Ten routing strategies for distributed task scheduling platform XXL job

Type of routing policy ：

1. first ： When this policy is selected , The first machine that will select the registered address of the actuator will execute , If the first machine fails , Then the scheduling task fails .

2. the last one ： When this policy is selected , The last machine with the registered address of the actuator will be selected to execute , If the last machine breaks down , Then the scheduling task fails .

3. polling ： When this policy is selected , It will poll and assign tasks according to the registered address of the actuator , If one of the machines fails , Scheduling task failed , The task will not shift .

4. Random : When this policy is selected , The task will be randomly assigned according to the registered address of the actuator , If one of the machines fails , Scheduling task failed , The task will not shift .

5. Uniformity HASH: When this policy is selected , Each task according to Hash The algorithm selects a certain machine . If that machine breaks down , Scheduling task failed , The task will not shift .

6. The least used ： When this policy is selected , Will give priority to the machine with the lowest frequency of use , If one of the machines fails , Scheduling task failed , The task will not shift .

7. Most recently unused ： When this policy is selected , Priority will be given to the machine that has not been used for the longest time , If one of the machines fails , Scheduling task failed , The task will not shift .

8. Fail over ： When this policy is selected , The heartbeat is detected in sequence , If one of the machines fails , Then it will be transferred to the next actuator , If the heartbeat test is successful , Will be selected as the target actuator and initiate scheduling .

9. Busy transfer ： When this policy is selected , Carry out idle detection in sequence , If one of the machines fails , Then it will be transferred to the next actuator , If idle detection is successful , Will be selected as the target actuator and initiate scheduling .

10. Broadcast by film ： When this policy is selected , Broadcast triggers all machines in the cluster to perform one task , At the same time, the system automatically transfers the partition parameters ; According to the partition parameters, the partition task can be developed . If one of the machines fails , Then the execution of the actuator fails , Will not affect other actuators .