Directed backup for massively parallel processing databases
Creating a data backup of data on a first computer system to restore to a second computer system, each of the first and second computer system including one or more nodes, each node configured to manage a subset of the data. Receiving, by the first computer system, identification of data to back up and node configuration information for the second computer system. Creating, by the first computer system, a backup of the data from the one or more nodes of the first computer system, configured in accordance with the node configuration information of the second computer system, such that the backed up data is directly manageable by the one or more nodes of the second computer system.
1. A method for creating a backup of a database on a first massively parallel processing (MPP) computer system to restore to a second MPP computer system, the method comprising:
receiving, by the first MPP computer system, identification of data to back up and node server configuration information for the second MPP computer system, wherein the node server configuration information includes how the data is partitioned among the node servers of the second computer system, and wherein each of the first and second MPP computer systems includes its own database storage media segmented into storage segments, and a plurality of node servers, each node server configured to manage a respective storage segment and a subset of the database stored on its respective storage segment, and wherein partitioning of the database among the node servers of the first MPP computer system is different than the partitioning of the database among the node servers of the second MPP computer system; and
creating, by the first MPP computer system, a backup of the identified data from the plurality of node servers of the first MPP computer system in accordance with the node server configuration information of the second MPP computer system, such that the backed up data is partitioned in accordance with the partitioning of the data among the node servers of the second MPP computer system and does not require a re-partitioning by the second MPP computer system for a restore of the data to the second MPP computer system.
2. The method according to claim 1 , wherein receiving further comprises:
receiving, by the first MPP computer system, an identifier for the second MPP computer system;
transmitting to the second MPP computer system, by the first MPP computer system, a request for the node server configuration information; and
receiving, by the first MPP computer system, the node server configuration information.
3. The method according to claim 2 , further comprises:
transmitting to the second MPP computer system, by the first MPP computer system, the backup of the data from the plurality of node servers of the first MPP computer system.
4. The method according to claim 1 , wherein partitioning the data from each of the plurality of node servers of the first MPP computer system executes in parallel.
5. The method according to claim 1 ,
wherein the node server configuration information for the second MPP computer system includes one or more of: an indication of a data storage format for data on the plurality of node servers of the second MPP computer system, and an indication of a data compression algorithm for data on the plurality of node servers of the second MPP computer system; and
wherein creating further comprises one or more of: formatting the data from the plurality of node servers of the first MPP computer system, in accordance with the received indication of the data storage format; and compressing the data from the plurality of node servers of the first MPP computer system, in accordance with the received indication of the compression algorithm.
6. The method according to claim 5 , wherein formatting the data from each of the plurality of node servers of the first MPP computer system executes in parallel, and/or compressing the data from each of the plurality of node servers of the first MPP computer system executes in parallel.