Redis (VII) optimization suggestions

1、 What to do if the cache is inconsistent with the database

   Assume that the master-slave structure is adopted 、 Read write separate database .
   If a thread A Delete cached data first , Then write the data to the main library , This is the time , The synchronization between master and slave is not complete , Threads B Failed to read data from cache , Read old data from the library , Then update to the cache , This is the time , What's in the cache is the old data .
   The reason for this inconsistency is ： Data inconsistency between master and slave databases , After adding the cache , The time of inconsistency between master and slave has been lengthened .
   Handling ideas ： After the data is updated from the library , Update the data in the cache at the same time , That is, when the data is updated from the database , Issue delete to cache , Discard the old data written during this period .

2、 How to solve the inconsistency between master and slave databases

   Scene description ： For master-slave Libraries , Read / write separation , If there is a time difference between master and slave update synchronization , It will lead to inconsistency between the master and slave databases .
   Handling ideas ：

• 1、 Ignore this data inconsistency , In the business with low data consistency requirements , Real time consistency is not necessary .
• 2、 Forced read main database , Use a highly available main library , Read and write the database in the main database , Add a cache , Improve the performance of data reading .
• 3、 Read the main library selectively , Add a cache , Used to record data that must be read from the main library , Specify which library to use 、 Which watch 、 Which primary key is cached key, Set the cache expiration time to synchronize the master and slave libraries . If there is this data in the cache , Read the main library directly , If there is no primary key in the cache , Read from the library .

3、Redis Common performance problems and solutions

  1、Master It's better not to do persistence work , Such as RDB Memory snapshot and AOF Log files .
  2、 If the data is important , Some Slave Turn on AOF Backup , The policy is set to synchronize once per second .
  3、 Try to avoid adding slave libraries to the main library with high pressure .
  4、 Master-slave replication should not adopt mesh structure , Try to be linear ,Master <— Slave1 <— Slave2 wait .