Fault tolerant mechanism record

When we use certain functions , Sometimes it may not be successful , So what strategies will we adopt to deal with this abnormal situation , Let's illustrate with an example

As shown in the figure above ,app2 Call... Over the network app1 Interface ,app1 It is a cluster composed of three machines

failfast Quick failure

When an error occurs in the system , Immediately feed back errors

eg1： When app2 Carry error Token visit app1 Interface ,app1 Immediately feed back error information

failsafe Fail safe

When an error occurs in the system , The failure will not cause impact or minimize the impact

eg1： When app2 Call the interface to query app1 Some interfaces , Such as query IPhone 10 The color of the terminal phone ,app1 Something's wrong inside , No error information is fed back at this time , It's going back to “ Unknown ” Information . about app2 Come on , It will not perceive the exception of interface call , We will continue to deal with it as usual

failover Fail over

When an error occurs in the system , Let's try other systems

eg1： When app2 Call the interface to query app1 Some interfaces , First call the interface app1（192.168.1.100）, If it fails , Let's try again app1（192.168.1.101）、app1（192.168.1.102）; The system handles a high availability state , Stop any of these apps , The system can also operate normally

failback Automatic fault recovery

When an error occurs in the system , The application self fixes the fault ; How to understand this self-healing process

• Network jitter problem , Sometimes the network may fail , Cause some exceptions in the program call , After the program receives the exception , You can retry calling the interface later , How long to wait and how many times to try , How to control is a means of self-healing
• Traffic degradation , When the network traffic increases sharply , Our system resources are insufficient , We can directly feed back to users and try later

eg1： When app2 Call the interface to query app1 An error occurred on some interfaces , Remaining interval 30s Try again , Retry no more than 3 Time

The above are several ways to deal with the error of the fault-tolerant mechanism

dubbo Fault tolerant solutions

Failover Cluster
Fail auto switch , When there is failure , Try another server again . Usually used for read operations , But retrying brings longer delays . It can be done by retries=“2” To set the number of retries ( Not including the first time ).

Failfast Cluster
Fast failure , Make only one call , Failure reports an immediate error . Usually used for non idempotent write operations , For example, adding records .

Failsafe Cluster
Failure safety , When something unusual happens , Direct to ignore . Usually used to write audit logs and other operations .

Failback Cluster
Automatic recovery from failure , Failed requests are logged in the background , Timing to resend . Typically used for message notification operations .

Forking Cluster
Call multiple servers in parallel , Return as soon as one succeeds . It is usually used for reading operations with high real-time requirements , But more service resources need to be wasted . It can be done by forks=“2” To set the maximum number of parallels .