Three expiration strategies

1. Delete regularly
meaning ： Set up key At the same time , For the sake of key Create a timer , Let the timer be in key When the expiration time of , Yes key To delete

advantage ：
Make sure the memory is released as soon as possible

shortcoming ：
1. If overdue key quite a lot , Delete these key It will take up a lot of CPU Time , stay CPU When time is tight ,CPU You can't spend all your time doing something important , And take the time to delete these key

2. Timer creation time consuming , If you set the expiration time for each key Create a timer （ There will be a lot of timers ）, Serious performance impact

3. Nobody uses

2. Lazy deletion

meaning ：
key Do not delete when expired , Get... From the database every time key Check if it's overdue , If overdue , Delete , return null.

advantage ：
The delete operation only occurs when fetching from the database key When it happens , And only delete the current key, So for CPU Time is less , And the deletion at this time has reached the point where we have to do something （ If it is not deleted at this time , We will get the expired key 了 ）

shortcoming ：
If a large number of key After exceeding the timeout period , For a long time , Have not been acquired , Then there may be a memory leak （ Useless garbage takes up a lot of memory ）

---

Delete periodically

meaning ：
Delete every once in a while ( stay redis.conf Profile Settings hz,1s Refresh frequency ) Be overdue key operation

advantage ：
1. By limiting the duration and frequency of delete operations , To reduce the number of delete operations CPU The occupation of time – Handle " Delete regularly " The shortcomings of
2. Periodically delete expired key– Handle " Lazy deletion " The shortcomings of

shortcoming
1. In terms of memory friendliness , Not as good as " Delete regularly "
2. stay CPU Time friendly , Not as good as " Lazy deletion "

difficulty
Reasonably set the execution time of delete operation （ How long does each deletion take ） And execution frequency （ How often do I delete ）（ This depends on the operation of the server ）

---

After reading the above three strategies, we can draw the following conclusions ：

1. Regular deletion and regular deletion are active deletion ：Redis Will regularly take the initiative to eliminate a number of past key

2. Lazy delete is passive delete ： It's only tested when it's used key Is it overdue , Delete after expiration

3. Inertia deleted as redis Server built-in strategy

Regular deletion can be done through ：

First of all 、 To configure redis.conf Of hz Options , The default is 10 （ namely 1 Seconds to perform 10 Time ,100ms once , The higher the value, the faster the refresh rate , most Redis The greater the performance loss ）

second 、 To configure redis.conf Of maxmemory Maximum , When used memory exceeds maxmemory When limited , Will trigger the active clean-up strategy