mysql cluster data disk increase more and more

Question

I have mysql cluster two management four api nodes and four data nodes. now I have a question,the data node disk space unceasing growth。The data nodes have 2TB of storage space.Datamemory configured 170GB of space,the indexmemory is configured with 30GB of space,FragmentLogFileSize configuration for 256 m, NoOfFragmentLogFiles is 512.How I should be able to make the disk space for data nodes no longer grow.

Answer 1

Each log part is configured to use 128 GByte of REDO log files. It doesn't say how many log parts you have. So I assume you use the default which is 4. That is you have 512 GByte of REDO log files.

The Local checkpoint files can store around two times the DataMemory at most (if all DataMemory is used). So this part can grow to about 350 GBytes.

So it should not grow beyond around 850 GBytes.

I presume since you say it is growing that you started without initialising REDO log files. This means that REDO log files will continue using disk space until they come to 512 GBytes.

If you want to decrease the amount of REDO logs you can change the configuration and perform an initial node restart for one node at a time. Make sure to do a backup before any such change.

The size of local checkpoint files is about the same as twice the size of the data in DataMemory used. So if you insert more data into NDB it will use more file space up to the size of the DataMemory.