Skip to content

cuMemImportFromShareableHandle Fails (CUDA error 101)

P2P fabric-handle import between containers on a multi-node NVLink (MNNVL) GPU fails with CUDA 'invalid device ordinal' due to a CUDA driver bug, not NCCL.

Quick answer

P2P fabric-handle import between containers on a multi-node NVLink (MNNVL) GPU fails with CUDA 'invalid device ordinal' due to a CUDA driver bug, not NCCL.

Hardware#cuda#fabric#mnnvl#gh200#error-101#driver-bug

What this failure is

cuMemImportFromShareableHandle Fails (CUDA error 101) is a Hardware failure seen during ML training runs. P2P fabric-handle import between containers on a multi-node NVLink (MNNVL) GPU fails with CUDA 'invalid device ordinal' due to a CUDA driver bug, not NCCL. Common tags: Cuda, Fabric, Mnnvl, Gh200.

Live diagnosis, no signup

Is this what broke your run? Paste your log.

You're reading about cuMemImportFromShareableHandle Fails (CUDA error 101). 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)

CUDA UMD bug in driver 570.00. Fabric handle export/import across container/PID namespaces breaks when both see the GPU as device 0. Not an NCCL bug. 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

  • P2P send/recv between containers on the same GH200 node fails
  • Both containers see the GPU as device 0
  • Error is in fabric handle import

Common symptoms and what they mean

SymptomWhy it happens
transport/p2p.cc NCCL WARN Cuda failure 101 'invalid device ordinal'CUDA UMD bug in driver 570.00
cuMemImportFromShareableHandle with CU_MEM_HANDLE_TYPE_FABRIC failsFabric handle export/import across container/PID namespaces breaks when both see the GPU as device 0
P2P ncclSend/ncclRecv across containers failsNot an NCCL bug

Which systems are affected

  • GH200 / MNNVL nodes
  • Containerized multi-process jobs sharing a GPU across PID namespaces
  • CUDA driver 570.00

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: transport/p2p.cc NCCL WARN Cuda failure 101 'invalid device ordinal'
  • ✓Verified signal present: cuMemImportFromShareableHandle with CU_MEM_HANDLE_TYPE_FABRIC fails
  • ✓Verified signal present: P2P ncclSend/ncclRecv across containers fails
  • ✓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

CUDA errors in context

CUDA reports errors asynchronously, so the traceback usually points at whatever line synchronised next rather than the one at fault. The hub covers every common CUDA error and how to make it report honestly.

Compare every cuda error side by side

Related failures to investigate next

Gpu Memory Clock Throttle often co-occurs with this failure because both share underlying assumptions about hardware; if you fix this entry but still see drift, validate the related primary cause next. gpu memory clock throttle →
Cuda Illegal Instruction often co-occurs with this failure because both share underlying assumptions about hardware; if you fix this entry but still see drift, validate the related neighboring cause next. cuda illegal instruction →
Gsp Rpc Timeout Xid119 often co-occurs with this failure because both share underlying assumptions about hardware; if you fix this entry but still see drift, validate the related neighboring cause next. gsp rpc timeout xid119 →

Root cause

  • CUDA UMD bug in driver 570.00
  • Fabric handle export/import across container/PID namespaces breaks when both see the GPU as device 0
  • Not an NCCL bug

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 Hardware errors

GPU Memory Clock Throttle

Hardware · medium

NVLink Error

Hardware · high

GPU Thermal Throttling

Hardware · high

GPU Power Cap Reached

Hardware · medium