How to solve cache avalanche, breakdown and penetration problems

Work together , Grow up together ！ This is my participation 「 Nuggets day new plan · 8 Yuegengwen challenge 」 Of the 1 God , Click to see the event details

Preface

Reids As a popular distributed cache , Cache penetration will be encountered in the actual production environment 、 Cache breakdown 、 Cache avalanche and other abnormal scenarios , In order to avoid the huge loss caused by the abnormality , We need to understand the causes of each exception and the corresponding solutions , To improve system reliability and availability .

Cache penetration

brief introduction

The data requested by the user does not exist in the cache or database , Users need to query the database every time they request data , This leads to frequent access to the background database , Database load pressure increases , This phenomenon is called cache penetration .

The reasons causing

explain ：

• Massive access to nonexistent key, Causes the database to process a large number of requests

Solution

There are two ways to solve cache penetration , The first is to cache empty objects , The second is to use a bloom filter .

Caching empty objects

When data cannot be found in the database , Caching empty objects , Then set the expiration time for the cache of the empty object , Check again next time Data time , Get directly from the cache , Thus, the purpose of reducing the pressure on the database is achieved .

Program drawback

• The cache layer needs to provide more memory space to cache empty objects , Waste more memory space .
• Even set a short expiration time when caching empty objects , It will also lead to data inconsistency during this period .

The bloon filter

The Bronx filter will be explained in detail in the subsequent articles , This article does not elaborate .

Cache breakdown

brief introduction

Redis When there are some hot data in , That is, there are a large number of requests for concurrent access key-value data . Extreme hot spot key-value When the data suddenly fails , Cache misses cause frequent access to the background database , This phenomenon is called cache breakdown .

The reasons causing

explain ： Hot data expired , Cause a large number of query requests to cross the cache , Direct query database .

Solution

There are two solutions to cache breakdown , The first is to set up key Never expire ; The second is to use distributed locks , Ensure that only one query request can reload hotspot data into the cache at the same time , such , Other threads only need to wait for the thread to finish running , You can restart from Redis Get data in .

Set up key Never expire

Setting hotspot key When , Do not give key Set the expiration time .

Distributed lock

For hot spots key Using distributed locks , When a large number of queries the same key The request of , Only one request can acquire the lock , Query the database , Then put the results into the cache , Then release the lock , here , Other requests waiting for locks can continue , At this time, there is data in the cache , So get the data directly from the cache and return , Does not query the database .

Cache avalanche

brief introduction

The data requested by the user does not exist in the cache or database , Users need to query the database every time they request data , This leads to frequent access to the background database , Database load pressure increases , This phenomenon is called cache penetration .

The reasons causing

explain ：

• The first is a large number Key Expired at the same time
• The second is Redis fault

Solution

For the first kind of large Key At the same time

• Set each key The expiration time of should be as different as possible , You can add a random length of time to the timeout , Make their failure points as evenly distributed as possible .

• The heat data can never expire

For the second kind of Redis failure

• use Redis Deployment of several high availability solutions , For the specific deployment plan, please refer to what I wrote before Redis High availability architecture .

summary

This article explains the avalanche of caching 、 breakdown 、 Causes and solutions of penetration problems , If you have any questions , Please feel free to give feedback .