Hi all,
I am using mysql-cluster 7.0.9 for testing.
There are two SQL nodes, two Data nodes and one management node in my
mysql-cluster.
I configure JDBC url to implement load balancing. The JDBC URL is like
this:
DS_mmvd_URL=jdbc:mysql:loadbalance://192.168.1.19:33306,192.168.1.
36:33306/mmvd?autoReconnect=true&autoReconnectForPools=true&loadBalanceB
lacklistTimeout=1000&failOverReadOnly=false&loadBalanceStrategy=random
I test it and it can implement load balancing.
But I find one big problem.
After I shutdown one SQL node, it can switch to the other available SQL
node.
But some data will be lost. For me, I insert 100 rows into the test
database, 20 rows are lost when I shutdown one SQL node while executing
the SQL operation.
I don't know the real reason. If it can implement load balancing by the
way of configuring the JDBC URL, it should also implement fail-over
without losing any data.
If it loses datas, it can't achieve genuine HA.
If it can't avoid losing datas, then how to implement genuine HA without
losing any data if one SQL node out of service?
I need your help, thanks.
Peter
