Abstract: A device input/output (“I/O”) monitoring mechanism serving as an interface between a computer operating system kernel and a device driver which obviates the necessity of implementing specific pseudo-device drivers for various peripheral devices and provides a standard interface between, for example, computer mass storage devices and a computer operating system. The I/O monitoring mechanism disclosed is of especial utility in the measurement of general storage device I/O performance and allows I/O statistics to be presented to application-level software operating in conjunction with a computer operating system which facilitates the identification of storage device “hot spots” within metadevices and sub-metadevices. The mechanism of the present invention is usable across a wide range of storage device peripheral hardware and device drivers and allows user-level program code to readily determine the associated physical sub-component.

Abstract: A device input/output (“I/O”) monitoring mechanism serving as an interface between a computer operating system kernel and a device driver which obviates the necessity of implementing specific pseudo-device drivers for various peripheral devices and provides a standard interface between, for example, computer mass storage devices and a computer operating system. The I/O monitoring mechanism disclosed is of especial utility in the measurement of general storage device I/O performance and allows I/O statistics to be presented to application-level software operating in conjunction with a computer operating system which facilitates the identification of storage device “hot spots” within metadevices and sub-metadevices. The mechanism of the present invention is usable across a wide range of storage device peripheral hardware and device drivers and allows user-level program code to readily determine the associated physical sub-component.

Abstract: A system, method and computer program product comprising a storage manager independent configuration interface translator which requests an opaque listing of the available storage devices from an associated metadisk driver and determines a subset of the listing meeting a preselected search criteria. The resultant opaque listing is then converted to a non-opaque format listing for presentation to a user of the computer system in a desired format such as through a command line or graphical user interface. A notification mechanism is also provided that presents information about storage devices which have in some way changed while in use.

Abstract: Apparatus and programmed method is disclosed, providing for the creation of a circular file of use in conjunction with the UNIX.TM. operating system or other layered and modular operating system architectures. The circular file uses an inode field to identify the file as circular and writes fill data to the file to allocate physical blocks to the file at the time of its creation. Thereafter, blocks already allocated to the file are re-allocated when more file space is needed and the file size does not increase.

Abstract: A device input/output ("I/O") monitoring mechanism serving as an interface between a computer operating system kernel and a device driver which obviates the necessity of implementing specific pseudo-device drivers for various peripheral devices and provides a standard interface between, for example, computer mass storage devices and a computer operating system. The I/O monitoring mechanism disclosed is of especial utility in the measurement of general storage device I/O performance and allows I/O statistics to be presented to application-level software operating in conjunction with a computer operating system which facilitates the identification of storage device "hot spots" within metadevices and sub-metadevices. The mechanism of the present invention is usable across a wide range of storage device peripheral hardware and device drivers and allows user-level program code to readily determine the associated physical sub-component.

Abstract: A metadisk driver functionally intermediate a computer operating system and one or more metadrivers and underlying layered drivers provides a driver rename/exchange function which does not depend on any particular driver having knowledge of the private data structures of any of the other drivers. The rename/exchange technique implemented thereby may be conducted while the underlying devices are on-line and comprise atomic operations which are, therefore, recoverable inasmuch as the operation will have either been completed or will not be committed in the event of any interruption.

Abstract: A system and method for file system fix-on-panic for a computer operating system which comprises an enhancement to the UNIX.RTM. File System ("UFS") that increases total system availability by detecting file system errors and determining whether on-line repair is possible and then locking, repairing and unlocking the failed file system. Availability of the entire computer system is increased since the mean time to failure for independent threads is increased by the amount of time up to the next failure and the mean time to repair for dependent threads is reduced to only the amount of time necessary for the repair. The system and method disclosed allows for repairs to be made during use, may be called from the user level and allows for blocking of only particular threads.

Abstract: Various versions of a computer file are provided without requiring copying the file or logging changed data, so that the files have consistent user data. A program module responds to a system call argument and allocates another node in file system tables and copies metadata information from the old node into the new node, so that both nodes contain the same data block allocation information. Portions of the module set shadow pointers in the old node to point to the new node, and in the new node to point to the old node. Changes to data are made with respect to the new node and fresh physical data blocks are allocated for the changed blocks. A method includes the steps of allocating a new node for storing file allocation information for another version of an existing file; logically connecting the new node to the existing node for file name and other purposes; and replacing node identification information in the directory with the new mode identification.