Scale expansion overview

For distributed databases with massive storage, the scale of table data increases over time, which is also inevitable that there will be insufficient capacity and computing power (CPU). At this time, a correct and efficient cluster expansion becomes indispensable. Although it is wisest to leave room for growth when the system is initially defined, it is usually unlikely to invest too early in resources. Therefore, we recommend that you plan and perform database scaling operations regularly.

In YMatrix, you can freely choose to use graphical interface expansion or command line online expansion. We recommend using a simple and easy-to-operate graphical interface. At present, YMatrix has supported online smooth expansion (reading and writing services are not affected by any impact during the expansion), flexible setting of data distribution in tables and other functions.

FAQ

Are the catalogs (catalogs) of new and old nodes consistent during the expansion process?
- catalog is always consistent, and catalog exists even if there is no data on the new node.
Impact on other queries during the expansion process
- Because it is an online expansion, all running queries will not be interrupted during the period. It is just that when adding a new node, the catalog is locked, so the DDL-related operations are not supported during this period. After adding a new node, the catalog lock will be released immediately.
- When data redistribution is added, the read operation will be blocked.