gpu-server-management – skill Khác cho agent AI

GPU Server Management

Provision, configure, and monitor NVIDIA GPU servers for AI inference and training workloads.

When to Use This Skill

Use this skill when:

• Setting up a new GPU server for LLM inference or model training
• Installing or upgrading NVIDIA drivers and CUDA toolkit
• Configuring Docker with NVIDIA Container Toolkit for GPU workloads
• Partitioning A100/H100 GPUs with MIG for multi-tenant workloads
• Troubleshooting GPU errors, driver issues, or thermal throttling

Prerequisites

• Ubuntu 22.04 LTS (recommended) or RHEL 8/9
• NVIDIA GPU (A10G, A100, H100, RTX 4090, or L40S recommended)
• Root or sudo access
• Internet access for package downloads

Driver Installation (Ubuntu)

# Inspect the installed driver/CUDA stack and running workloads first.
# Driver removal requires an approved maintenance window and recovery plan;
# do not purge the existing GPU stack as a default installation step.

# Add NVIDIA package repository
distribution=$(. /etc/os-release; echo $ID$VERSION_ID)
curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | \
  sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg

curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | \
  sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
  sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list

sudo apt update

# Historical version example only: select a reviewed driver/CUDA pair from the
# current compatibility matrix for this exact OS/GPU before installation.
sudo apt install -y nvidia-driver-560 cuda-toolkit-12-6

# Install NVIDIA Container Toolkit (Docker GPU support)
sudo apt install -y nvidia-container-toolkit
sudo nvidia-ctk runtime configure --runtime=docker
sudo systemctl restart docker

# Verify
nvidia-smi
nvcc --version
docker run --rm --gpus all nvidia/cuda:12.6.0-base-ubuntu22.04 nvidia-smi

Post-Install Configuration

# Enable persistence mode (reduces driver initialization latency)
sudo nvidia-smi -pm 1

# Set power limit (reduce heat/noise on inference servers)
sudo nvidia-smi -pl 350          # watts; check TDP for your GPU model

# Preserve ECC by default; changes require a workload-specific integrity
# assessment and approved maintenance window.

# Enable P2P for multi-GPU NVLink training
sudo nvidia-smi topo -m          # check NVLink topology

GPU Health Monitoring

# Real-time monitoring (like htop for GPUs)
watch -n 1 nvidia-smi

# Detailed stats
nvidia-smi --query-gpu=index,name,temperature.gpu,utilization.gpu,\
utilization.memory,memory.used,memory.free,power.draw,clocks.current.graphics \
--format=csv --loop=1

# DCGM — production monitoring daemon (for clusters)
sudo apt install -y datacenter-gpu-manager
sudo systemctl start dcgm
dcgmi discovery -l                # list GPUs
dcgmi diag -r 1                  # quick health check
dcgmi diag -r 3                  # full diagnostic (takes ~20 min)

# Check GPU errors (XID errors — important for stability)
sudo dmesg | grep -i "NVRM\|nvidia\|XID"
nvidia-smi --query-gpu=ecc.errors.corrected.volatile.total \
  --format=csv,noheader

Prometheus GPU Metrics (DCGM Exporter)

# Deploy DCGM Exporter for Prometheus scraping
docker run -d \
  --name dcgm-exporter \
  --gpus all \
  --cap-add SYS_ADMIN \
  -p 9400:9400 \
  --restart unless-stopped \
  nvcr.io/nvidia/k8s/dcgm-exporter:latest

# Key metrics exposed:
# DCGM_FI_DEV_GPU_UTIL          - GPU utilization %
# DCGM_FI_DEV_MEM_COPY_UTIL     - Memory bandwidth utilization
# DCGM_FI_DEV_FB_USED           - Framebuffer memory used (MB)
# DCGM_FI_DEV_SM_CLOCK          - SM clock speed (MHz)
# DCGM_FI_DEV_GPU_TEMP          - Temperature (°C)
# DCGM_FI_DEV_POWER_USAGE       - Power draw (W)
# DCGM_FI_DEV_XID_ERRORS        - XID error count (0 = healthy)

MIG Partitioning (A100/H100)

MIG (Multi-Instance GPU) allows slicing one GPU into isolated smaller GPUs.

# Enable MIG mode (requires reboot or restart of all processes)
sudo nvidia-smi -mig 1
sudo systemctl restart nvidia-persistenced

# List available MIG profiles (A100 80GB example)
nvidia-smi mig -lgip
# 1g.10gb   — 1 slice,  10GB (max 7 instances)
# 2g.20gb   — 2 slices, 20GB (max 3 instances)
# 3g.40gb   — 3 slices, 40GB (max 2 instances)
# 7g.80gb   — full GPU, 80GB (max 1 instance)

# Example only if this exact GPU reports the profile and placement as valid:
sudo nvidia-smi mig -cgi 2g.20gb,2g.20gb,2g.20gb -C

# List created instances
nvidia-smi mig -lgi
nvidia-smi mig -lcgi

# Use in Docker
docker run --gpus '"device=MIG-GPU-xxx/0/0"' ...

# Disable MIG
sudo nvidia-smi mig -i 0 -dci
sudo nvidia-smi mig -i 0 -dgi
sudo nvidia-smi -mig 0

Kernel & OS Tuning for GPU Servers

# Increase file descriptor limits
echo '* soft nofile 1048576' | sudo tee -a /etc/security/limits.conf
echo '* hard nofile 1048576' | sudo tee -a /etc/security/limits.conf

# Disable transparent huge pages (reduces latency jitter)
echo never | sudo tee /sys/kernel/mm/transparent_hugepage/enabled
echo never | sudo tee /sys/kernel/mm/transparent_hugepage/defrag

# Persistence: prepare a dedicated reviewed systemd unit/drop-in, preserving
# existing boot configuration. Do not overwrite /etc/rc.local.
# Validate on the actual kernel/workload and retain a rollback plan.

# PCIe performance mode
sudo nvidia-smi --auto-boost-default=0
sudo nvidia-smi --auto-boost-permission=0

Multi-GPU Topology Check

# Check NVLink and PCIe topology
nvidia-smi topo -m
# Output shows interconnect type:
# NV# denotes the number of bonded NVLinks, not the NVLink generation.
# PHB = PCIe bus (slower; avoid for tensor parallel training)
# PIX = same PCIe switch (fast)

# Bandwidth test between GPUs
/usr/local/cuda/samples/bin/x86_64/linux/release/p2pBandwidthLatencyTest

Common Issues

| Issue | Cause | Fix |

|-------|-------|-----|

| nvidia-smi: command not found | Driver not installed | Follow driver installation steps above |

| Driver version mismatch | CUDA/driver incompatibility | Check compatibility matrix at developer.nvidia.com |

| GPU temperature >85°C | Poor airflow or fan failure | Check nvidia-smi -q -d TEMPERATURE; reseat cooler |

| XID 79 errors | GPU hardware error | Run dcgmi diag -r 3; may need GPU replacement |

| failed to open device in container | Container toolkit not configured | Run nvidia-ctk runtime configure --runtime=docker |

| Low PCIe bandwidth | Wrong slot or power limit | Check nvidia-smi -q | grep PCIe; use x16 slot |

Best Practices

• Always enable persistence mode (nvidia-smi -pm 1) — reduces first-request latency.
• Monitor XID errors; persistent XID 79/94 indicates hardware failure.
• For training: use NVLink-connected GPUs; for inference: PCIe is usually fine.
• Set up DCGM alerts on temperature >80°C and power draw near TDP.
• Use MIG for multi-tenant inference to provide GPU isolation between models.

Related Skills

• vllm-server (vllm-server) - LLM inference on GPUs
• llm-fine-tuning (llm-fine-tuning) - GPU training setup
• linux-hardening (linux-hardening) - Secure the host OS
• prometheus-grafana (prometheus-grafana) - Metrics dashboards

Limitations

• Infrastructure commands can disrupt services: confirm target host/scope and have backups/snapshots before mutating state.
• Docs-only import: upstream scripts and templates not bundled.

Power/clock/THP and MIG changes can interrupt active jobs; inspect exact device IDs, supported profiles and workloads first, then apply only the approved settings. nvidia-smi topo -m observes topology and does not enable P2P. Review and pin repository keys/packages/container digests before privileged installation; the historical package URLs/versions are not a current compatibility guarantee. See the [NVIDIA MIG profiles](https://docs.nvidia.com/datacenter/tesla/mig-user-guide/supported-mig-profiles.html).