International Business Machines Corporation

Xue Qiang Zhou

Yongjie Gong

Yang Liu

Jing Jing Zhang

Shuang Shuang Li

Mei Mei

Nicolas Marc Clayton

Gang Lyu

Chuan Li

Sarvesh S Patel

Kushal S Patel

Chao Li

Duo Chen

Shan Fan

Embodiments of the present disclosure relate to a new approach for workload balancing in a system. A computer-implemented method comprises obtaining a configuration of a system, wherein the configuration includes a connection relationship between components in the system. The method further comprises determining a workload threshold of a given component in the components based on the configuration, wherein a workload threshold of an upstream component of the given component is associated with a workload limit of the upstream component and a workload limit of the given component. The method further comprises determining a utilization rate of the given component based at least partially on a workload amount of the given component, the workload threshold of the given component, and the workload threshold of the upstream component. Accordingly, a more accurate utilization rate of each component in a system can be determined by considering the bottleneck of other components in the system, and thus workloads of various components in the system can be balanced from the perspective of the overall system.

Workload balancing to achieve a global workload balance

Method, system and computer program product are disclosed for an error recovery process. One method comprises: program instructions are executable by the processor to perform; determining a component of the computer system related to the error event; determining a quality of service requirement for the component; determining an error recovery process for the error event based on the quality of service requirement; and executing the error recovery process for the error event.

Error recovery process

The rebuilding process comprises the steps of detecting a failure on a first server that stores a first data and processes input/output requests from an application, redirecting incoming input/output requests from the application to a second server storing a second data. The second data being mirrored from the first data. The process further comprises forming, by the first server, a data loss information that identifies the lost data units in the first data. The process further comprises rebuilding, by the first server, the lost data units in the first data, by using the data that is retrieved from the second data and based on the data loss information by the second server. The process further comprises syncing up data units, other than the lost data units in the first data, with corresponding data in the second data in the second server during the rebuilding process.

Data rebuilding

In a hierarchical storage system, blocks of data selected for auto-tiering migration, are selected based on dynamically adjusted group sizes. Contiguous blocks are organized into default groups. I/O activity of the blocks in a group is monitored. Based on the I/O activity, the default groups may be sub-divided into smaller sub-groups or combined into larger groups, to separate as much as practical, contiguous series of cooler blocks and contiguous series of hotter blocks into respective focused (concentrated) groups or sub-groups. The concentrated group or sub-group may then be migrated according to the average I/O activity of the included blocks. Group configurations are continually and dynamically adjusted according to changing I/O conditions.

Automatic tiering of storage using dynamic grouping

A method for determining path health to conduct a plurality of Input-Output (IO) operations along a healthy path in a network is provided. The present invention may include receiving an IO request from a user and sending the received IO request on a first path. The present invention may include determining a first IO response has exceeded a threshold time on the first path. The present invention may include determining the first path has degraded based on the exceeded threshold time. The present invention may include generating a duplicate IO request and sending on a second path. The present invention may include receiving the duplicated IO response before receiving the original IO response. The present invention may include determining a health state associated with the slower path. The present invention may include refreshing a path state machine based on the determined health state associated with the slow path.

Handling path issues for storage copy services

The present disclosure relates to a method and an apparatus for failure detection in a storage system. One embodiment provides a method for failure detection in a storage system. The method includes determining an amount of data received by a plurality of switches in the storage system within a predetermined time window so as to obtain a plurality of data amounts; determining a count of check errors detected by the plurality of switches in the data to obtain a plurality of check error counts; and calculating a failure risk for the plurality of switches based on the plurality of data amounts and the plurality of check error counts. Another embodiment of the present invention provides a corresponding apparatus.

Method and apparatus for failure detection in storage system

Disclosed is a method, system and computer program product for forming a consistency group to define a dataset of point-in-time data to be copied from a source volume to a target volume; setting copy priorities for respective data in the dataset; and transmitting the respective data in the order of the copy priorities from the source volume to the target volume, wherein the copy priorities are dynamically adjusted in response to arrival of write data during transmission of the data.

Reducing write collisions in data copy

A method for improving I/O performance in synchronous data replication environments is disclosed. In one embodiment, such a method includes receiving write data into a primary write cache of a primary storage system. The method synchronously mirrors the write data from the primary write cache to a secondary write cache of a secondary storage system. The method is further configured to detect when the primary write cache is full. When the primary write cache is full, the method temporarily uses primary read cache of the primary storage system to store incoming write data. This incoming write data is mirrored from the primary read cache to the secondary write cache of the secondary storage system. A corresponding system and computer program product are also disclosed herein.

Performance in synchronous data replication environments

A method for efficiently managing extents of backup volumes in a tiered storage system is disclosed. In one embodiment, such a method includes identifying a lifespan associated with a backup volume. The lifespan indicates how long extents of the backup volume are retained in a tiered storage system before being released. The method further assigns a temperature value to the extents that determines where the extents are stored in the tiered storage system. The method utilizes the lifespan to determine the temperature value, where the temperature value is inversely proportional to the lifespan. The method places the extents on tiers of the tiered storage system in accordance with their temperature value, such that higher temperature extents are placed on higher tiers of the tiered storage system, and lower temperature extents are placed on lower tiers of the tiered storage system. A corresponding system and computer program product are also disclosed.

Comingling conventional and backup volumes in tiered storage systems

Performance-degrading hardware components in computer storage systems are identified. A timed-out input/output (I/O) operation is detected. Hardware components that include a plurality of types of hardware components involved with the timed-out I/O operation are identified. Involve counts of the hardware components are increased. A plurality of suspicious hardware components are identified based, at least in part, on a hardware hierarchy including one or more types of hardware components. There is at most one suspicious hardware component for each type of hardware component. A performance-degrading hardware component is identified from among the plurality of suspicious hardware components based, at least in part, on the involve counts and the hardware hierarchy. The performance-degrading hardware component is in a layer of the hardware hierarchy that is lower than layers including other suspicious hardware components.

Identifying performance-degrading hardware components in computer storage systems

Embodiments of the present disclosure relate to data management in a multitier storage system (MSS), the MSS comprises a storage virtualization controller (SVC) having at least one storage pool, and the storage pool comprises at least one logic volume, and the at least one logic volume comprises at least one tier. In response to the completion of a data migration from a source physical space of a first tier to a destination physical space of a second tier in a logic volume of a first storage pool, it is determined there is a free physical space in the source physical space. In response to the result of the determination indicating there exists a free physical space and the source physical space being thin provisioned, the free physical space of the source physical space is released to the storage array by the SVC.

Data management in a multitier storage system

A computer-implemented method, computer program product and system for dynamic granularity of a data storage by identifying tiers of data including units of data with a unit heat index, corresponding to an activity level metric and identifying a tier heat index for each of the tiers of data corresponding to an average of the unit heat index for the units of data within the tiers of data. Furthermore determining a granularity index based on a size of the units of data; calculating a pairing index by comparing the unit heat indexes from the corresponding tiers of data and calculating an efficiency index by comparing the unit heat indexes indicating a percentage of the of units of data within a first tier of data with a higher unit heat index than all of the units of data within a another tier of data with a lower tier heat index.

Dynamic granularity for storage optimization

A computer system can dynamically manage the performance of a data storage system that includes sets of data storage devices. One set of the data storage devices can include read-intensive (RI) solid-state drives (SSDs). The computer system can detect, in response to a specified number of input/output operations per second (IOPS) of data contained on the RI SSDs exceeding the access density of the RI SSDs divided by the portion of data locations of the RI SSDs containing the data, a free area within the RI SSDs having space for additional data. Additional data within the data storage system that has a required access time less than a specified access time threshold and an access frequency less than a specified access frequency threshold can also be detected. In response, the detected additional data can be moved to the free area within the RI SSDs.

Growing community of inventors

Shanghai, China

Xue Qiang Zhou