What are MemStore and StoreFile in HBase?

10 months ago

Noah Thompson

2 minutes

In HBase, MemStore is a memory storage area used to store newly written data and updates. When a write request arrives at HBase, the data is first written to the MemStore and then asynchronously written to the StoreFile on disk in the background. MemStore holds recent data written in memory, and when it reaches a certain size, a flush operation is triggered to write the data to the StoreFile on disk.

StoreFile is a persistent storage file in HBase used to store written data. It is stored in HDFS in the HFile format, allowing it to be read and queried by HBase. Once data from MemStore is flushed into StoreFile, MemStore is cleared and ready for the next writing operation. Data in StoreFile is accessed and queried by HBase through a Scanner.