniques for code load processing. While performing code load processing of a set of modules of a same module type, it is determined that a first module in the set of modules is not in an operational state. It is determined that a second module is a redundant module for the first module. In response to determining that the second module is in an operational state and has already completed code update, the code load processing is continued. In response to determining that the second module is in an operational state and has not already completed code update, it is determined whether there is a third redundant module that is in an operational state. In response to determining that there is a third redundant module that is in an operational state, the code load processing is continued.

Cloud infrastructure for reducing storage facility code load suspend rate by redundancy check

A system for managing a code load for a storage system is disclosed. The system can include instantiating a code load. The code load can include a first update for a first component and a second update for a second component. The system can include monitoring the operational state of the first and second components in response to instantiating the code load. The system can also include determining to perform the first update in response to a triggering event. The system can also include performing the first update in response to determining to perform the first update.

Managing a code load

Provided are a computer program product, system, and method for applying a platform code level update to update a source partition in a computing system. Computational resources in the computing system are allocated to a destination partition. A code load is applied to the destination partition to implement an updated platform code level comprising an update to the platform code level on the source partition while the source partition is operational and processing computing requests. Submission of new transactions to the source partition is blocked in response to applying the code load to the destination partition. An operational environment and system state at the source partition are migrated to the destination partition in response to blocking submission of new transactions to the source partition. Transactions are directed to the destination partition intended for the source partition in response to migrating the operational environment and system state to the destination partition.

Applying a platform code level update to an operational node

A sequence for distributing at least one of a plurality of code packages to the at least one facility according to different states of a fixed state machine is set. The at least one of the plurality of code packages is maintained in at least one staging area in a valid, dormant mode while the fixed state machine is stopped. Pursuant to a resumption of the fixed state machine at a subsequent time, a current code package is swapped with the at least one of the plurality of code packages in the at least one staging area to activate the at least one of the plurality of code packages.

Updating elements in data storage facility using predefined state machine over extended time period

Provided are a computer program product, system, and method for performing code load operations on managed components in a system. A first node group comprising at least one computational node in the computer system performs code load operations for the managed components. Status of the code load operations at the managed component is written to a status data structure while the code load operations are being performed at the managed component. The first node group transfers control of the code load operations for the managed components to a second node group comprising at least computational node in the computer system while the code load operations are occurring at the managed components. A second node group reads the status data structure for the managed components to determine the status of the code load operations and continue the code load operations at managed components.

Performing code load operations on managed components in a system

A method and system for managing a code load for a storage system is disclosed. The method and system can include instantiating a code load. The code load can include a first update for a first component and a second update for a second component. The method and system can include monitoring the operational state of the first and second components in response to instantiating the code load. The method and system can also include determining to perform the first update in response to a triggering event. The method and system can also include performing the first update in response to determining to perform the first update.

A setup module organizes a single software image for a management command. A process module creates a plurality of processes independently executing the management command on each of the plurality of devices from a management console. Each process employs the software image. A termination module ends the management command after each process has completed on each of the plurality of devices.

Concurrent management console operations

Reducing storage facility code load suspend rate by redundancy check

Various embodiments for automated error recovery in a computing storage environment by a processor device are provided. In one embodiment, if a failure is caused by an operation of a hardware management console (HMC) and a malfunction of a current network connection, a cleanup operation is performed on at least a portion of a current HMC configuration, an alternative network connection to the current network connection is made, and a retry operation is performed.

Growing community of inventors

Lincoln, CA, United States of America

Robin Han