January 2019
Intermediate to advanced
286 pages
7h 41m
English
For consistency, a majority of the servers should agree on a state change for any given data. In that case, it might happen that some servers lag behind a lot, and it has to continuously play the game of syncing transactions from other servers, which are serving transactions very fast. Due to the slowness of some of the servers, the entire cluster may be slow. To make sure that this situation doesn't occur, group replication has come up with a flow control mechanism.
In this mechanism, the servers that can execute more quickly reduce their speed to allow the slow servers to catch up quickly, rather than burdening them continuously. When the group replication flow control mode is set to QUOTA, and when the faster node realizes ...