A method for processing a series of database transactions according to an embodiment includes recording a first transaction requested to a first database by a first application, among the plurality of applications, in association with a transaction group ID, recording a second transaction requested to a second database by a second application called by the first application, among the plurality of applications, in association with the transaction group ID, rolling back the second transaction targeted to the second database in response to a determination that an error has occurred during processing of the second transaction, identifying the first transaction based on the transaction group ID and rolling back the first transaction targeted to the first database.

An example file server manager updates a selected share of a destination distributed file server based on a snapshot of at least a portion of a selected share of a source distributed file server. The selected share of the destination distributed file server is updated while the source distributed file server serves client requests for storage items of the selected share of the source distributed file server. The file server manager receives a request to failover from the source distributed file server to the destination distributed file server and configures the destination distributed file server to service read and write requests for storage items of the selected share of the destination distributed file server. The file server manager further redirects client requests for storage items of the selected share of the source distributed file server to the destination distributed file server by updating active directory information.

In an example embodiment, information about one or more failed payment attempts via an electronic payment processing system is obtained. One or more features are extracted from the information. Then, for each of a plurality of potential candidate retry time points, the one or more features and the potential candidate retry time point are fed into a dunning model, the dunning model trained via a machine-learning algorithm to produce a dunning score indicative of a likelihood that a retry attempt at an input retry time point will result in a successful payment processing. The dunning scores for the plurality of potential candidate retry time points are used to select a desired retry time point. Then the electronic payment processing system is caused to attempt to reprocess a payment associated with one of the failed payment attempts at a time matching the desired retry time point.

Live recovery generates a new “recovery VM” that operates as an ongoing “live” production platform. A previously created non-cloud-native backup copy is the data source for the recovery VM. Live recovery restores data blocks from the backup copy on backup media directly to cloud-based virtual disk(s) assigned to the recovery VM. As a result, the cloud-based recovery VM can become fully operational in the cloud computing environment on a going-forward basis. The advantage of live recovery over a traditional restore is that live recovery provides a cloud-based VM that begins operating well before the backup copy is fully restored. This is accomplished by temporarily mounting a “temp-mounted VM” in the cloud while the backup copy is methodically restored in the background. VM reads and writes begin issuing from the temp-mounted VM and writes are retained on completion. Downtime is minimized when switching from the temp-mounted VM to the recovery VM.

A method for processing data exactly once using transactional stream writes includes receiving, from a client, a batch of data blocks for storage on memory hardware in communication with the data processing hardware. The batch of data blocks is associated with a corresponding sequence number and represents a number of rows of a table stored on the memory hardware. The method also includes partitioning the batch of data blocks into a plurality of sub-batches of data blocks. For each sub-batch of data blocks, the method further includes assigning the sub-batch of data blocks to a buffered stream; writing, using the assigned buffered stream, the sub-batch of data blocks to the memory hardware; updating a storage log with an intent to commit the sub-batch of data blocks using the assigned buffered stream; and committing the sub-batch of data blocks to the memory hardware.

One embodiment of the present invention sets forth a technique for optimizing data in a dataset. The technique includes determining, based on one or more attributes of a dataset, an optimization that is associated with at least one of a file encoding, a file size, and a sort column. The technique also includes identifying a plurality of candidate configurations associated with the dataset and corresponding to the optimization, and for each candidate configuration, generating a corresponding set of evaluation metrics associated with the first optimization. The technique further includes determining, based on the sets of evaluation metrics corresponding to the plurality of candidate configurations, a set of configurations in the plurality of candidate configurations to be applied to the dataset. Finally, the technique includes modifying the dataset based on the set of configurations.

Systems and methods for storage snapshots for nested virtual machines. An example method may comprise running, by a host computer system, a hypervisor managing a first virtual machine associated with a first virtual device. Responsive to creating a second virtual machine by the hypervisor, requesting, by the first virtual machine, a first snapshot of the first virtual device. The hypervisor generates the first snapshot of the first virtual device and forwards the first snapshot of the first virtual device to the second virtual machine.

A remote management system includes a device management service having multiple instances that performs parallel distributed processing through the instances and a worker service having multiple instances that performs parallel distributed processing through the instances. Each instance of the device management service acquires the status of processing related to the recovery of the relevant instance from counter information, and stores the status of the processing related to the recovery of the relevant instance in the counter information when having performed the processing. Each instance of the worker service acquires the status of processing related to the recovery of an instance of the device management service from the counter information, and stores the status of the processing related to the recovery of the instance of the device management service in the counter information when having performed the processing.

Described is a system (and method) that provides the ability to create an endpoint to allow cloud-based components to access services directly using network infrastructure of a cloud system. To provide such an ability, access connections from components of a production system to the cloud system may be monitored to derive a storage service and a datastore based on portions of a domain name. The derived storage service and datastore are then used to determine configuration settings required to automatically create a service endpoint. The service endpoint may then be deployed within the cloud system allowing the cloud-based component to access the storage service directly. Accordingly, the system provides the ability to automatically configure and deploy service endpoints by leveraging information derived from monitoring network access connections between a production environment and a cloud environment.

Techniques for managing a disk involve acquiring a message for an access operation for a disk, the message including a first status code at an operating system level for the access operation. The techniques further involve acquiring a second status code at a disk hardware level for the access operation if it is determined that the first status code indicates that the access operation fails. The techniques further involve determining, according to a handling policy corresponding to the second status code, whether the disk will be marked as faulty. The techniques further involve managing the disk based on a count of failed access operations for the disk if it is determined that the disk is not marked as faulty. Such techniques may quickly determine a specific reason for a failure of a disk access operation, making it possible to solve problems quickly, save time and improve the user experience.