vLLM Startup Free Memory Below Utilization Target
Free memory on device ... on startup is less than desired GPU memory utilization
Free GPU memory at startup is below the reservation vLLM requested. Compare reported free and requested GiB per worker before KV cache sizing.
Source checked against vLLM v0.32.0. Log wording and available flags may vary by release.
First action
Check the named GPU and its resident processes with nvidia-smi. If sharing is intentional, lower --gpu-memory-utilization enough for the reservation to fit, then check that the remaining budget can still serve your workload.
Verify the change. Repeat the same launch and compare free and requested GiB for each tensor-parallel worker. If the next error concerns cache blocks, investigate the new cache budget separately.
Diagnose your vLLM log offline
Synthetic demonstration samples
Synthetic demonstration samples. Not customer telemetry.
Upload your log file. Large files are processed in your browser. Only selected evidence is submitted when you run diagnosis.
Your logs are processed entirely in your browser and are never uploaded by this tool.
Add host kernel / dmesg or nvidia-smi context (optional)
100% local browser engine (same deterministic core as the Open VSX & VS Code extension). 0 bytes of log data uploaded.
Handling a production incident? Replay it with your team.