Linux server development, MySQL cache strategy

Recommend a free open course of zero sound College , Personally, I think the teacher spoke well , Share with you ：Linux,Nginx,ZeroMQ,MySQL,Redis,fastdfs,MongoDB,ZK, Streaming media ,CDN,P2P,K8S,Docker,TCP/IP, coroutines ,DPDK Etc , Learn now
MySql What are the ways to improve read and write performance ？

• Connection pool Blocking io+ Thread pool
• Asynchronous connection Non blocking io
• sql Execute punishment ： Carry out in a timely manner + Precompile execution
• Read / write separation
• Cache scheme
  Master-slave replication solves the problem of single point of failure .

Reading strategy

• First look at cache There's no data in it , There is a direct return .
• without , Go again MySql Take the data from the library .
• If MySql There are data in it. , The cache to cache
• without , That is, there is no

Final consistency and strong consistency .

Write strategy

• Delete first cache, And then delete MySql
• Delete first cache, Revise MySql. Modify cache directly after improvement , Set expiration time , modify MySql Remove the expiration time .
• Delete first cache, Add again MySql. Modify cache directly after improvement , Set expiration time , modify MySql Remove the expiration time .

Asynchronous operation is compared with synchronous operation , It mainly saves the time of network transmission .

Cache penetration

redis and MySql None of them exist ,redis No data will always be requested MySql, and MySql There is nothing but to return . When a large number of requests accumulate to MySql, Finally, I was overwhelmed .、
Solution ：

• Find out MySql non-existent , take redis Set up <key,nil> Set expiration time , The next visit is not visit MySql, It is easy to cause many invalid redis Cache data .
• Set up the bloon filter , The non-existent hotspot data is directly filtered out .

Cache breakdown

redis No, ,MySql There's data . When there are a large number of concurrent requests , Request to pile up MySql Too big . One of them Key A lot of requests .

• Lock , Operate with lock , Sleep without lock .
• Will be a lot hot key Set not to expire .

Cache avalanche

redis Downtime , All data are empty , Cause collapse MySql. Multiple Key A lot of advice .

• Highly available cluster solutions , Like sentinel mode ,cluster Pattern .
• Set the random period value or other mechanism to stagger the expiration time .
• redis Turn on persistence .