Double-write consistency means that when we update the data in the database, the data in redis must also be updated synchronously. The process of reading data using redis. When the user accesses the data, the data will be read from the cache first. If the cache is hit, the data in the cache will be directly returned to the user. If there is no data in the cache, the database will be queried first. Save the queried data to the cache and then return it to the user.
1. Update the cache first, then update the database
2. Update the database first , then update the cache
3. Delete the cache first, then update the database
4. Update the database first, then delete the cache
1. Update the cache first, then update the database
The problem is obvious. If the cache is updated successfully but the database update fails, dirty data in the cache will be caused.
2. Update the database first, then update the cache
If the concurrency is high, the following situation may exist. Thread A updates the database. If due to network or other reasons, Thread A has not had time to update the cache. At this time, a process B updates the database and updates the cache. Only then does process A update the cache. This will cause thread B to lose its update to the cache, like a transaction loss situation
3. Delete the cache first, and then update the database
This strategy may have avoided the cache loss in Strategy 2, but under high concurrency conditions , there will also be inconsistencies. For example, when thread A performs a write operation, it first deletes the cache and then prepares to communicate with the new database. At this time, thread B performs a write operation without hitting the cache, and then queries the database. At this time, the old value is read. , and save the queried old value to the cache. Then thread A completes the update of the database. At this time, the database and cache are inconsistent again. Solution: We only need to re-thread A. After completing the update of the database, Delaying and then deleting the cache is also called delayed double deletion. The delay time here must be greater than the time of a read operation of the business.
4. Update the database first, then delete the cache
No matter how high the concurrency is, there will be inconsistencies, such as when thread A reads data. , when preparing to write to the cache, thread B updated the database, and then performed the delete cache operation. At this time, thread A wrote the old value into the cache, although the probability of this happening is relatively low, because the write operation time is greater than the time of a read operation. Solution: Delayed double deletion. Delayed double deletion is still a problem. What should I do if deleting the cache fails? Of course, delete it again and continue deleting in a loop. After the deletion fails, we can put the key to be deleted into the queue, and then try to delete it repeatedly until the deletion is successful.
The above is the detailed content of Ensure the double-write consistency of mysql and redis. For more information, please follow other related articles on the PHP Chinese website!