Why Does an Application Move Back to the Original Queue After the ResourceManager Restarts?
Question
After an application is moved from one queue to another, the application moves back to the original queue after the ResourceManager restarts.
Answer
This behavior is due to a usage limitation of the ResourceManager. When an application is moved to a different queue during running, the ResourceManager does not store the updated queue information in its state store. As a result, if the ResourceManager restarts, the application is reassigned to its original queue.
Assume that a user submits a MapReduce task to the leaf queue test11. If the user deletes the leaf queue test11 while the task is still running, the system automatically reassigns the submission queue to lost_and_found. This queue serves as a fallback for tasks that are no longer associated with any valid queue. As a result, the task is suspended. To resume the task, the user moves the task to the leaf queue test21. Although the task is running in the leaf queue test21, after the ResourceManager restarts, the submission queue appears as lost_and_found instead of test21.
If a task is still in progress when the ResourceManager restarts, the ResourceManager's state store retains only the original queue to which the task was initially submitted. To resolve this issue, reassign the application to the desired queue after the ResourceManager restarts. This ensures that the updated queue information is correctly written to the ResourceManager's state store.
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbot