International Business Machines Corporation

E8 Storage Systems Ltd.

Amazon Technologies, Inc.

Eb Storage Systems Ltd.

Alex Friedman

Constantine Gavrilov

Evgeny Budilovsky

Barak Pinhas

Artemy Voikhansky

Alex Liakhovetsky

Lior Chen

Alex Snast

Aharon Novogrodski

Yoni Divinsky

Itay Gonshorovitz

Sergei Dyshel

Ofir Dahan

Andrew Boyer

Swapnil Vinay Dinkar

Yonatan Divinsky

Hen Guetta

A storage object and an associated permissions record is stored at a storage server. The permissions record indicates that some storage consumers are not permitted to perform a type of I/O operation on the storage object. In response to detecting that an event of a deletion triggering type with respect to the records, a modified version of the permissions record is stored at the server, indicating that the storage consumers remain prohibited from performing the I/O operations. In response to receiving a command to perform a particular I/O at the server after the modified version has been stored, the modified version is used to process the command.

Reservation persistence in distributed block storage systems

A thread of a storage server determines that a command to modify a reservation record which comprises access permissions of an object stored at the server has been received. A version of the record is stored in an in-memory cache. The thread acquires a first lock associated with the record in exclusive mode, and a second lock in shared mode. The thread reads the cached version, constructs a modified version indicated in the command, and releases the second lock without updating the cache. After entries representing the modified version are stored in an operations journal, the thread acquires the second lock in exclusive mode and updates the cache. The thread then releases both locks. An indication that the command has succeeded is provided without verifying that the second version is stored at a metadata store.

Enhancing i/o performance using in-memory reservation state caching at block storage services

A method for data storage, in a system that includes multiple servers, multiple multi-queue storage devices and at least one storage controller that communicate over a network, includes receiving in a server, from an application running on the server, a request to access data belonging to one or more stripes. The stripes are stored on the storage devices and are shared with one or more other servers. In response to the request, the following are performed at least partially in parallel: (i) requesting one or more global locks that prevent the other servers from accessing the stripes, and (ii) reading at least part of the stripes from the storage devices speculatively, irrespective of whether the global locks are granted. Execution of the request is completed upon verifying that the speculatively-read data is valid.

Shared volumes in distributed RAID over shared multi-queue storage devices

A computer program product, system, and method are provided for upgrading a kernel or kernel module with a configured persistent memory. A persistent memory space is configured in the memory to store application data from applications in user mode. A kernel executing in the memory is prevented from accessing the persistent memory space. A service is called to load an updated kernel in the memory to replace the kernel, wherein the applications have access to the persistent memory space after the updated kernel is loaded. The service may comprise a kernel execution mechanism that directly loads the updated kernel into the memory without a full reboot of the computer system. An extended memory kernel service may be loaded during a boot operation to reserve the persistent memory space as an extended memory space for use by the applications and prevent the kernel from accessing the persistent memory space.

Upgrading a kernel or kernel module with a configured persistent memory unused by the kernel

Methods, apparatus and computer program products implement embodiments of the present invention that enable a computer system comprising networked computers to self-heal from a boot failure of one of the computers. In some embodiments, upon detecting a first computer having a memory and a physical boot device failing to successfully load a boot image, a second computer defines a virtual boot device in communication with the first computer and the second computer and having a boot image for the first computer. Subsequent to defining the virtual boot device, the second computer can configure the first computer to boot from the virtual boot device.

Self-healing using a virtual boot device

A method for data storage includes, in a system that includes multiple servers, multiple multi-queue storage devices and at least one storage controller that communicate over a network, storing data blocks by the servers on the storage devices. A shared data structure, which is accessible to the servers using remote direct memory access and which maps hash values calculated over the data blocks to respective storage locations of the data blocks on the storage devices, is maintained. The data blocks stored on the storage device are deduplicated, by looking-up the shared data structure by the servers without executing code on a processor of the storage controller.

Deduplication in a highly-distributed shared topology with direct-memory-access capable interconnect

In an approach for distributed management of a network, a processor discovers a network, wherein the network is comprised of a plurality of nodes and an at least one switch. A processor identifies a first port associated with the at least one node. A processor identifies a link between the first port and a second port within the network. A processor determines the at least one link is properly configured. A processor selects a first node of the plurality of nodes to be a master node, wherein the master node is configured to manage the network. A processor initializes a routing table, wherein the routing table contains locations of each node in the plurality of nodes and a location of the at least one switch. A processor adjusts by the first node, the routing table based on the change to the network.

Distributed subnet manager for InfiniBand networks

In an approach for managing read-modify-write operations within a network. Aspects of an embodiment of the present invention include an approach for managing read-modify-write operations within a network, wherein the approach includes a processor for determining that a lock associated with a component of a network is set. A processor identifies a key associated with the lock. A processor determines that the key associated with the lock has not been altered since a previous time cycle. A processor responsive to determining that the key associated with the lock has not been altered, determines that a lock timeout has expired for the key, wherein the lock timeout specifies, at least, a period of time. A processor responsive to determining that the lock timeout has expired for the key, releases the lock.

Cooperative MKEY locking for managing infiniband networks

A method for data storage includes, in a system that includes multiple servers, multiple multi-queue storage devices and at least one storage controller that communicate over a network, storing data by the servers in a storage process that (i) caches the data in a cache memory that is accessible to the servers and to the storage devices and (ii) de-stages the cached data from the cache memory to the storage devices. A progress of the storage process is continually journaled, by the servers, in a plurality of server-specific journals residing in the cache memory. In response to a failure occurring during the storage process, the data is recovered by the storage controller based on the server-specific journals.

Write cache and write-hole recovery in distributed raid over shared multi-queue storage devices

A computer program product, system, and method are provided for upgrading a kernel or kernel module with a configured persistent. A persistent memory memory space is configured in the memory to store application data from applications in user mode. A kernel executing in the memory is prevented from accessing the persistent memory space. A service is called to load an updated kernel in the memory to replace the kernel, wherein the applications have access to the persistent memory space after the updated kernel is loaded. The service may comprise a kernel execution mechanism that directly loads the updated kernel into the memory without a full reboot of the computer system. An extended memory kernel service may be loaded during a boot operation to reserve the persistent memory space as an extended memory space for use by the applications and prevent the kernel from accessing the persistent memory space.

A method for data storage includes, in a system that includes multiple servers, multiple multi-queue storage devices and at least one storage controller that communicate over a network, running, in a server among the servers, multiple data-path instances (DPs) that operate independently of one another and issue storage commands for execution in the multi-queue storage devices. The storage commands, issued by the multiple DPs running in the server, are multiplexed using an Input-Output Multiplexer (I/O MUX) process. The multiplexed storage commands are executed in the multi-queue storage devices.

Storage in multi-queue storage devices using queue multiplexing and access control

Apparatus and computer program products implement embodiments of the present invention that enable a computer system comprising networked computers to self-heal from a boot failure of one of the computers. In some embodiments, upon detecting a first computer having a memory and a physical boot device failing to successfully load a boot image, a second computer defines a virtual boot device in communication with the first computer and the second computer and having a boot image for the first computer. Subsequent to defining the virtual boot device, the second computer can configure the first computer to boot from the virtual boot device.

Growing community of inventors

Herzliya, Israel

Alex Friedman