International Business Machines Corporation

Pure Storage, Inc.

Cleversafe, Inc.

Andrew George Peake

Jason K Resch

Teague Scott Algie

Ilya Volvovski

Manish Motwani

Mark Duane Seaborn

Andrew D Baptist

Thomas Franklin Shirley, Jr

Thomas Darrel Cocagne

Sebastien Vas

Sanjaya Kumar

Bart Richard Clifone

Methods and systems for use in a storage network to prioritize storage units for data retrieval operations. In various examples, a device obtains resource utilization information and pending resource demand information for a plurality of storage units of a storage network. The device further groups, based on the resource utilization information, the storage units into an underutilized resource group and an overutilized resource group, and further issues one or more high priority read slice access requests to the storage units of the underutilized resource group. The read slice access requests correspond to a pending resource demand for data retrieval. In response to determining that the high priority read access information does not include at least a read threshold number of read slice requests, the device further issues one or more additional read slice requests to one or more storage units of the overutilized resource group.

Prioritizing storage units for data retrieval operations

Methods and systems for use in a storage network to prioritize storage units for data storage operations. In various examples, a device obtains data storage cost information relating to the storage network and resource utilization information for a plurality of storage unit sets. The device also determines pending resource demand information for storage units of each of the plurality of storage unit sets. The device further groups, based on at least one of the resource utilization information or the data storage cost information, the storage units into an underutilized resource group and an over utilized resource group and selects, based on at least one of the resource utilization information or the pending resource demand information, a storage unit set of the plurality of storage unit sets for storing new data. The device further issues high priority write access information to storage units of the selected storage unit set.

Prioritizing storage units for data storage operations

A method begins by a computing device of a dispersed storage network (DSN) selectively bringing online and taking offline storage units of a set of storage units of the DSN. When bringing a first storage unit of the set of storage units online in accordance with the selectively bringing online and taking offline storage units, the method continues with the computing device determining a rebuilding approach of the first storage unit. The method continues with the computing device bringing the first storage unit online in accordance with the rebuilding approach.

Rotating offline storage units in a dispersed storage network

A method includes detecting a storage error associated with a first memory device of a storage unit of a set of storage units, where data is error encoded into a set of encoded data slices and stored in a plurality of memory devices of the set of storage units, and where the plurality of memory devices includes the first memory device. The method further includes determining attributes associated with the first memory device and determining attributes of other memory devices of the plurality of memory devices. The method further includes selecting a memory device from the other memory devices based on the attributes of the memory device comparing favorably to the attributes associated with the first memory device. The method further includes rebuilding an encoded data slice associated with the storage error and storing the rebuilt encoded data slice in the selected memory device.

Maintaining failure independence for storage of a set of encoded data slices

Methods and apparatus for isolating the introduction of software defects in a dispersed storage network (DSN) are disclosed. In various embodiments, a search strategy is employed whereby after identifying a test failure in a current version of the memory software code, a sequence of interim versions of the code between the current version of the memory software code and a previous successfully tested version of the code is determined. A first version of the memory software code is selected from the sequence of interim versions (e.g., from the middle of or approximately in the middle of the sequence) and tested. When testing of the first version does not result in a test failure, a second version of the memory software code is tested, the second version selected from a sub-sequence of the sequence of interim versions between the first version of the code and the current version of the code.

Isolating the introduction of software defects in a dispersed storage network

A method begins by one or more processing modules of one or more computing devices of a dispersed storage network (DSN) receiving a delete data request for a data object of a set of data objects, where each data object of the set of data objects has substantially identical data to other data objects of the set of data objects and where only one plurality of sets of encoded data slices corresponding to one data object of the set of data objects is stored in memory of the DSN. The method continues with the one or more processing modules determining whether the data object is associated with a data ID and when the data object is associated with a data ID, identifying linking information associated with the data object. Linking information is updated to indicate that the set of data objects includes one fewer copy of the data object and then it is determined whether the Data ID is associated with a link-object and when the Data ID is associated with a link-object the link-object is deleted from the DSN memory.

Detection and correction of copy errors in a distributed storage network

A method for execution by a dispersed storage and task (DST) processing unit that includes a processor includes generating storage unit heat data based on a plurality of temperature readings received from each of a plurality of storage units, where the storage unit heat data indicates a first hot storage unit. A pair of storage units is selected from the plurality of storage units based on the storage unit heat data, where the pair of storage units includes the first hot storage unit and a second storage unit. A data swap request is generated for transmission to the pair of storage units, where the data swap request includes an instruction to transfer at least one first data slice from the first hot storage unit to the second storage unit and to transfer at least one second data slice from the second storage unit to the first hot storage unit.

Relocating storage unit data in response to detecting hotspots in a dispersed storage network

A dispersed storage and task (DST) processing unit receives a data access request corresponding to a data segment. Range availability information is obtained for a plurality of DST execution units of the DST network. A subset of the plurality of DST execution units is selected based on the range availability information and a threshold number corresponding to the data access request. Execution unit access requests are generated and sent to the subset of the plurality of DST execution units corresponding to a plurality of slices of the data segment, wherein the execution unit access requests include address information that is based on the range availability information.

Dispersed storage based on range availability and methods for use therewith

A method for execution by a computing device of a dispersed storage network. The method begins by determining whether frequency of access to a set of encoded data slices exceeds a frequently accessed threshold. The method continues, when the frequency of access exceeds the frequently accessed threshold, by determining an access amount indicative of a degree that the frequency of access exceeds the frequently accessed threshold. The method continues by generating a number of additional encoded data slices and storing the number of additional encoded data slices in a number of additional storage units, wherein the set of storage units and the number of additional storage units produce an expanded set of storage units. The method continues by sending a plurality of data access requests to subsets of the expanded set of storage units in a distributed manner to improve processing efficiency of the plurality of data access requests.

Generating additional slices based on data access frequency

A method begins with determining a memory device of a dispersed storage network (DSN) has an issue, where a data segment of a data object is dispersed storage error encoded to produce a set of encoded data slices that is stored in a set of storage units of the DSN, where a storage unit includes a plurality of memory devices that includes the memory device, and where the memory device stores an encoded data slice of the set of encoded data slices. The method continues by identifying an issue with the memory device and by identifying sets of encoded data slices. The method continues by generating an additional encoded data slice for each of the sets of encoded data slices to produce a group of encoded data slices and storing the group of encoded data slices in at least one other memory device of the DSN.

Expanding slice count in response to low-level failures

A method begins by a dispersed storage (DS) processing module of a dispersed storage network (DSN) obtaining status information from a set of distributed storage units (SUs) and determining that a plurality of the SUs are currently inactive. The method continues with the DS processing module selecting one or more inactive SUs for reactivation and issuing a request to change activation status to each. When a favorable response is received from the one more SUs, the method continues with the DDS processing module determining that encoded data slices (EDSs) stored in a first SU include errors and determining to rebuild the EDSs stored in the first SU that include one or more errors by issuing a request to change activation status to second SU that is currently inactive. Upon receiving a favorable response from the second SU, the EDSs with errors are rebuilt using EDSs from the second SU.

Rotating inactive storage units in a distributed storage network

A method begins by a computing device of a dispersed storage network (DSN) selectively bringing online and taking offline storage units of a set of storage units of the DSN. When bringing a first storage unit of the set of storage units online and taking a second storage unit of the set of storage units offline in accordance with the selectively bringing online and taking offline storage units, the method continues with the computing device determining a rebuilding approach of the first storage unit and a first to second storage unit transition. The method continues with the computing device bringing the first storage unit online in accordance with the rebuilding approach and the first to second storage unit transition and taking the second storage unit offline in accordance with the first to second storage unit transition.

Rotating offline DS units

A method begins by determining to rebuild a slice or slices to at least one of multiple distributed storage network (DSN) memory locations. The method continues by calculating a future risk estimation of each one of the multiple DSN memory locations, the future risk estimation including one or more risk factors. The method continues by determining a selection of which of the multiple DSN memory locations to rebuild and rebuilding the slice in the DSN memory of the selection.

Prioritizing rebuilding based on a longevity estimate of the rebuilt slice

A dispersed storage network facilitates isolating the introduction of software defects in dispersed storage units. A search strategy is employed whereby after identifying a test failure in a current version of the memory software code, a code version since a previous successfully tested version is identified. An interim version that represents the point at which approximately one half the changes were introduced is then tested. When there is a test failure, the next interim version selected for testing represents the point at which approximately one half the changes were introduced between the first interim version tested and the current version. If no failure, a next interim version is tested that represents the point at which approximately one half the changes were introduced between the previous successfully tested version and the first interim version tested.

Growing community of inventors

Chicago, IL, United States of America

Andrew George Peake