Skip to content

NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled

Traffic spread across multiple NICs arrived out of order. RoCE is highly sensitive to reordering, go-back-N retransmission means a reordered flow collapses throughput rather than degrading gracefully, and can stall the pipeline entirely. This entry explains how to confirm the cause, apply the fix, and separate it from adjacent nccl-tuning failures.

Quick answer

NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled means Traffic spread across multiple NICs arrived out of order. RoCE is highly sensitive to reordering, go-back-N retransmission means a reordered flow collapses throughput rather than degrading gracefully, and can stall the pipeline entirely. Preserve the first preceding error, then run the targeted control below.

Communication#nccl-tuning#nccl#cross#nic#roce#packet

What this failure is

The literal signature is "NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled". It is a communication failure associated with NCCL algorithm, buffer, and multi-NIC tuning. The line identifies the failing operation or subsystem, while the surrounding evidence decides whether it is the initiating fault or a downstream symptom.

Live diagnosis, no signup

Is this what broke your run? Paste your log.

You're reading about NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled. 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 of full access?

Request a free trial code for unlimited diagnoses, alerts, history, and follow-up questions. No credit card.

Request 14-day trial

Why it happens (the mechanism)

Traffic spread across multiple NICs arrived out of order. RoCE is highly sensitive to reordering, go-back-N retransmission means a reordered flow collapses throughput rather than degrading gracefully, and can stall the pipeline entirely. The failure becomes visible at this call site because the operation first requires the missing resource, valid state, healthy peer, or correct result. Earlier log lines and a known-good control carry more causal value than the final wrapper exception.

What you'll observe

  • The workload stops or loses forward progress after emitting "NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled".
  • A retry on the same configuration reproduces the failure because the causal state has not changed.
  • The outer framework exception can hide the rank, node, allocation, or dependency that failed first.
  • Increasing timeouts or reducing workload size can suppress the symptom without correcting the cause.

Common symptoms and what they mean

SymptomWhy it happens
NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalledTraffic spread across multiple NICs arrived out of order. RoCE is highly sensitive to reordering, go-back-N retransmission means a reordered flow collapses throughput rather than degrading gracefully, and can stall the pipeline entirely.
The same operation fails at a consistent stage of NCCL algorithm, buffer, and multi-NIC tuning.The decisive evidence is the first log line that precedes "NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled" and differs from a healthy run.
The first related warning appears before the final exception and names the causal subsystem.A nearby failure remains a competing hypothesis until a control separates configuration, capacity, transport, and hardware causes.
A known-good control changes one variable and either reproduces or clears the failure.Traffic spread across multiple NICs arrived out of order. RoCE is highly sensitive to reordering, go-back-N retransmission means a reordered flow collapses throughput rather than degrading gracefully, and can stall the pipeline entirely.

Which systems are affected

  • NCCL algorithm, buffer, and multi-NIC tuning
  • production-shaped multi-accelerator workloads
  • containerized and bare-metal deployments of the same stack

How to confirm this is the problem

Apply the following checklist to a small reproduction: each box below is a positive signal that you are looking at this exact failure rather than a sibling in the same taxonomy.

  • Find the first occurrence of "NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled" and preserve at least 100 lines before it.
  • Identify which rank, node, device, or process emitted the first related warning.
  • reordering across NICs usually means ECMP hashing is spreading one QP's packets over multiple paths. Confirm the fabric hashes on the full 5-tuple and that both NICs are in the same layer-2 domain with matching MTU. Check port counters for out-of-sequence on the receiving HCA.
  • Repeat the same input after the targeted change and require the signature to disappear.
  • Resume from latest checkpoint only after the control passes.

Example training logs (fingerprint)

training.log (synthetic fingerprint)
NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled

Timestamps and exact values vary across runs, but the pattern. An info-level start, an early WARN, an ERROR carrying the symptom. Is the actual fingerprint you should alert on. The Denpex platform flags this combination automatically.

The fix and the prevention pattern

The root cause is on this page and stays free. A free account adds the exact remediation steps, keeps your diagnoses instead of discarding them, and unlocks the fix on every entry in the encyclopedia.

Sign up free. Unlock the full analysis

No credit card · 3 free diagnoses · Instant access

Why the recommended fix works

disable multi-NIC striping for a single flow, export NCCL_CROSS_NIC=0, and re-measure. If throughput recovers, reordering was the cause. This changes the condition in the causal diagnosis instead of hiding the outer exception. The repeated control proves ownership before recovery from latest checkpoint.

Code examples

snippet
# Preserve evidence before restarting
NCCL_DEBUG=INFO NCCL_DEBUG_SUBSYS=INIT,NET,COLL torchrun train.py
ibstat
ethtool -S <interface>

# Find the exact signature in the complete log
rg -n -F -- "NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled" <log-file>

Adapt the snippet to your framework. The same pattern holds for PyTorch Lightning, Hugging Face Trainer, DeepSpeed, Megatron-LM, and vLLM training wrappers. Where the wrapper exposes a config flag (for examplelr_scheduler_type in Trainer), prefer the flag over the imperative API to keep the schedule declarative and reproducible.

Best practices by model family

Model / StackRecommendationNotes
First responsePreserve the first failureKeep the context before "NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled" so aggregation does not erase causality.
ConfirmationChange one variableUse a known-good node, rank, input, or configuration as the control.
RecoveryResume from latest checkpointResume only after the literal signature no longer appears in the same control.

With the fix vs without the fix

DimensionWith the fixWithout the fix
EvidenceFirst preceding error and one controlled comparisonOnly the final aggregated exception
Fixdisable multi-NIC striping for a single flow, export NCCL_CROSS_NIC=0, and re-measure. If throughput recovers, reordering was the cause.Retrying the unchanged workload
Exit criterion"NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled" is absent in the repeated controlThe job happened to run once

Real engineering notes

Treat "NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled" as a search key and an investigation checkpoint, not as proof of every cause associated with the phrase. The high-value evidence is what changed immediately before it and whether the failure follows the workload, node, or configuration.

Visual fingerprint

Decision path for NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled
literal error captured
        |
        v
find first preceding failure
        |
        v
run one known-good control
        |
        +-- follows workload --> inspect input or configuration
        +-- follows node ------> inspect hardware or platform
        +-- disappears --------> validate the targeted fix
The control separates workload, configuration, and node ownership before recovery from latest checkpoint.

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

NCCL errors in context

NCCL is where a distributed job reports failure, which is not the same as where it failed. The hub lists every common NCCL error next to what it actually indicates, and the environment variables that tell them apart.

Compare every nccl error side by side

Root cause

  • Traffic spread across multiple NICs arrived out of order. RoCE is highly sensitive to reordering, go-back-N retransmission means a reordered flow collapses throughput rather than degrading gracefully, and can stall the pipeline entirely.
  • The decisive evidence is the first log line that precedes "NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled" and differs from a healthy run.
  • A nearby failure remains a competing hypothesis until a control separates configuration, capacity, transport, and hardware causes.

The fix and how to prevent it

Unlock the full remediation runbook

14 days on the Scale plan, up to 50 diagnoses a day. Step-by-step remediation, the RMA evidence payload, and multi-node correlation on your own logs. No card, and it does not roll into a subscription.

We send a single-use code to that address. Company addresses only, the free diagnoses above stay open to everyone, and keeping trials to work email is how we keep them open.

Frequently asked questions

Twelve targeted questions that engineers and on-call staff most commonly ask about this failure.

What does "NCCL multi-NIC packet reordering detected RoCE packets arrived out of sequence pipeline stalled" mean?
Traffic spread across multiple NICs arrived out of order. RoCE is highly sensitive to reordering, go-back-N retransmission means a reordered flow collapses throughput rather than degrading gracefully, and can stall the pipeline entirely.
Is this line always the root cause?
No. It can be the direct failure or the point where an earlier failure becomes visible. The first preceding error and a controlled comparison decide which.
What should I collect before restarting?
Collect complete log context, the emitting rank or node, component versions, resolved configuration, and the diagnostic output shown above.
What is the fastest confirmation?
reordering across NICs usually means ECMP hashing is spreading one QP's packets over multiple paths. Confirm the fabric hashes on the full 5-tuple and that both NICs are in the same layer-2 domain with matching MTU. Check port counters for out-of-sequence on the receiving HCA.
How do I prevent it from recurring?
keep a given queue pair pinned to one NIC and one path; use NCCL_IB_HCA to bind ranks to the NIC on their own NUMA node.

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.