Abstract: A method for managing data in a computer system comprises identifying a data-set that has data unit protection that is suspected of having been corrupted, suspending a data unit protection process for the data-set; deferring re-establishing the data unit protection process for the data-set until issuance of a request for an I/O operation for the data unit; and re-establishing the data unit protection process for the data-set in response to the request and before the I/O operation is conducted. The data has a data unit protection invalidity indicator for indicating that the data-set has data unit protection that is suspected of having been corrupted; and the data unit protection process is suspended in the presence of the data-set protection invalidity indicator. Meta-data protection that is suspected of having been corrupted may be used to identify a corrupted data-set.

Abstract: A method for managing data in a computer system comprises identifying a data-set that has data unit protection that is suspected of having been corrupted, suspending a data unit protection process for the data-set; deferring re-establishing the data unit protection process for the data-set until issuance of a request for an I/O operation for the data unit; and re-establishing the data unit protection process for the data-set in response to the request and before the I/O operation is conducted. The data has a data unit protection invalidity indicator for indicating that the data-set has data unit protection that is suspected of having been corrupted; and the data unit protection process is suspended in the presence of the data-set protection invalidity indicator. Meta-data protection that is suspected of having been corrupted may be used to identify a corrupted data-set.

Abstract: A scheduler for selecting a logical volume for scheduling pending I/O tasks for job generation. The scheduler includes a main scheduler that selects from among multiple sub-schedulers that operate according to different scheduling approaches, e.g., high priority read I/O request scheduling, urgency of non-read I/O tasks, as well as load balancing and round robin schemes. In selecting a sub-scheduler, the main scheduler considers the amount of time allocated to performing read I/O requests, as well as the level of urgency associated with the logical volumes and the disk adapter that manages those logical volumes for non-read I/O tasks.