International Business Machines Corporation

Benedict Michael Rafanello

Mark Alan Peloquin

Cuong Huu Tran

Cristi Nesbitt Ullmann

Steven L Pratt

Barry Lowell Russell

Vincent Brian Rapp

John Cameron Stiles

A mechanism is provided by which a logical volume manager may allow features to be added and modified without having to modify the logical volume manager code. The present invention provides an logical volume manager engine, which attempts to load feature plug-ins that are stored as dynamic link libraries. The logical volume manager also includes application program interfaces to make it possible for user interfaces and other programs that communicate with the LVM Engine to communicate with features created by third parties. The logical volume manager is also modified to eliminate assumptions about the number and order of features being applied to the volume. Features may be selected and ordered by a user at the time of volume creation.

Method and an apparatus for logical volume manager plug-ins

Logical disks are provided, which are created by device managers. The logical volume manager is modified to use only logical disks. The device managers may be plug-in modules. This allows new device managers to be added to the logical volume manager at any time without changing the code in the logical volume manager. Anything that can be made to appear as a logical disk through the use of a device manager plug-in may then be used with the logical volume manager. A device manager for network attached storage may be written allowing network attached storage devices to be treated as local disk drives by the logical volume manager, thereby allowing all of the capabilities of the logical volume manager to be used with the network attached storage devices.

Method and an apparatus to extend the logic volume manager model to allow device management plug-ins

A database is employed for the recording of assigned computer disk drive letters, but database is distribute across multiple drives and media, rather than centralizing the database as prior art systems have done. Fragments of the database are stored within the unused portion of disk space located after partition tables but before the beginning of a partition, thereby forming one or more Shadow Partition Tables. In these Shadow Partition Tables, the drive letter assignments or logical volume assignments for the particular piece of media and partition are stored. A method for automatically determining relationship between entries in the normal partition tables and the Shadow Partition Tables is employed which is compatible with industry standard disk format definitions, while allowing enhanced drive letter assignment management especially useful for removable disk media.

Sticky drive letters for computer media

A method and apparatus to dynamically order features and manage features, especially aggregators, during creation of a logical volume is provided. The method and apparatus make use of a partition/aggregate list to identify partitions and/or aggregates that make up a logical volume that is to be created. In addition, the partition/aggregate list identifies features to be applied to the partitions and/or aggregates as well as the order in which these features are to be applied. The order in which these features are to be applied is designated by a current feature indicator (CFI). Based on a current CFI count of an entry in the partition/aggregate list, application of a feature corresponding to the current CFI count is attempted. If successful, the current CFI count is incremented and the process returns to a first entry in the partition/aggregate list. If unsuccessful, the current CFI count is not incremented and the process continues on to the next entry in the partition/aggregate list.

Method and an apparatus for volume creation in the presence of multiple aggregators

A method and apparatus to dynamically order features and resolve conflicts in a logical volume management environment is provided. The method and apparatus classifies features of a logical volume into partition level, aggregate level and volume level classes. Based on these classes and the attributes associated with each feature, ordering of the features in a feature stack is performed and conflicts between features identified for conflict resolution. In addition, the apparatus and method provides a mechanism by which a default ordering of features selected by a user may be generated for a logical volume. The user may accept this default ordering or edit the ordering as long as the user does not generate any conflicts. Any conflicts generated will be reported to the user and the attempted ordering will not be allowed.

Method and an apparatus to dynamically order features and to resolve conflicts in a multiple-layer logical volume management environment

A method and an apparatus for enabling and disabling features in a logical volume management environment is presented. This operation can be performed at boot time or at run time. The user is shown a list of features as well as the current state for each feature. If the feature is currently enabled, the user can select to disable the feature, provided this is a safe operation given the current features for the volume. If the feature is disabled, the user has the following choices: re-enable, disable and retain in the feature stack, disable and remove from the feature stack, and permanently remove the disabled feature from the volume. Disabling or re-enabling features can be very useful to accommodate a changing operating environment. For example, the feature Volume Mirroring allows a replication of the volume at a remote site. If due to network difficulties that remote site becomes inaccessible, then this feature would have to be disabled. At a later time when the site becomes accessible again, the feature can be re-enabled.

System and method for selectively enabling and disabling plug-ins features in a logical volume management enviornment

A system and method for control of data and boot records associated with multi-layer logical volumes allows the logical volume data area containing the fake EBR to be deterministically and efficiently accessed by using a &ldquo;top down&rdquo; control methodology. Control of the logical volume manager (&ldquo;LVM&rdquo;) data area is distributed. An LVM data area is placed on each aggregate as well as on each partition, and the fake EBR is placed in the LVM data area of the topmost aggregate. The topmost aggregator controls access to the fake EBR, and only this aggregator must provide API functionality to access the fake EBR. This allows the fake EBR to be found and accessed only by new software applications which are &ldquo;aware&rdquo; of the fake EBR, while allowing legacy aggregators to continue to function correctly and allowing legacy software applications to continue to interwork with legacy aggregators through their unchanged APIs.

Top-down control of multiple layer aggregation logical volume management data and boot record

A system and method for a multi-layer logical volume management (&ldquo;LVM&rdquo;) system which extends the single-layer logical volume model of current technology to handle multiple levels of aggregation more efficiently. Multiple levels of aggregation allows multiple types of aggregators, such as drive linking, mirroring, and software RAID, to be used together to bypass the limitations inherent in each individual aggregation technology. An LVM data area which stores information about the organization of the multi-layer logical volume is stored within the last partition of the logical volume. A fake EBR, which describes the logical volume as if it were a single disk drive, is located in the LVM data area of one of the partitions in the logical volume. A &ldquo;bottom-to-top&rdquo; control methodology is used to locate the LVM data area containing the fake EBR, including setting a special flag in the LVM data area containing the fake EBR. Aggregators search for this flag when creating aggregates, and set a flag in the LVM data area of the aggregate being created if any of the items being aggregated has this flag set. Thus, a path is built from the topmost aggregate to the partition containing the fake EBR. This path allows I/O requests to the EBR to be intercepted and easily redirected to the fake EBR, thereby improving the efficiency with which this is executed in comparison to previous methods.

Control of multiple layer aggregation logical volume management data and boot record

A database distributed across multiple drives and media, rather than being centralized. Fragments of the database are stored within the unused portion of disk space located after partition tables but before the beginning of a partition, thereby forming one or more hidden databases. The method and system disclosed is compatible with industry standard disk format definitions.

Method and system for a distributed hidden database across a plurality of computer media

A data processing system and method are disclosed for storing logical volume information used by a logical volume manager to create and manage a logical volume. The data processing system includes a plurality of storage devices. A plurality of physical partitions within the plurality of storage devices are specified. Each one of the physical partitions includes a plurality of sectors. The specified physical partitions are logically combined to create the logical volume. A first plurality of the sectors are reserved in each of the physical partitions for the logical volume information. The first plurality of sectors are located at the top of each of the physical partitions. The logical volume information is stored in the first plurality of sectors in each of the physical partitions. In this manner, the logical volume information is stored at the top of each of the physical partitions.

Data processing system and method including a logical volume manager for storing logical volume data

An enhancement to existing logical volume managers which provides the functionality to expand logical volumes employing multiple levels of aggregation. It includes functionality to determine if a logical volume employing multiple levels of aggregation can be expanded. If it can be expanded, the aggregator or feature which will perform the expansion is determined. Once this has been done, the features and aggregators that need to be applied to the partitions being added to the volume are determined. These features and aggregators are then applied to the partitions being added to the volume, with user input as required. Finally, the aggregator which will perform the actual expansion is called, the volume is expanded, and all of the remaining features and aggregators on the volume are given the opportunity to update their respective data as appropriate to accommodate the new size of the volume

System and method for volume expansion in the presence of multiple plug-in features

A programable apparatus and method for assigning drive letters without conflicts to resource devices in a data processing system based upon three signature numbers called a resource serial number, a boot drive serial number and a sequence number, where the signature numbers are assigned by the data processing system's operating system to the resource devices.

Foreign drive determination and drive letter conflict resolution

A method for routing an input/output request for a particular logical volume. In a preferred embodiment, partitions are assigned to logical volumes, wherein an arbitrary number of partitions can be assigned to a logical volume and the partitions can be of an arbitrary size. Each logical volume is divided into a plurality of zones, wherein the zones for a particular logical volume are substantially equal in size. A zone table is created whose entries contain data about one or two partitions that correspond to that zone. Responsive to receiving an input/output request directed to a logical volume, the zone table entries for a first and a last sector listed in the input/output request are used to determine to which partitions in the logical volume the input/output request should be routed.

Method and apparatus for improving performance of drive linking through use of hash tables

Where a number n of read attempts are required to successfully read a data sector, with the first n-1 attempts returning a disk drive read error, the number of attempts required is compared to a predefined threshold selected to indicate that the sector is unreliable and is in danger of becoming completely unrecoverable. If the threshold number of attempts is not exceeded, the sector is presumed to still be good and no further action need be taken. If the threshold number of attempts was equaled or exceeded, however, the unreliable or failing sector is relocated to a reserved replacement sector, with the recovered data written to the replacement sector. The failing data sector is remapped to the replacement sector, which becomes a fully functional substitute for the failing sector for future reads and writes while preserving the original user data. Data within a failing sector is thus preserved before the sector becomes completely unrecoverable.

Recovering and relocating unreliable disk sectors when encountering disk drive read errors

A high performance file system (HPFS) is validated by checking the directory entries to the extent possible, saving the information required to check F-Nodes and queuing detected errors together with corrective actions. The F-Nodes are then checked in order, minimizing the head motion required to read the F-Nodes and also reducing the I/O time required for the validation process. The detected errors may then be processed, corrective action taken, and the affected DIRBLK written. To reduce the amount of memory required, the directory structure of an HPFS storage device may be processed employing a breadth-first, level-by-level approach. Checking of the directory entries may be further segmented into multiple threads to take advantage of the ability of RAID systems to issue multiple read requests.

Growing community of inventors

Round Rock, TX, United States of America

Benedict Michael Rafanello