Redis' cache penetration, cache breakdown, cache avalanche

Prior to redis In the article of , We said redis Master-slave replication of , Now let's talk about redis Frequent concentration problems

1, Cache penetration , It refers to data not in the cache and not in the database , This will result in cache misses , Because there is no such data in the database , So it will not be written to the cache after the request , This will lead to direct access to the data source , Result in crushing the data source

2, Cache breakdown , Refer to key The corresponding data exists , But it expired in the cache . At this time, a large number of highly concurrent data requests come , It will lead to direct access to the data source because there is no cache , After accessing, you need to write the data back to the data source , This will easily overwhelm the data source

3, Cache avalanche , This is when the cache server restarts or the cache fails within a time period , This will invalidate all caches , To crush the data source

Solution

For cache penetration , I know about two ways

The first is to use a large enough bitmap( use bitmap, Because it can save more space ) Put all possible key It's all in it , In this way, non-existent requests will be filtered out

The second is to use a null value storage , After a nonexistent request calls , Not in cache , Will go to the data source to check , At this time, because there is no data source, we will not write back to the cache , But to avoid cache penetration , We can also save this null value , It's just that the expiration time is set shorter .

Cache breakdown is now commonly used in the way of locking , Xiaobian is a little sleepy , I'll share these with you next time .