Hi,
i was so confused about the node recovery principles. In my cluster,there are 2 data nodes in one nodegroup,and tables are all in memory.When one data node crashed,all operations go to the surviving node.
My question are:
1. Then i restart the crashed node,how does it catch up with the surviving node? i Know that the crashed node would read the LCP and the redo log already resided in the crashed node,
but how to get the newer data already changed in the surviving node since the crash?
2.ok,after case 1,the starting node are retrieving the newer data from the surviving node, but at the moment, that data are being or to be changed .
After the case 1,the starting node still can not catch up with the surviving node.
i was so confused about the node recovery principles. In my cluster,there are 2 data nodes in one nodegroup,and tables are all in memory.When one data node crashed,all operations go to the surviving node.
My question are:
1. Then i restart the crashed node,how does it catch up with the surviving node? i Know that the crashed node would read the LCP and the redo log already resided in the crashed node,
but how to get the newer data already changed in the surviving node since the crash?
2.ok,after case 1,the starting node are retrieving the newer data from the surviving node, but at the moment, that data are being or to be changed .
After the case 1,the starting node still can not catch up with the surviving node.