Abstract: Programming methods and products are described for developing and producing installation and/or upgrade software for systems where multiple versions of software are available. Information is stored representing the version of software that is being installed, the version of software being installed having an updated feature. Groups of commands enable the transition from a present version of the software to another version of software having another version of the feature. An example of such a feature is deployment of logical volume partitions. A state machine is executed that consults the stored information to access one or more of the groups of commands. The one or more groups of commands is executed to transition from the present version of the software to the version of software that is being installed having the updated feature. If an error occurs during the execution of a group of commands, then recovery information including remediation commands is executed.

Abstract: Software configuration changes are made in a digital computer by executing selected opcodes to add new features. When executing each opcode, a persistent record is kept of execution of each elementary command of the opcode. To recover from a failure of execution of an opcode, remediation opcodes are executed for a known error, or else the persistent record of execution is accessed to identify executed commands of the opcode, and the executed commands are reverted in reverse order by undoing changes made to the software configuration by the executed commands of the opcode. The resulting configuration of the digital computer is compared to a backup copy of the configuration existing at the start of execution of the opcode. If the reversion of the executed commands of the opcode has not reproduced the backup configuration, then the configuration of the digital computer is restored to the backup configuration.

Abstract: Errors arising in logical volume partitions are detected and corrected automatically well before the errors cause a failure to access data. A database is created of configuration information about the logical volume partitions defined in physical volume headers, and the database is stored in storage separate from the headers. Thereafter the errors are detected by capturing configuration information about the logical volume partitions from the headers, and comparing the captured configuration information to the configuration information in the database. The errors are automatically corrected in the corrupt headers. If requested, a report is generated of the volumes and disk partitions that are inaccessible or missing or have the wrong size. The configuration check is performed periodically as a background task and during a re-boot after a system crash or interruption of a configuration process or when invoked prior to a recovery install or database restore.