Redisson distributed lock source code analysis for high-level use of redis

1. What is a distributed lock
1. Lock synchronized lock Stand-alone lock
2. Some of our mutually exclusive resources cannot be executed in parallel. We need something to ensure that the lock executed in serial is synchronized
3、Distributed lock: This lock storage must be independent of our business service
2. Design a lock by yourself, how to achieve
Requirements:
1. Only one thread can execute mutual exclusion at the same timeresources
2. When other threads are executing, if there is a thread executing, either wait or report an error

implementation steps:
1. There is a mark to mark whether there is a thread executing the toilet pit door lock state=1 Someone is doing it=0 No one is doing it
2. This mark must be visible to ensure mutualThis mark is volatile
that can get the latest result. 3. Obtaining this mark cannot simultaneously preempt the successful safe cas

3. How to do distributed locks in Redis
1. Mark the k-v structure key as a mark. If the key exists, it means that someone is doing it. If it does not exist, it means that no one is doing it.
2. Visibility listThe thread must be set before you can get
3. Ensure atomic setnx single-line execution
if(exists("pay:11") == 0) {
set("pay:111",1);
}
4. We hope that no other instructions can be inserted between exists and set2 instructions. To ensure the atomicity of multiple instructionsSexual mulit open transaction - exec commit discard rollback

Four. Transaction
1. The command is atomic, but the subsequent logic cannot be determined according to the intermediate command

Five, lua script
1. The host redis must support lua script perfectly
2. eval "return 'huihui'" 0
3. eval "if KEYS[1]=='1' then ARGV[1] end return ARGV[2]" 1 1 'huihui' 'xiaohuihui' The first parameter is the number of KEYs, the second is the value of KEY, the third is the value of ARGV1, and the fourth is the value of ARGV2Value
4. How to do duplicate locks?The same thread can be locked and added multiple times
4.1 Mutually exclusive key Big key
4.2 Know the thread information field
4.3 Save the number of flushes value hincrby thread-safe
4.
5.If the 30s lock will expire, but the business will fail if the lock is not executed.

Six, Redission lock principle source code analysis
1.lua + hash to do it
if (redis.call('exists', KEYS[1]) == 0)
then " +
"redis.call('hincrby', KEYS[1], ARGV[2], 1); " +
"redis.call('pexpire', KEYS[1], ARGV[1]); " +
"return nil; " +
"end; " +
"if (redis.call('hexists', KEYS[1], ARGV[2]) == 1) then " +
"redis.call('hincrby', KEYS[1], ARGV[2], 1); " +
"redis.call('pexpire', KEYS[1], ARGV[1]); " +
"return nil; " +
"end; " +
"return redis.call('pttl', KEYS[1]);",If there is an expiration date?
3. Check this thing regularly, if it is, I will reset the expiration time for it! If it is not, I will not set it
4. The hashwheelTimer is also a scheduled task

7. Source code analysis of the time wheel algorithm principle of watchDog watchdog mechanism

1. Interlocking: Give you multiple master nodes + the same lock at the same time
When Redis is doing distributed locks, there will be multiple losses ap model
Reduce losses

Interview questions:
1. There is a foot bath shop and their technicians are arranged in this way. There are 8 technicians in total.
Technician 1: Responsible for guests 0-3
Technician 2: Responsible for 3-6 guests
Technician 3: responsible for 6-9 guests
Technician 4: responsible for 9-12 guests
Technician 5: responsible for 12-15 guests
Technician 6: responsible for 15-18 guestsGuest
Technician 7: In charge of guests 18-21
Technician 8: In charge of guests from 21-24
If there is a day off at 6:00 in the morning, I want to book a massage for 9 hours later, who will I do?Technician for me?
6 o'clock 6+9=15 hours to go to massage 6 after 15 o'clock, there are already people in line, you have to line up too

If the current time is 21:00, I would like to make an appointment for which technician will come to service after 6 hours
21+6 27 hours 2