Abstract: The method includes identifying a log stream from a first device, wherein the first device is of an unknown device type. The method further includes identifying features in the log stream from the first device. The method further includes determining a probability that the log stream from the first device was created by a specific device type based on matching of features of the identified log stream from the first device and features in log streams from a plurality of known device types.

Abstract: Systems and methods commence upon detecting certain computing system specification change events. Changes are recorded in resource usage intent specification records that characterize desired states associated with respective resources of the computing system. The change event causes generation of a data state snapshot of data stored in the computing system. A state tuple is constructed to associate the desired state with the data state. The state tuple is committed to a version control system. At any time, a user or process can select a selected system state tuple from committed tuples stored in the version control system. The tuple is analyzed to determine its associated desired state and its data state. Actions are taken to restore the data state of the tuple to the computing system, and then bringing the computing system to the state of the tuple so as to reproduce the desired system state in the computing system.

Abstract: Methods and apparatus for identifying consistent snapshot points for file systems in a distributed storage service in which a client may access a file system in backend storage via two or more access nodes. A snapshot coordinator initiates a two-phase commit to perform a snapshot epoch transition. The coordinator sends prepare messages to the nodes. In response, a node sends a response to the coordinator, stops issuing writes to the file system, sets a timer, and waits for a commit message. Upon receiving a response from all nodes, the coordinator sends a commit message. Upon receiving the commit or the timer expiring, the nodes increment an epoch counter and begin issuing writes to the file system tagged with the new epoch and indicating that a copy-on-write operation be performed. The transition is successful if all nodes respond to the request message; otherwise, the two-phase commit is retried.

Abstract: A method, apparatus, computer-readable medium, and/or system may comprise a user device configured to receive, from a user, a selection of a plurality of dependent computing tasks for a first computing task. The user device may transmit, to a computing device, the selection of the plurality of dependent computing tasks for the first computing task. The system may comprise a computing device comprising a processor and memory storing computer-executable instructions that, when executed by the processor, cause the computing device to receive, from the user device, the selection of the plurality of dependent computing tasks for the first computing task. The computing device may store, at a task registry and in association with the first computing task, an indication of the plurality of dependent computing tasks. The computing device may cause transmission of the first computing task from the task registry to a task queue.

Abstract: The present disclosure relates to packet-aware fault tolerance methods of virtual machines applied to cloud services. One such method includes using a primary input processing unit to receive an uploaded packet and replicating the uploaded packet to a backup input processing unit. The method can further include utilizing a primary execution unit to operate the uploaded packet to generate an output packet, which is then transmitted to a processing unit. A backup execution unit can be used to operate the uploaded packet to generate a backup output packet. The backup output packet can be transmitted to a backup processing unit. A fault-tolerance control component can receive the primary output packet from the primary output processing unit. The primary fault-tolerance control component can be used to compare a primary packet checksum of the primary output packet is consistent with a backup packet checksum.

Abstract: A system includes a dynamic random-access memory (DRAM); and a storage device comprising a power source and a persistent store. The storage device is configured to provide reserve power to the DRAM. Data stored in the DRAM is transferred to a reserved storage in the persistent store of the storage device in a power loss event using the reserve power.

Abstract: An implementation of the disclosure provides a system for generating low impact snapshots comprising a memory to store application data and a processing device coupled to the memory. The processing device is to identify a plurality of transactions within a cloud computing environment. Each of the transactions comprising an operation associated with a storage device of the cloud computing environment. A transaction queue that is storing identifiers of high and low priority database operations to be performed by an application on the storage device is monitored. (Currently amended) The monitoring indicates how many transactions are currently being processed. A snapshot policy that stores conditions in which a snapshot can be generated without adversely impacting performance of the storage device is evaluated based on the monitored transaction queue.

Abstract: A cache is maintained with write order numbers that indicate orders of writes into the cache, so that periodic partial flushes of the cache can be executed while maintaining write order consistency. A method of storing data into the cache includes receiving a request to write data into the cache, identifying lines in the cache for storing the data, writing the data into the lines of the cache, storing a write order number, and associating the write order number with the lines of the cache. A method of flushing a cache having cache lines associated with write order numbers includes the steps of identifying lines in the cache that are associated with either a selected write order number or a write order number that is less than the selected write order number, and flushing data stored in the identified lines to a persistent storage.

Abstract: Data integrity is maintained in an environment in which multiple instances of a software component each can include multiple processes capable of accessing a shared resource. Each instance of the software component can be either an “active” instance (which is allowed to update data) or a “standby” instance (which is not allowed to update data). At any point in time, the environment does not comprise more than one active instance. There can, however, be more than one standby instance. Preferably, a single standby instance can become a new active instance in the event of a complete failure of the active instance, i.e. a failure associated with every process of the current active instance.

Abstract: An example information encryption method that includes acquiring to-be-encrypted information and converting the to-be-encrypted information into a polynomial of a predetermined format; extracting biometric information, and acquiring biometric data; and substituting the biometric data into the polynomial for calculation to acquire a value of the polynomial and using a two-dimensional dataset including the biometric data and the value of the polynomial corresponding to the biometric data as first encrypted information. The techniques of the present disclosure improve the security of information encryption, and reduce the risk of illegal decryption of encrypted information.

Abstract: Processing of transactions within a computing environment is facilitated by taking actions to increase the chances of successfully executing a transaction. A counter is maintained that provides a count of how often a transaction has aborted. The counter increments the count each time the transaction is aborted, and it is reset to zero upon successful completion of the transaction or an interruption leading to no more re-executions of the transaction. If the count reaches a threshold value, then an interrupt is presented and transaction execution is unsuccessful. However, before the count reaches the threshold, a number of actions may be taken to increase the chances of successfully executing the transaction. These actions include actions to be performed within the processor executing the transaction, and/or actions to be performed against conflicting processors.

Abstract: A method of evaluating real-time computer faults and using a fault evaluation (FE) platform is provided. The method includes ingesting log data associated with a computer system, the log data includes a plurality of fault events, a fault severity identifier associated with at least one fault event of the plurality of fault events. The method also includes selecting, from the plurality of fault events, a fault event set which corresponds to a time window and includes the at least one fault event. The method further includes generating a fault score for the at least one fault event and an aggregate fault score. The method also includes determining that the aggregate fault score exceeds a predefined threshold, and providing, to a configuration management platform, instructions to initiate a hardware component remediation process.

Abstract: A computer system includes transactional memory to implement a nested transaction. The computer system generates a plurality of speculative identification numbers (IDs), identifies at least one of a software thread executed by a hardware processor and a memory operation performed in accordance with an application code. The computer system assigns at least one speculative cache version to a requested transaction based on a corresponding software thread. The speculative ID of the corresponding software thread identifies the speculative cache version. The computer system also identifies a nested transaction in the memory unit, assigns a cache version to the nested transaction, detects a conflict with the nested transaction, determines a conflicted nesting level of the nested transaction, and determines a cache version corresponding to the conflicted nesting level. The computer system also invalidates the cache version corresponding to the conflicted nesting level.

Abstract: Examples described herein include receiving an operation pipeline for a computing system and building a graph that comprises a model for a number of potential memory side accelerator thread assignments to carry out the operation pipeline. The computing system may comprise at least two memories and a number of memory side accelerators. Each model may comprise a number of steps and at least one step out of the number of steps in each model may comprise a function performed at one memory side accelerator out of the number of memory side accelerators. Examples described herein also include determining a cost of at least one model.

Abstract: An in-memory database system stores the entirety of a database that is being operated on by a database system within main memory. The database includes a table that comprises a plurality of rows including a particular row. The database table comprises a first version of the particular row having a first valid time, and a second version having a second valid time. Index(es) are associated with the database table. Each index references the plurality of rows, including referencing the first and second versions of the particular row. A transaction acting on the first version of the particular row is executed. The first version of the particular row is visible to the transaction based on the first valid time and the second version of the particular row being not visible to the transaction based on the second valid time.

Abstract: In various embodiments a distributed computing node in a plurality of distributed computing nodes executes transactions in a distributed processing system. In one embodiment, a transaction commit message is received from a client computing node for a transaction. The transaction commit message includes at least an identifier of the transaction and a transaction sequence for the transaction. The transaction sequence indicates a sequence of execution for the transaction on the plurality of distributed computing nodes. An entry within the transaction sequence associated with the distributed computing node is identified. The entry includes a sequence number for executing the transaction on the distributed computing node with respect to other transactions. The transaction is executed based on the sequence number in the entry.

Abstract: Methods, systems and computer program products are provided for managing at least one cache for a database. In one method, an operation record is inserted in an operations log for each write operation to the database. The operation record includes a Log Sequence Number (LSN). At least one cache is updated using operation records in the operations log. To update a cache, unapplied operations in the operations log for the cache are identified using an applied LSN associated with the cache. The applied LSN denotes an LSN of a last operation applied to the cache. The unapplied operations are then applied to the cache and the applied LSN is updated with an LSN corresponding to the last unapplied operation of the unapplied operations that were applied to the cache during updating the cache.

Abstract: Techniques are provided to configure a graphics processing unit (GPU) service platform to provide GPU processing and checkpointing services to a client system. For example, when the GPU service platform receives a service request from a first client system for GPU processing services, a GPU server node is allocated to execute a first GPU processing task associated with the service request using one or more GPU devices. When a GPU checkpointing operation is performed during execution of the first GPU processing task to generate a checkpoint image of a current state of the first GPU processing task, the GPU server node is configured to execute at least a portion of a second GPU processing task associated with a second service request received from a second client system, using the one or more GPU devices concurrently with performing the GPU checkpointing operation for the first GPU processing task.

Abstract: A computer system includes transactional memory to implement a nested transaction. The computer system generates a plurality of speculative identification numbers (IDs), identifies at least one of a software thread executed by a hardware processor and a memory operation performed in accordance with an application code. The computer system assigns at least one speculative cache version to a requested transaction based on a corresponding software thread. The speculative ID of the corresponding software thread identifies the speculative cache version. The computer system also identifies a nested transaction in the memory unit, assigns a cache version to the nested transaction, detects a conflict with the nested transaction, determines a conflicted nesting level of the nested transaction, and determines a cache version corresponding to the conflicted nesting level. The computer system also invalidates the cache version corresponding to the conflicted nesting level.

Abstract: This invention relates to a system, method and computer program product for a drive array controller (for example a RAID6 controller) comprising: a failed strip identifier for identifying a failed strip in a critical stride strip in drive array when the drive array has two or more failed drives, a critical stride is a stride with two (for example) failed strips, a failed strip is a strip in a failed drive, a failed drive is a drive that can no longer be reliably used; a safe strip identifier for identifying a safe strip in a safe stride, a safe stride is a stride with no failed strips (or at least two less than a critical stride), a safe strip is a strip in a safe stride; a drive array writer for writing a recalculated failed strip to the drive array location for the safe strip; and a lending table writer for writing the failed strip location and associated safe strip location to a lending table.

Abstract: A test coordinator for an application whose data is to be stored at a journal-based multi-data-store database using an optimistic concurrency control protocol identifies a test descriptor. The test descriptor indicates a particular order of a plurality of events associated with the application's data, such as reads from various data stores and propagations of committed writes from the database journal to the data stores. The test coordinator uses a set of control mechanisms to cause the events to occur in a selected execution environment in the particular order, and stores an indication of a state reached by the application in response to the events.

Abstract: In an information processing apparatus connected to an external apparatus, when information indicating a change in a folder indicates that a new file has been appended to the folder, a document management application updates a screen so that existing files are arranged in order according to a specified sorting method and a new file is appended to a position, not according to the specified sorting method, at which the new file can be easily recognized.

Abstract: A network security assessment apparatus includes a memory and a processor. The memory stores first, second, and third values of a metric. The metric indicates one or more of a number of network security breaches and a number of fraudulent transactions. The processor chronologically orders the first, second, and third values for the metric over a period of time and compares the first, second, and third values against a threshold to produce first, second, and third results. If the first result is different from the second result or if the second result is different from the third result, the processor increments a volatility count. The processor determines, based on the volatility count, that a control should be implemented to mitigate one or more of the number of network security breaches and the number of fraudulent transactions and in response to that determination, the processor implements the control.

Abstract: A computer-implemented method for generating an updated action history for display in a graphical user interface. The method may include identifying a last matching action between two action histories, disabling a subset of actions in an action history, inserting, subsequent to the last matching action in an updated action history, an action corresponding to user-triggered input, enabling a subset of actions in the updated action history that are listed in the action history after the last matching action and displaying the updated action history in the graphical user interface, the updated action history including one or more indications for each of the actions associated with inserting the action corresponding to the user-triggered input.

Abstract: An in-memory database system stores the entirety of a database that is being operated on by a database system within main memory. The database includes a table that comprises a plurality of rows including a particular row. The database table comprises a first version of the particular row having a first valid time, and a second version having a second valid time. Index(es) are associated with the database table. Each index is implemented as a lock-free data structure and references the plurality of rows, including referencing the first and second versions of the particular row. A transaction acting on the first version of the particular row is executed. The first version of the particular row is visible to the transaction based on the first valid time and the second version of the particular row being not visible to the transaction based on the second valid time.

Abstract: Disclosed herein are system, method, and computer program product embodiments for performing garbage collection on a multi-version concurrency control information in the database management system. An embodiment operates by determining, using multi-version concurrency control (MVCC) information, when a row manipulated by a write transaction is visible to a plurality of readers accessing a table that includes the row. The MVCC information for the row includes at least a creation timestamp, a destruction timestamp and a row state. Once the row is visible to the plurality of readers, garbage collecting at least the creation timestamp or the destruction timestamp in the MVCC information. After the creation timestamp or destruction timestamp have been garbage collected, the plurality of readers use the row state to determine accessibility of the row in the table.

Abstract: An in-memory database system includes database table stored in system memory. The database table comprises a plurality of rows including a particular row. Data corresponding to each row is stored entirely in the system memory. The database table comprises a first version of the particular row having a first valid time, and a second version having a second valid time. Index(es) are associated with the database table. Each index is implemented as a lock-free data structure and references the plurality of rows, including referencing the first and second versions of the particular row. A first transaction acting on the first version of the particular row is executed. The first version of the particular row is visible to the first transaction based on the first valid time and the second version of the particular row being not visible to the first transaction based on the second valid time.

Abstract: An example information encryption method that includes acquiring to-be-encrypted information and converting the to-be-encrypted information into a polynomial of a predetermined format; extracting biometric information, and acquiring biometric data; and substituting the biometric data into the polynomial for calculation to acquire a value of the polynomial and using a two-dimensional dataset including the biometric data and the value of the polynomial corresponding to the biometric data as first encrypted information. The techniques of the present disclosure improve the security of information encryption, and reduce the risk of illegal decryption of encrypted information.

Abstract: A database upon opening each of at least one cursor for a transaction, assigns each statement forming part of a transaction a statement sequence number (SSN) involving at least one row of the database. Thereafter, temporary timestamps are assigned to the at least one row using a transaction identifier and the corresponding SSN. Further, active cursors are monitored such that a final timestamp is prevented from being associated with each row until it has no remaining cursors. Related apparatus, systems, techniques and articles are also described.

Abstract: A system and method is illustrated for utilizing an agent in intelligent Continuous Data Protection (CDP) snapshot based backups. The system and method utilizing the agent receiving a previous backup bookmark, obtaining a data change rate by calculating a number of bookmarks since the previous backup bookmark, comparing the data change rate with a threshold, initiating a backup when the data change rate exceeds the threshold, and labeling the backup with a bookmark.

Abstract: A system for logging database transactions comprises a memory (comprising a direct access storage device) operable to store a database, a database log, and a plurality of log buffers associated with the database log. Each log buffer is operable to store at least one log record associated with a database transaction. The system further comprises a processor operable to log a first record of a first database transaction using the database log and write the database log to the storage device. The processor is further operable to log a second record of a second database transaction in a first of the plurality of log buffers substantially concurrently to the database log being written to the storage device.

Abstract: According to one embodiment, a shared storage system includes a plurality of host servers, a plurality of storage devices, and a management server. The management server manages each of a plurality of logical units using first address management information, and manages a revision of the first address management information using first revision data. The host servers each hold respective copies of the first address management information and first revision data. When a first host server has requested a first storage device to execute access, based on a copy (second revision data) of the first revision data, the first storage device executes the requested access on condition that the second revision data coincides with revision data notified by the management server.

Abstract: A method, medium, and system to receive a request to perform a log recovery to restore multiple database services; determine log backup entries corresponding to a target log position for a first database service of the multiple database services; read from a sequential stream device, by the first database service, the log backup entries corresponding to the target log position for the first database service; inform a second database service of the multiple database services that the first database service has concluded executing the log backup entries corresponding to the target log position for the first database service from the sequential stream device; assuring that no resources of the streaming device are blocked by the first database service; and read log backup entries of the second database service corresponding to a target log position for the second database service from the sequential stream device.

Abstract: Various technologies described herein pertain to performing time travel debugging. A computer-executable program can be executed. The computer-executable program can be executable under control of a virtual machine. The virtual machine can interact with a browser system during execution of the computer-executable program. Moreover, nondeterministic events can be logged via an interrogative virtual machine interface (VMI) during the execution of the computer-executable program. The nondeterministic events can be logged as part of event logs. Moreover, the interrogative VMI is between the virtual machine and the browser system. Further, snapshots of the virtual machine can be captured during the execution of the computer-executable program. The snapshots can be captured via the interrogative VMI. At least a portion of the execution of the computer-executable program can be replayed based at least in part on a snapshot and at least a portion of the event logs.

Abstract: A transaction model for data stores using a distributed file system is disclosed. In various embodiments, a transaction-related metadata is sent to each of a plurality of segments to which a corresponding assigned portion of a query plan is assigned. A metadata modification record is received from zero or more of the segments, reflecting changes, if any, to system metadata required to be made to reflect changes to data made by that segment in processing the portion of the query plan assigned to that segment. It is determined whether a metadata modification record has been received from each of the segments in the plurality of segments. In the event a metadata modification record is determined to have been received from each of the segments in the plurality of segments, a system metadata store is updated to reflect the respective metadata modification records received from the plurality of segments.

Abstract: The present invention provides a method of and apparatus for implementing changes within a data system. The method receiving a request including one or more objects to be imported into a data system, creating a back up request comprising a copy of the objects included within the request, implementing the back up request in the system to generate a modified back up request including a log of the objects modified in the request, storing the modified back up request. This allows a log of changes to objects in the data system to be generated using a copy of a request including a list of changes to the objects. Once the modified back up request has been generated the request may be imported into the system.

Abstract: There is provided an information processing apparatus including a control unit. The control unit adds type information indicating a first type in a first operation request in which operation target storage devices are specified by unique identifiers, respectively, used in a first storage apparatus and outputs the first operation request to the first storage apparatus through the network. When the control unit transmits the first operation request to a second storage apparatus and receives an error response indicating a type error from the second storage apparatus after transparent failover is executed, the control unit adds type information indicating the second type in a second operation request in which operation target storage devices are specified by unique identifiers, respectively, used in the second storage apparatus and outputs the second operation request to the second storage apparatus through the network.

Abstract: A method and system for large data and distributed disk cache processing in a Pneuron platform 100. The system and method include three specific interoperable but distributed functions: the adapter/cache Pneuron 14 and distributed disk files 34, a dynamic memory mapping tree 50, and distributed disk file cleanup 28. The system allows for large data processing considerations and the ability to access and acquire information from large data sets 102a, 102b and rapidly distribute and provide the information to subsequent Pneurons 104 for processing. The system also provides the ability to store large result sets, the ability to deal with sequential as well as asynchronous parallel processing, the ability to address large unstructured data; web logs, email, web pages, etc., as well as the ability to handle failures to large block processing.

Abstract: Embodiments of the present disclosure include a method, a system, and a computer program product for processing multiple transactions within a transaction execution environment. The transaction execution environment includes a first transaction that is assigned a first transaction identification number, and a second transaction that is assigned a second transaction identification number which process on a processor. The processor processes the first transaction and the second transaction on one or more threads of the processor. A first rollback save state is created for the first transaction and a second rollback save state is created for the second transaction. An interference is detected prior to ending a transaction, between the processing of the first transaction and the processing of the second transaction. In response to the interference, the second transaction is rolled back to the second rollback save state.

Abstract: A system includes reception of an instruction to load a page from a database datastore into a database cache, allocation of the page within the cache, determination of a resource type associated with the page, and allocation of a control block for the page within the cache based on the determined resource type, wherein the allocated control block in a case that the resource type is a first resource type is different from the allocated control block in a case that the resource type is a second resource type.

Abstract: A distributed data store may provide volume recovery access to multiple recovery agents. A data volume may be maintained for a storage client at the distributed data store. Write access to the data volume may be granted according to a single writer consistency scheme. When a recovery event is detected for the data volume, the data volume may be made available to multiple recovery agents that may perform respective recovery operations. Upon first completion of a recovery operation for the data volume, granting access to the data volume according to the single writer consistency scheme may be resumed. In some embodiments, the distributed data store may be a log-structured data store.

Abstract: Methods and apparatus to restore resources to a device according to an order of the resources requested by a user are described. The restoration may be interrupted in response to a user request to update an order to restore remaining resources yet to restore. The remaining resources may be restored according to the updated order.

Abstract: A pluggable database is transported between a source DBMS and a destination DBMS, in a way that minimizes downtime of the pluggable database. While a copy of the pluggable database is being made at the destination DBMS, transactions continue to execute against the pluggable database at the source DBMS and change the pluggable database. Eventually, the transactions terminate or cease executing. Redo records generated for the transactions are applied to the copy of the pluggable database at the source DBMS. Undo records generated for at least some of the transactions may be stored in a separate undo log and transported to the destination DBMS. The transported pluggable database is synchronized at a destination DBMS in a “pluggable-ready state”, where it may be plugged into the destination container DBMS.

Abstract: A method for recovering system status consistently to a designed recovering time point in a distributed database, wherein the distributed database comprises a primary server and multiple region servers, comprising: when the region servers detect a change of system status, analyzing an event for the change to generate an event log and storing the event log to the database; after a preset condition, the region servers generate a snapshot respectively; when the primary server receives the instruction to recover the system status to a designed time point, indicates the region servers to implement: reading the event log and snapshots stored in region servers; finding the snapshot closest to the designed time point; finding the event log and snapshot corresponding to an time interval between the time recorded in the snapshot and the designed time point to recover the system status to the designed time point.

Abstract: A method begins by identifying a storage unit of a set of storage units within a dispersed storage system as being likely to fail. The method continues by determining an approach for minimizing rebuilding the encoded data slices that are stored on an identified storage unit in an event of a failure of the identified storage unit. The approach is at least one approach option from a list of approach options that includes fostering new encoded data slices, replicating existing encoded data slices, and re-dispersing existing encoded data slices. The method continues by implementing the approach regarding the identified storage unit and when the identified storage unit fails, retrieving the encoded data slices that were stored on the identified storage unit based on the approach.

Abstract: According to one aspect of the invention, a request to generate a state checkpoint of a computer is initiated within a user-level software entity, such as a virtual machine. Upon sensing the request, a checkpointing mechanism generates and stores at least one checkpoint, each checkpoint comprising a representation of the total state of the computer system. Upon sensing a state restoration request corresponding to one of the checkpoints, the checkpointing mechanism restores the checkpointed state in the computer, which can then resume operation from the restored total state. According to another aspect of the invention, a total checkpointed state is exported to another computer, where the state can be modified, for example, debugged, and then loaded into either the originally checkpointed computer (which, again, may be a virtual machine), or some other computer.

Abstract: A method is used in managing files of file systems. A group transaction record for an indirect data block of a file of a file system is created. The indirect data block of the file includes a set of data blocks of the file. A set of transaction entries is added to the group transaction record. Each transaction entry of the set of transaction entries includes information regarding an operation performed on a data block of the set of data blocks. The group transaction record is stored in a journal. The journal includes metadata transactions upon a plurality of files of a plurality of file systems.

Abstract: In an embodiment, a request is received that specifies a function and a specified key value. Rows from all tables that are accessible from the specified key value are transitively searched, wherein the transitively searching further comprises finding values in a plurality of pairs of columns, wherein found rows that are found by the transitively searching comprise values in a respective first column of the plurality of pairs of columns that satisfy a dependency relationship with values in a respective second column of the plurality of pairs of columns. The function is executed against only the found rows.

Abstract: In an embodiment, a request is received that specifies a function and a specified key value. Rows from all tables that are accessible from the specified key value are transitively searched, wherein the transitively searching further comprises finding values in a plurality of pairs of columns, wherein found rows that are found by the transitively searching comprise values in a respective first column of the plurality of pairs of columns that satisfy a dependency relationship with values in a respective second column of the plurality of pairs of columns. The function is executed against only the found rows.

Abstract: According to an aspect, a method for checkpointing in a hybrid computing node includes executing a task in a processing accelerator of the hybrid computing node. A checkpoint is created in a local memory of the processing accelerator. The checkpoint includes state data to restart execution of the task in the processing accelerator upon a restart operation. Execution of the task is resumed in the processing accelerator after creating the checkpoint. The state data of the checkpoint are transferred from the processing accelerator to a main processor of the hybrid computing node while the processing accelerator is executing the task.