Abstract: Techniques are disclosed for backup and restore in a distributed database utilizing consistent database snapshots. In particular, a distributed database system configured in accordance with an embodiment of the present disclosure includes a plurality of interconnected database nodes that collectively define a database having no single point of failure and that can be “viewed” by a SQL client as a single, logical database. In the course of executing such transactions, the distributed database system enables clients to declare snapshots that, when the transaction is committed, cause the distributed database system to logically freeze a consistent and complete copy of the state of the database at the point in time the transaction completed. Such snapshots enable clients to execute so-called “point-in-time” queries against those past states to return a result set that is essentially a reconstructed version of the database “as of” a particular point in time.

Abstract: Techniques are disclosed for disconnected operation in a distributed database system. In an embodiment, the distributed database system implements a disconnected mode of operation allowing isolated regions of database nodes to provisionally commit transactions, with the global requirements of those transactions later satisfied by a transparent healing process after network connectivity is reestablished between the regions. The healing process, in turn, enables the distributed database system to construct a consistent global state of the database that accounts for the transactions provisionally-committed in each isolated region during the disconnected mode. Once the healing process completes, database clients “see” a healed version of the database that simulates or otherwise closely approximates a state of the database had the distributed database system performed the transactions during a normal, fully-connected, mode of operation.

Abstract: Techniques are disclosed for implementing a unified partitioning scheme within distributed database systems to allow a table to be horizontally partitioned and those partitions stored on and serviced by a storage group. A storage group is a subset of storage manager (SM) nodes, and each SM node is configured to persist database data in durable storage. The distributed database system assigns each storage group to a subset of SM nodes. The distributed database system can address each storage group using a symbolic mapping that allows transactions to identify a particular storage group, and to direct read and write operations to a subset of SM nodes servicing that storage group. An administrator can update this mapping on-the-fly to cause the distributed database system to dynamically adjust an implemented partitioning scheme without necessarily interrupting on-going database operations.

Abstract: In a distributed database, many nodes can store copies, or instances, of the same record. If the record is split on one node, it should be split on the other nodes to maintain consistency, concurrency, and correctness of the data in the distributed database. In some distributed databases, the records are locked during the update process to ensure data integrity. Unfortunately, locking the records can increase latency, especially for larger databases. But if the records aren't locked and a node fails as a record is being split and updated simultaneously, the split and update may not propagate throughout the distributed database, leading to a loss of data integrity. Exchanging messages about the status of record splitting and forwarding updates internally reduces the likelihood of a loss of data integrity due to a node failure.

Abstract: Methods and systems to detect and resolve failure in a distributed database system is described herein. A first node in the distributed database system can detect an interruption in communication with at least one other node in the distributed database system. This indicates a network failure. In response to detection of this failure, the first node starts a failure resolution protocol. This invokes coordinated broadcasts of respective lists of suspicious nodes among neighbor nodes. Each node compares its own list of suspicious nodes with its neighbors' lists of suspicious nodes to determine which nodes are still directly connected to each other. Each node determines the largest group of these directly connected nodes and whether or not it is in that group. If a node isn't in that group, it fails itself to resolve the network failure.

Abstract: Methods and systems to detect and resolve failure in a distributed database system is described herein. A first node in the distributed database system can detect an interruption in communication with at least one other node in the distributed database system. This indicates a network failure. In response to detection of this failure, the first node starts a failure resolution protocol. This invokes coordinated broadcasts of respective lists of suspicious nodes among neighbor nodes. Each node compares its own list of suspicious nodes with its neighbors' lists of suspicious nodes to determine which nodes are still directly connected to each other. Each node determines the largest group of these directly connected nodes and whether or not it is in that group. If a node isn't in that group, it fails itself to resolve the network failure.

Abstract: A distributed database processing system for a database composed of data records organized into tables which processes unique index atoms consistently and concurrently. Each attempt to insert a new key value into such a unique index atom in any given node is routed to a unique index atom chairman for processing. The chairman determines whether the request will be granted. If the request is not granted, the requesting unique index atom continues to try to submit the insert. If the request is granted, the requesting unique index atom modifies the content thereof and broadcasts a replicated given unique index atom all other nodes that contain the replicated unique index atom.