Rubrik, Inc.

Brian Yin

Pin Zhou

Systems and methods for shard consistency in a clustered database are described. The system identifies an addition of a first shard to a clustered database on a source storage platform. The clustered database includes at least two shards and the first shard includes a first node. The system communicates a first node agent to the first node. The node agent being is configured to execute on the first node to extract a data image and an operation log at the source storage platform and stream the data image and the operation log to the secondary storage platform. The system versions a first version of the clustered database. The versioning includes storing the data image and operation log for the first node at the second storage platform. The timestamp associated with the first shard is different than the timestamp for other shards in the clustered database.

Systems and methods for shard consistency in a clustered database

Systems and methods to process a topology change in a clustered database are described. The system communicates a query to a source storage platform enquiring of a topology change in a clustered database stored on the source storage platform. The system receives a message, at a control computer responsive to communicating the query. The message includes node identifiers identifying nodes including a first node identifier identifying a first node included in the clustered database. The system automatically identifies an addition of the first node to the clustered database on the source storage platform. The system initializes the first node on the source storage platform by communicating a first node agent to the first node. The first node agent is configured to execute on the first node to extract the data image at the source storage platform and stream the data image to the secondary storage platform.

Systems and methods to process a topology change in a clustered database

Systems and methods for node consistency in a clustered database are described. The system automatically identifies an addition of a first node to a clustered database on a source storage platform. The system initializes the first node on the source storage platform communicating a first node agent to the first node. The first node agent is configured to execute on the first node to extract the operation log at the source storage platform and stream the operation log to the secondary storage platform. The system versions a first version of the clustered database. The versioning of the first version including not storing a version of the operation log associated with the first node at the secondary storage platform responsive to identifying, during the first interval of time, the synchronizing of the operation log at the first node having not been synchronized, in a prior version of the clustered database, with the operation log at the storage platform.

Growing community of inventors

San Jose, CA, United States of America

Brian Yin