Off node data compaction

A system for compacting datafiles, comprising:

• a first data store;
• a database node attached with the first data store, the database node configured to:
• store datafiles in the first data store;
• and
• generate a snapshot of the first data store including the datafiles;
• a second data store attached with a compaction processing node, the second data store configured to store the snapshot;
• and
• the compaction processing node, the compaction processing node being an initiated instance of a cloud computing infrastructure, the compaction processing node including first compute resources that are separate from second compute resources of the database node, the compaction processing node configured to:
• generate a compacted datafile by compacting the datafiles of the snapshot in the second data store;
• store the compacted datafile in the second data store;
• and
• shut down after generating the compacted datafile;
• and
• wherein the database node is further configured to:
• attach with the second data store including the compacted datafile subsequent to the compacted datafile being stored in the second data store;
• and
• write a second datafile in the attached second data store, the second data file being stored in the second data store with the compacted data file.

A non-transitory computer readable medium storing instructions that when executed by a processor cause the processor to:

• store datafiles in a first data store attached with the processor;
• generate a snapshot of the first data store including the datafiles;
• store the snapshot in a second data store attached with a compaction processing node, the compaction processing node being an initiated instance of a cloud compute infrastructure, the compaction processing node including first compute resources that are separate from the processor;
• responsive to the compaction processing node generating a compacted datafile by compacting the datafiles of the snapshot in the second data store, attach with the second data store including the compacted datafile and shut down the compaction processing node;
• and
• write an additional datafile in the attached second data store, the additional data file being stored in the second data store with the compacted data file.