Skip to content

Zombie GPU Process / Orphaned Job

Zombie processes hold GPU memory after the main job exits. Denpex auto-detects and kills them.

Quick answer

Zombie processes hold GPU memory after the main job exits.

Reliability#zombie-process#gpu-memory#process-cleanup#orphaned#reliability

What this failure is

Zombie GPU Process / Orphaned Job is a Reliability failure seen during ML training runs. Zombie processes hold GPU memory after the main job exits. Denpex auto-detects and kills them. Common tags: Zombie Process, Gpu Memory, Process Cleanup, Orphaned.

Live diagnosis, no signup

Is this what broke your run? Paste your log.

You're reading about Zombie GPU Process / Orphaned Job. Paste your own crash log or traceback below and get the real root cause for YOUR run, not this generic entry. No account, no card. Logs are masked at ingress and never saved to account history.

training_logs.txt
No log to hand? Try one:

3 free diagnoses/day

Want 14 days on the Scale plan?

Request an evaluation code. A verified workplace organization activates up to 50 diagnoses a day, alerts, history, and follow-up questions. No credit card or automatic subscription.

Evaluate one incident

Why it happens (the mechanism)

Training killed without SIGTERM propagated to child processes. NCCL communicator not finalized holding GPU memory. Python multiprocessing spawn leaves child processes. SSH session disconnected mid-training, leaving orphaned CUDA context. Taken together, these mechanisms explain why the failure is reproducible, why it tends to surface on specific workloads or scales, and why generic mitigation attempts often fall short without addressing the underlying cause.

What you'll observe

  • nvidia-smi shows GPU memory used but process is invisible
  • New jobs fail with CUDA OOM after killing previous job
  • SSH sessions show defunct Python processes

Common symptoms and what they mean

SymptomWhy it happens
nvidia-smi shows CUDA processes that won't dieTraining killed without SIGTERM propagated to child processes
Training fails: address already in useNCCL communicator not finalized holding GPU memory
GPU memory 0 bytes allocated by invisible processesPython multiprocessing spawn leaves child processes
Stale processes left after SLURM job cancellationSSH session disconnected mid-training, leaving orphaned CUDA context

Which systems are affected

  • SLURM job cancellation without cleanup
  • Kubernetes pod evictions
  • Ray actor crashes
  • Ctrl+C training without trap handlers

How to confirm this is the problem

Use this checklist to test the hypothesis against a small reproduction. No single line proves the root cause, so preserve the preceding events and compare one variable at a time.

  • ✓Reproduce the failure from a clean checkpoint/seed: the symptom must appear without warm-up state from a previous run.
  • ✓Verified signal present: nvidia-smi shows CUDA processes that won't die
  • ✓Verified signal present: Training fails: address already in use
  • ✓Verified signal present: GPU memory 0 bytes allocated by invisible processes
  • ✓Verified signal present: Stale processes left after SLURM job cancellation
  • ✓A targeted fix from the "How to fix it" section eliminates or substantially reduces the symptom within one validation pass.

The fix and the prevention pattern

The root cause is on this page and stays free. A free account adds the exact remediation steps, saved history, and the fix on every entry in the encyclopedia.

Sign up free. Unlock the full analysis

No credit card. Daily allowance follows verified trust tier. Instant access.

Diagnose this failure in VS Code

Select the traceback or open the failed terminal, then run Denpex locally to see the initiating rank, collateral failures, exact fix, and verification command without uploading the log.

Install the free VS Code extension

Related failures to investigate next

Bytedance Sequence Length Straggler often co-occurs with this failure because both share underlying assumptions about reliability; if you fix this entry but still see drift, validate the related primary cause next. bytedance sequence length straggler →
Training Checkpoint Corruption On Write often co-occurs with this failure because both share underlying assumptions about reliability; if you fix this entry but still see drift, validate the related neighboring cause next. training checkpoint corruption on write →
Wandb Tensorboard Failure often co-occurs with this failure because both share underlying assumptions about reliability; if you fix this entry but still see drift, validate the related neighboring cause next. wandb tensorboard failure →

Root cause

  • Training killed without SIGTERM propagated to child processes
  • NCCL communicator not finalized holding GPU memory
  • Python multiprocessing spawn leaves child processes
  • SSH session disconnected mid-training, leaving orphaned CUDA context

The fix and how to prevent it

Evaluate Denpex on your own logs

Request a Scale evaluation code. A verified workplace organization activates 14 days with up to 50 diagnoses a day. Every account keeps its current diagnosis allowance and gets a verification path. No card or automatic subscription.

We send a single-use code tied to that address. Static provider and TLD rules do not reject valid addresses. Account trust determines the benefit after signup.

References

Don't just read the fix, diagnose your run

The encyclopedia tells you what went wrong. Denpex tells you what went wrong in YOUR training run. With your logs, your config, and your stack.

Diagnose your logs freeBrowse all failures

Related Reliability errors

Sequence Length Imbalance Causing Distributed Training Stragglers

Reliability · high

Silent Data Corruption from GPU Hardware Faults Causing Loss Spikes and Model Divergence

Reliability · critical

NIXL Firmware Page Registration Fan-Out Triggers Host OOM Kills on HGX H200 and B200

Reliability · critical

MTTF Scaling Inversely with GPU Count in Large ML Research Clusters

Reliability · high