Abstract: An approach is provided for managing a failure of a critical high availability (HA) component in a HA system. Critical HA components are identified. Categories are assigned to the identified components and weights are assigned to the categories. A current value indicating a performance of a component included in the identified components is obtained by periodically monitoring the components. A reference value for the performance of the component is received. A deviation between the current value and the reference value is determined. Based on the deviation, the component is determined to have failed. Based in part on the failed component, the categories, and the weights, a health index is determined in real-time. The health index indicates in part how much the component having failed affects a measure of health of the HA system.

Abstract: A method and system of distributed parallel processing. There are a plurality of distributed parallel processing units (DPPUs). Each DPPU is configured to receive data related to a condition of the network. The type of data received by each DPPU is disparate for each DPPU. Each DPPU analyzes its data. Upon determining that a predetermined condition is met or a predetermined threshold is exceeded, the disparate data is transformed into a common format using an appropriate driver of the configuration module. The common format data is sent to a storage device of a first DPPU of the plurality of DPPUs.

Abstract: An apparatus includes a processing unit including a configuration memory and self-scrubber logic coupled to read the configuration memory to detect compromised data stored in the configuration memory. The apparatus also includes a watchdog unit external to the processing unit and coupled to the self-scrubber logic to detect a failure in the self-scrubber logic. The watchdog unit is coupled to the processing unit to selectively reset the processing unit in response to detecting the failure in the self-scrubber logic. The apparatus also includes an external memory external to the processing unit and coupled to send configuration data to the configuration memory in response to a data feed signal outputted by the self-scrubber logic.

Abstract: Techniques for load balancing and fault tolerant service are described. An apparatus may comprise load balancing and fault tolerant component operative to execute a load balancing and fault tolerant service in a distributed data system. The load balancing and fault tolerant service distributes a load of a task to a first node in a cluster of nodes using a routing table. The load balancing and fault tolerant service stores information to indicate the first node from the cluster of nodes is assigned to perform the task. The load balancing and fault tolerant service detects a failure condition for the first node. The load balancing and fault tolerant service moves the task to a second node from the cluster of nodes to perform the task for the first node upon occurrence of the failure condition.

Abstract: Techniques are described for identifying unhealthy nodes in a multi-node system. One or more parameters of each node is monitored, then compared with the values for the same parameter running on other nodes in the multi-node system. Based on the comparison, a determination is made whether a node is healthy. If the multi-node system comprises one or more nodes with differing capabilities, an adjustment is performed to account for the differing capabilities of each respective node. Further provided are methods of taking remedial action upon a determination that a node is unhealthy. A tuner is used to modify values of health parameters until the node is performing similarly to its peers.

Abstract: A memory management method, a memory storage device and a memory control circuit unit are provided. The memory management method includes: detecting a replacement physical unit number of a rewritable non-volatile memory module; adjusting an available capacity of the rewritable non-volatile memory module from a first available capacity to a second available capacity if the replacement physical unit number meets a default condition. Therefore, a lifespan of the memory storage device may be prolonged by adjusting the available capacity of the rewritable non-volatile memory module.

Abstract: A Peripheral Component Interconnect Express (PCIe) network system with fail-over capability and an operation method thereof are provided. The PCIe network system includes a management host, a PCIe switch, a first non-transparent bridge, and a second non-transparent bridge. The upstream port of the PCIe switch is electrically coupled to the management host. The first non-transparent bridge is disposed in the PCIe switch for electrically coupling to the first PCIe port of a calculation host. The first non-transparent bridge can couple the first PCIe port of the calculation host to the management host. The second non-transparent bridge is disposed in the PCIe switch for electrically coupling to the second PCIe port of the calculation host. The second non-transparent bridge can couple the second PCIe port of the calculation host to the management host.

Abstract: A system that uses a persistent main memory to preserve events that await logging in a persistent store. Each event is written into the persistent main memory so as to be loggable in case of recovery. For instance, the event may be written into a log cache structure, along with other state which identifies that the event is in the particular log cache structure, the location of the event within the particular log cache structure, and the order of the event. To recover, the log in the persistent store is evaluated to identify the end of the stored log. The tail of the log is identified in the persistent main memory by identifying any log cache structures that are after the end of the stored log and which are validly recoverable. The log cache structure contents are then serialized one log cache at a time, earliest first.

Abstract: In some example embodiments, a method comprises: receiving, by a first node of a plurality of nodes in a distributed database system on a shared disk cluster infrastructure, a transaction request to perform a user database transaction a data item in a user database on a shared disk; acquiring, by the first node, a transaction lock for the data item; storing a lock file for the user database transaction in a lock information database on the shared disk, the lock file comprising lock information for the transaction lock and an indication of a status of the user database transaction, and the lock information comprising an identification of a location of the data item; and storing a transaction record of the user database transaction in the user database on the shared disk subsequent to the storing of the lock file in the lock information database on the shared disk.

Abstract: A standby database cluster takes on the role of the primary database cluster if the primary database cluster becomes unavailable using the following steps: (i) operating a database management system (DBMS) including an initial primary cluster and a plurality of standby clusters; (ii) communicating to a set of client driver(s) connecting a first application to the initial primary cluster an identity of the plurality of standby clusters; (iii) on condition that the initial primary cluster becomes unavailable, assigning a selected standby cluster of the plurality of standby clusters to be assigned as a new primary cluster in place of the initial primary cluster; and (iv) in response to assignment of the new primary cluster, seamlessly moving the first application from the initial primary cluster to the new primary cluster without any substantial human intervention.

Abstract: The technology disclosed enables understanding the user experience of accessing a web page under high loads. A testing system generates a simulated load by retrieving and loading a single web object. A performance snapshot is taken of accessing an entire web page from the server under load. The performance snapshot may be performed by emulating a browser accessing a web page's URL, the web page comprising multiple objects that are independently retrieved and loaded. The simulated load is configured with a number of users per region of the world where the user load will originate, and a single object from the web page to retrieve. Performance data such as response time for the single object retrieved, number of hits per second, number of timeouts per sec, and errors per second may be recorded and reported. An optimal number of users may be determined to achieve a target user experience goal.

Abstract: Systems, methods, and computer-readable storage media for detecting smoke in a computer system. The system detects smoke via a first smoke detector located at an airflow inlet of the system or a second smoke detector located at an airflow outlet of the system. Next, the system determines whether the smoke originates from inside of the system or outside of the system to yield a determination of an originating location of the smoke. Based on the determination of the originating location of the smoke, the system then generates a response to the smoke.

Abstract: A crash dump system in a device includes firmware that determines when an atypical reboot of the device is about to occur and both sets a flag indicating there was an atypical reboot and stores in nonvolatile memory crash dump data based on the type of event that caused the atypical reboot. The crash dump data includes a reboot code indicating the type of event that was detected. When the device is subsequently restarted due to the reboot, the operating system detects that the flag was set, and in response to the flag being set collects and analyzes the crash dump data from the reserved memory. Crash analysis data is generated based on this analysis and sent is to a remote service for possible further action (e.g., analysis, notification of a developer, and so forth).

Abstract: The present disclosure includes detecting a failure associated with a first storage location on which a first agent virtual computing instance (AVCI) is deployed, wherein the first AVCI is being executed by a first hypervisor, stopping the execution of the first AVCI, determining whether a second AVCI that provides services analogous to the first AVCI is being executed by a second hypervisor and is deployed on a second storage location, creating a linked clone of the second AVCI on the second storage location responsive to the second AVCI being executed by the second hypervisor and deployed on the second storage location, redeploying the first AVCI on the second storage location responsive to the second AVCI not being executed by the second hypervisor or not deployed on the second storage location, and deleting files of the first AVCI from the first storage location after the failure is corrected.

Abstract: A computer-implemented method for testing a hybrid architecture instruction set is provided. The method includes defining a first instruction definition table for a first base architecture, and defining a second instruction definition table for a second base architecture, wherein the first base architecture is different than the second base architecture. The method also includes defining a delta table, wherein the delta table defines architecture specific behavior, and generating a hybrid architecture table based on the delta table and at least one of the first instruction definition table or the second instruction definition table. The method includes executing a test based on the hybrid architecture table, wherein the hybrid architecture table is for a hybrid architecture that is compatible between the first base architecture and the second base architecture.

Abstract: The present application describes techniques for providing dynamic at-risk system reports at the site of the at-risk system. The ARS may send information about system health (e.g., values for system health parameters) to a support entity, and may receive a risk assessment from the support entity. The ARS may, alternatively or in addition, receive support information from the support entity. Risk assessments and recommendations may be surfaced on a user interface of the ARS. New risk profiles, risk assessments, and support recommendations may be developed and loaded onto the support entity to provide dynamically updatable support solutions.

Abstract: In one aspect, a method includes mirroring a logical unit between a first site and a second site, tracking the changes to the logical unit on a backlog mirrored between the first site and the second site, replicating a virtual machine from the first site to a third site, storing virtual machine disks on a file system on the mirrored logical unit and resuming replication of the virtual machine to the third site from the second site using the mirrored logical unit after failure of the first site.

Abstract: Detecting error states when interacting with web applications is performed by accessing a first web page of a web application, determining that the first web page includes an input validation operation, configuring an input to cause the input validation operation to effect an error state, providing the input to the first web page, thereby effecting the error state, identifying a feature that is absent from the first web page before the input is provided to the first web page and present in the first web page after the input is provided to the first web page, and detecting that a second web page of the web application is in an error state if the feature is present in the second web page.

Abstract: Data files in a distributed system sometimes becomes unavailable. A method for fault tolerance without data loss in a distributed file system includes allocating data nodes of the distributed file system among a plurality of compute groups, replicating a data file among a subset of the plurality of the compute groups such that the data file is located in at least two compute zones, wherein the first compute zone is isolated from the second compute zone, monitoring the accessibility of the data files, and causing a distributed task requiring data in the data file to be executed by a compute instance in the subset of the plurality of the compute groups. Upon detecting a failure in the accessibility of a data node with the data file, the task management node may redistribute the distributed task among other compute instances with access to any replica of the data file.

Abstract: A plurality of storage nodes in a single chassis is provided. The plurality of storage nodes includes a first plurality of storage nodes configured to communicate together as a first storage cluster and a second plurality of storage nodes configured to communicate together as a second storage cluster. Each of the first and second pluralities of storage nodes has nonvolatile solid-state memory for user data storage and each of the first and second pluralities of storage nodes is configured to distribute user data and metadata associated with the user data throughout a respective plurality of storage nodes such that a respective storage cluster maintains ability to read the user data, using erasure coding, despite a loss of one or more of the respective plurality of storage nodes.