Seagate Technology Incorporated

Western Digital Technologies, Inc.

Robert Lester

Clark Edward Lubbers

Timothy G Lieber

Evan Richardson

Austin Striegel

Donald Penza

Bernd Lamberts

Remmelt Pit

Zarko Popov

Robert S Gittins

Todd McNally

Jared Tesone

An approach to identifying poorly performing data storage devices (DSDs) in a data storage system, such as hard disk drives (HDDs) and/or solid-state drives (SSDs), involves retrieving and evaluating a respective set of log pages, such as SCSI Log Sense counters, from each of multiple DSDs. Based on each respective set of log pages, a value for a Quality of Service (QoS) metric is determined for each respective DSD, where each QoS value represents an average percentage of bytes processed without the respective DSD performing an autonomous error correction. In response to a particular DSD reaching a predetermined threshold QoS value, an in-situ repair may be determined for the particular DSD or the particular DSD may be added to a list of candidate DSDs for further examination, which may include an FRPH examination for suitably configured DSDs.

Use of error correction-based metric for identifying poorly performing data storage devices

An approach to identifying problematic data storage devices, such as hard disk drives (HDDs), in a data storage system involves retrieving and evaluating a respective recovery log, such as a media error section of a device status log, from each of multiple HDDs. Based on each recovery log, a value for a Full Recoveries Per Hour (FRPH) metric is determined for each read-write head of each respective HDD. Generally, the FRPH metric characterizes the amount of time a head has spent performing recovery operations. In response to a particular head FRPH reaching a pre-determined threshold value, an in-situ repair can be determined for the HDD in which the head operates. Similarly, in the context of solid-state drives (SSDs), a latency metric is determinable based on time spent waiting on resolving input/output (IO) request collisions, on which an in-situ repair can be based.

Use of recovery behavior for prognosticating and in-situ repair of data storage devices

A dynamic scalable error correction coding (ECC) scheme for a data storage system involves a system controller predicting a type and/or amount of ECC needed to reconstruct data to be stored on a particular data storage device(s) based on operational data integrity information accessed from the array of data storage devices. Thus, redundancy does not need to be allocated unless required. The devices may be logically grouped into subsets according to common characteristics, whereby the prediction made for a device in a subset may be based on the data integrity information from that subset, as well as from other relevant subsets.

System-level error correction coding allocation based on device population data integrity sharing

To provide enhanced operation of data storage devices and systems, various systems, apparatuses, methods, and software are provided herein. In a first example, a data storage system is presented. The data storage system includes data storage devices configured for storage and retrieval of data, power supply modules configured to provide power to at least the plurality of data storage devices, and a management controller configured to monitor power properties associated with the power supply modules to detect a degraded power condition. Responsive to the degraded power condition, the management controller is configured to establish adjustments to performance of the data storage devices to mitigate the degraded power condition.

Enhanced power control of data storage arrays

A data storage system and associated method are provided wherein a policy engine continuously collects qualitative information about a network load to the data storage system in order to dynamically characterize the load and continuously correlates the load characterization to the content of a command queue of transfer requests for writeback commands and host read commands, selectively limiting the content with respect to writeback commands to only those transfer requests for writeback data that are selected on a physical zone basis of a plurality of predefined physical zones of a storage media.

Preferred zone scheduling

As apparatus and associated method for a dual active-active array storage system with a first controller with top level control of a first memory space and a second controller with top level control of a second memory space different than the first memory space. A seek manager residing in only one of the controllers defines individual command profiles derived from a combined list of data transfer requests from both controllers. A policy engine continuously collects qualitative information about a network load to both controllers to dynamically characterize the load, and governs the seek manager to continuously correlate each command profile in relation to the load characterization.

Request priority seek manager

A data storage system and associated method implement a HASTE with a policy engine that continuously collects qualitative information about a network load to the data storage system in order to dynamically characterize the load, and continuously correlates a command profile to a data storage device of the data storage system in relation to the characterization.

Host adaptive seek technique environment

Apparatus and method for command queue management of back watered requests. A selected request is released from a command queue, and further release of requests from the queue is interrupted when a total number of subsequently completed requests reaches a predetermined threshold.

Command queue management of back watered requests

Apparatus and method for efficiently arranging and searching data in a memory space, such as a cache memory of a data storage array controller. A data structure comprises a skip list of nodes having an array of forward pointers. Each node has a node level derived from an index at which the node is stored in a table in the memory space. The total available number of nodes is preferably selected to be less than half of the largest power of 2 that can be expressed by a number of bits of the index, and the nodes are preferably stored at only even or only odd indices of the table. In such case, a free list of nodes is preferably generated from an array of pairs of counts and indices to identify the available nodes. Additional table structures can further be provided to enhance data arrangement and searching functions.

Skip list with address related table structure

A data storage system and associated method are provided wherein a policy engine continuously collects qualitative information about a network load to the data storage system in order to dynamically characterize the load and continuously correlates the load characterization to a command queue depth of data transfer requests.

Command queue loading

A plurality of data storage devices are arranged to form a multi-device array space, and a controller controls access to the array space. The controller is configured to index sets of operational performance data from the plurality of data storage devices into a unified data log. The controller further detects a failure trend of one or more data storage devices by analyzing at least one of a vertical block of data and a horizontal block of data in the unified data log. The vertical data block of data is associated with one of the data storage devices at a plurality of indices. The horizontal block of data is associated with a plurality of the data storage devices at an index.

Growing community of inventors

Colorado Springs, CO, United States of America

Robert Lester