locked
Run multiple tasks concurrently on a node using Python RRS feed

  • Question

  • Hello,

    I have seen that you can allow a node to run multiple tasks concurrently in .NET by setting the CloudPool.MaxTasksPerComputeNode property.

    How can this be set in Python?

    Also, is there a time limit for a job? I have read that a job cannot run for more than 7 days straight. Is that true?

    Thanks

    Wednesday, November 22, 2017 8:00 AM

Answers

  • Hi John,

    The docs for the python SDK are here: http://azure-sdk-for-python.readthedocs.io/en/latest/ref/azure.batch.html

    Specifically for the question you asked you can refer to this: http://azure-sdk-for-python.readthedocs.io/en/latest/ref/azure.batch.models.html#azure.batch.models.PoolAddParameter -- specifically the max_tasks_per_node field. Set that when adding the pool to configure the max number of tasks per node.

    There is no max runtime for a job, but there is a max runtime for an individual task. The time between an individual task addition and that tasks completion must be <= 7 days. After completion a task can exist indefinitely, the limit only exists for tasks in "Active" or "Running" state.

    Hope that helps,

    -Matt

    Wednesday, November 22, 2017 8:40 AM