What Should I Do If Jobs Fail to Be Submitted to YARN After JobGateway Starts for the First Time

Question

When JobGateway starts for the first time, jobs fail to be submitted to YARN, and no launcher-job ID is generated.

On the YARN page, no launcher-job queue is displayed.

Check whether the launcher-job information is displayed on the FusionInsight Manager page.

Check all JobServer logs and locate "Create and modify launcher-job tenant failed."

Answer

1. Log in to the active management node as user omm and run the following command to obtain the tenant ID of the launcher-Job tenant:

   curl -ki -L -u Username:Password "https://$(hostname -i):9022/mrsmanager/api/v2/clusters/1/tenants/resource_allocation/super_scheduler?resource_pool_name=default&parent_tenant_name=root&limit=10&offset=0&filter="

2. Run the following command to delete launcher-job. The tenant ID of launcher-job is the one obtained in 1.

   curl -k -i --basic -u admin:{Password} -X DELETE 'https://{omsIp}:28443/web/api/v2/clusters/1/tenants/{Tenant ID of launcher-job}'

3. Log in to FusionInsight Manager, choose Cluster > Services > JobGateway, click Instances, and perform a rolling restart on all JobServer instances.
4. After the restart is complete, log in to the native YARN page and check whether the launcher-job queue is generated and whether the job can be submitted successfully.