Maya process memory stays high after File New Scene: leak or allocator retention?

Question

During long Maya sessions, resident memory climbs steadily and is not returned to the OS after closing a heavy scene with File > New Scene. The only eventual failure condition is a platform-dependent out-of-memory crash, so there is no error string to anchor diagnosis. The unresolved question is how to classify this growth before it reaches that point.

Two documented contributors complicate the picture: the undo queue, whose length is configurable via the undoInfo command and which grows without bound under default settings, and the evaluation/dependency-graph caches retained in Parallel and Evaluation Manager modes. On top of that, Maya's memory allocators and loaded plug-ins are known to retain arenas even after scene data is destroyed, so Task Manager or top numbers alone cannot distinguish a genuine leak from fragmentation and retention. Third-party plug-ins and Python API reference cycles (including legacy MScriptUtil misuse) add further suspects.

Assume a recent Maya version on Windows or Linux, with behavior potentially differing by evaluation mode and OS allocator.

Is there a supported way to attribute Maya's resident memory to the undo queue, evaluation caches, or plug-in allocations, short of external heap profiling?
Does bounding the undo queue via undoInfo and flushing caches reliably cap growth, or is a full restart the only guaranteed reclamation?
What minimal reproduction best separates a true plug-in leak from allocator retention?

Tasadduq Burney · Answer

Answer the exact question first
1. Is there a supported way to attribute Maya’s resident memory to the undo queue, evaluation caches, or plug‑in allocations, short of external heap profiling?

Undo queue: Query the current size with undoInfo -display. You can set a hard limit with undoInfo -maxUndoLevels 50 or undoInfo -unlimited false to force a cap.
  Evaluation caches: In the UI, open Window → Settings/Preferences → General → Evaluation and click Clear All Caches. From the command line you can run evalInfo -clearCaches to flush DG/Parallel caches.
  Plug‑in allocations: Maya does not expose per‑plugin memory usage directly. The closest workaround is to launch Maya with the -noexec flag to skip user scripts and plug‑ins, then compare RSS before and after a New Scene. If the difference disappears, the culprit is likely a plug‑in.

These commands give you a rough attribution, but they do not provide a fine‑grained breakdown like a heap profiler would.

Does bounding the undo queue and flushing caches reliably cap growth?
Setting a low maxUndoLevels and clearing caches will stop the rate of RSS growth, but Maya’s internal allocators (C++ arenas, Python memory manager) often keep freed blocks for reuse. As a result, the OS may still report a high RSS even after the undo data is gone. Only a full process restart guarantees that all retained arenas are returned to the OS.

Minimal reproduction to separate a true plug‑in leak from allocator retention

Launch Maya with maya -noexec (no user scripts, plug‑ins, or startup files). Record RSS with ps -o rss= -p $$ (Linux) or Get-Process -Id $PID | Select-Object WorkingSet64 (PowerShell).
  Open a heavy test scene, perform a series of operations, then choose File → New Scene.
  undoInfo -maxUndoLevels 20
undoInfo -unlimited false
evalInfo -clearCaches

Measure RSS again. The difference should be close to the size of the undo data plus caches.
  Repeat the same steps in a clean Maya instance that loads all your plug‑ins (no -noexec). If RSS remains high after the same sequence, the extra memory is likely held by a plug‑in or a Python reference cycle.

By comparing the two runs, you isolate allocator retention (present in both) from plug‑in leakage (present only in the full‑load run).

Missing diagnostic detail that could change the recommendation
Do you have any third‑party plug‑ins or custom Python scripts loaded during the session? If so, they might be the source of the persistent memory, and the recommendation would shift toward disabling them or investigating their own memory handling.

Maya process memory stays high after File > New Scene: leak or allocator retention?

1 answer

Answer the exact question first

Does bounding the undo queue and flushing caches reliably cap growth?

Minimal reproduction to separate a true plug‑in leak from allocator retention

Missing diagnostic detail that could change the recommendation

Your perspective belongs here.

0 question comments