Hi, I currently work with 3 node xtrab cluster.
- Node 1 is ONLY used for write operations
- Nodes 2 & 3 are used for read operations
Everything works fine normally, but I'm currently having problems with the following situation:
Whenever I insert a large amount of rows into a table (>100K), although it is Innodb, and no other process is using it, all other DML operations get stuck for serveral seconds (even minutes sometimes), this occurs when the insert is finished in the writing server and it has to replicate the change in the other nodes.
Is there anyway to avoid this situation?
Bruno