How Do I Do If a Compute Instance Fails 30 Seconds After It Is Started?

Question

A HetuEngine compute instance enters the faulty state about 30 seconds after it is started.

Answer

When starting a compute instance, HetuEngine sends a command to Yarn to start the corresponding application. If HetuEngine does not receive a response from Yarn within 30 seconds, HetuEngine ends the request due to timeout.

If the response message for Yarn to start the application cannot be received within 30 seconds due to machine performance or network environment problems, you can prolong the timeout period.

1. Log in to FusionInsight Manager.
2. Choose Cluster > Services > HetuEngine and click Configurations then All Configurations.
3. Search for the application.customized.properties parameter, add the custom parameter yarn.application.start.timeout, set the timeout interval as required (enter only digits without the unit second), and save the configuration.
4. Click the Instance tab, select all HSBroker instances, click More, and select Restart Instance to restart the HSBroker instances as prompted.