1, single table data size 1 billion;
2, the user of the system is small, almost no concurrent operations;
3, the business is given priority to with analysis, the data used in the single business size 1 million,

From the cost and complexity of consideration, giving up depots;
Current solution is to design a table partition, use table partitioning the data scale down to 10 million, combining reading and writing separation scheme improve performance;

Friends have a similar case, a better solution or suggestion,

CodePudding user response:

Used mycat, is a mysql database cluster middleware, studied the no, don't know the original poster is appropriate for the building

CodePudding user response:

The volume of this table is fixed?
If is not fixed, new every day?
If the partition, according to what?

"The data used in the single business scale 1 million"
According to what is on the screen?

CodePudding user response:

reference 1/f, February 16 response:

used mycat, is a mysql database cluster middleware, studied the no, don't know the original poster is appropriate for the host's scene

Not seen mycat, feel with
I this QPS and TPS system business requirements are very low, is the amount of data is large, the monthly data increment in 50 million

CodePudding user response:

refer to the second floor of gypsy song response:

the data volume is fixed this table?
If is not fixed, new every day?
If the partition, according to what?

"The data used in the single business scale 1 million"
According to what is on the screen?

Every time the operation of the business analysis using data, is about 1 million

CodePudding user response:

Curious bosses finally how to implement,