Inventory the six second level capabilities of Huawei cloud gaussdb (for redis)

Abstract ： Count Gauss Redis Second level capability , Including expansion 、 Backup 、 Delete 、 Start up, etc .

This article is shared from Huawei cloud community 《​ ​ Hua Wei Yun GaussDB(for Redis) Uncover secrets 20 period ： Six second level capability inventory ​​》, author ： gaussian Redis The official blog .

introduction

People often use time indicators to evaluate the database capacity , Such as RTO/RPO、 Expansion takes time 、 How long the change operation will affect the business , With the increasingly urgent demand for high availability , These are important references for evaluating the database capability . stay KV Database domain , Hua Wei Yun GaussDB(for Redis) Infrastructure based on the separation of storage and calculation , There are... In many key evaluation dimensions “ Second level ” Ability , Today I will show you about .

GaussDB(for Redis) Second level capability inventory

Seconds to start

Community Edition Redis In the scenario of downtime recovery , Need to rely on AOF or RDB File loads all data into memory , As the amount of data increases , This time-consuming process often takes several minutes or more .

GaussDB(for Redis) There is no need to load full data when the node is pulled up , You can provide services quickly , The startup time is very little affected by the amount of data , It can also be started in seconds under the large data scale , Experience silky .

Seconds to delete

Community Edition Redis In the delete / Expired data , Especially big key It will cause severe access congestion .

and GaussDB(for Redis) It has fundamentally solved the problem of "big..." key Delete / Out of date operational hazards , In the bottom layer, the “ Mark deletion , Asynchronous recycling ” The logic of , Delete any data / Be overdue , Both are executed successfully and immediately 0 Blocking , So it doesn't affect business access at all . Delete in actual measurement / One big overdue hash key（ contain 1000w Elements ） when ,GaussDB(for Redis) It only takes milliseconds .

chart ：GaussDB(for Redis) Big key Second delete

Second backup

Community Edition Redis The full amount of data needs to be dump To disk , Efficiency is affected by disk IO Performance impact , It takes a long time . And with “fork problem ”, Cause performance jitter , As a result, the capacity utilization rate is only 50%.

GaussDB(for Redis) Backup snapshot technology based on the underlying file system , Record the data status at a certain point in time , No time-consuming data copy operations , Truly take a snapshot at the second level . The customer's data snapshot file is automatically uploaded to Huawei cloud OBS In barrel , Later, it can be used to restore to the new instance with one click , Quite convenient .

chart ：GaussDB(for Redis) One click recovery of point in time snapshots

Second expansion

Community Edition Redis Calculation 、 Storage is strongly coupled . such as 64G The specifications are full , When the need arises 128G Specification time , The upgrade process is not only time-consuming , Serious impact on business ; And the price doubled , There is a lot of waste of computing power .

GaussDB(for Redis) Provide the following 3 A flexible way to change “ Capacity expansion ” The way ：

• 1） The data is almost full ？ One click capacity expansion , Seconds to complete , It doesn't affect the business

GaussDB(for Redis) The underlying storage resources are provided and used in the form of quotas , The data will soon fill the scene , Just one click for capacity expansion , Get more capacity in seconds . such as 64G Expand to 128G, Seconds to complete , It doesn't affect the business . At the same time, there is no need to pay for unnecessary computing power , The expansion price increases slightly , Cost has great advantages .

chart ：GaussDB(for Redis) Storage capacity

• 2） Sudden peak traffic ？ One click plus nodes , Business is only affected by seconds

GaussDB(for Redis) The most convenient way to expand the calculation force is to add nodes , Load rebalancing can be completed in seconds , Business is temporarily affected .

chart ：GaussDB(for Redis) Computing capacity expansion - Add a node

• 3） Do activities periodically ？ Node specifications are scalable on demand , Avoid wasting your computing power

In some periodic activities, there are scenes of securities grabbing , Customers often need to raise and lower the specifications according to the cycle , Carrying planned traffic peaks , At the same time, avoid the usual waste .GaussDB(for Redis) Support to upgrade node specifications , such as 16U l 32U, Customers can flexibly expand capacity on demand .

chart ：GaussDB(for Redis) Computing capacity expansion - Node specification change

Second recovery from failure

GaussDB(for Redis) There is a global HA Control ability , Compute node fault second level detection , After the fault is found, the second level takeover method is also adopted , Restore the normal access of the service side at the second level .

Double active second level RPO

GaussDB(for Redis) Support region Inside / Span region disaster , That is, build a stable master-slave instance , Provide enterprise level high availability . The high-performance master-slave instance synchronization is realized through the high-performance data synchronization module , Test under the business pressure of high pressure writing , with RPO=10s Ultra high reliability performance .

summary

Through the above simple inventory, it can be seen that ,GaussDB(for Redis) Start up 、 Big key Delete 、 It has second level capability in various application scenarios such as capacity expansion and change , Minimize business impact , Provide enterprise level security for users 、 Stable 、 Reliable use experience .

appendix

• The author of this article ： Huawei cloud database GaussDB(for Redis) The team
• Hangzhou / Xi'an / Shenzhen resume delivery ：[email protected]
• More product information , Welcome to the official blog ：​ ​​​bbs.huaweicloud.com/blogs/248875

​ ​ Click to follow , The first time to learn about Huawei's new cloud technology ~​​