Short deion explaining the high-level reason for the new issue. <h1 dir="aut

Great, got a fix! WIll release soon. See <a class="issue-link js-issue-link" data-erro

Thanks <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-u

Workflow stuck with parallelizable node about hamilton HOT 2 OPEN

legout commented on July 17, 2024

Workflow stuck with parallelizable node

from hamilton.

Comments (2)

elijahbenizzy commented on July 17, 2024 1

Great, got a fix! WIll release soon. See #949.

from hamilton.

elijahbenizzy commented on July 17, 2024

Thanks @legout! You've identified what appears to be a bug. I'm digging in -- first clue (+ context) is the following:

The executor first breaks it up into "tasks"
It then executes tasks, walking through the graph
Usually this is: (1) node tasks for each of the non-parallel block tasks, then multi-node tasks for everything inside a "parallelizable block"
In this case, it's not creating a task for the first one (so it never gets started). This is getting added to the middle task. It thus never reaches the end, so it just spins.

Looking into it more, hoping I can find a quick solution

Debugging information:

Tasks:

[TaskSpec(base_id='expand-numbers',
          spawning_task_base_id=None,
          nodes=[<numbers {'module': 'temporary_module_b26db111_76fe_4d83_8d44_4fc58b9542f6'}>],
          purpose=<NodeGroupPurpose.EXPAND_UNORDERED: 'expand_unordered'>,
          available_nodes={'numbers'},
          outputs_to_compute=['numbers'],
          overrides={},
          adapter=<hamilton.lifecycle.base.LifecycleAdapterSet object at 0x2a27fb810>,
          base_dependencies=['block-numbers']),
 TaskSpec(base_id='block-numbers',
          spawning_task_base_id='expand-numbers',
          nodes=[<add1 {'module': 'temporary_module_b26db111_76fe_4d83_8d44_4fc58b9542f6'}>,
                 <random_int {'module': 'temporary_module_b26db111_76fe_4d83_8d44_4fc58b9542f6'}>,
                 <add2 {'module': 'temporary_module_b26db111_76fe_4d83_8d44_4fc58b9542f6'}>],
          purpose=<NodeGroupPurpose.EXECUTE_BLOCK: 'execute_block'>,
          available_nodes={'add1', 'random_int', 'add2'},
          outputs_to_compute=['random_int', 'add2'],
          overrides={},
          adapter=<hamilton.lifecycle.base.LifecycleAdapterSet object at 0x2a27fb810>,
          base_dependencies=['expand-numbers']),
 TaskSpec(base_id='collect-numbers',
          spawning_task_base_id='expand-numbers',
          nodes=[<collect_numbers {'module': 'temporary_module_b26db111_76fe_4d83_8d44_4fc58b9542f6'}>],
          purpose=<NodeGroupPurpose.GATHER: 'gather'>,
          available_nodes={'collect_numbers'},
          outputs_to_compute=['collect_numbers'],
          overrides={},
          adapter=<hamilton.lifecycle.base.LifecycleAdapterSet object at 0x2a27fb810>,
          base_dependencies=['block-numbers'])]