test server

config/config.json

@@ -0,0 +1,181 @@
+{
+  "server": {
+    "name": "AI Server RTX 2060",
+    "version": "1.0",
+    "description": "Локальный AI сервер оптимизированный для RTX 2060",
+    "installation_date": "2024-12-22"
+  },
+  
+  "hardware": {
+    "gpu": {
+      "model": "RTX 2060",
+      "vram_gb": 6,
+      "cuda_cores": 1920,
+      "recommended_batch_size": 8,
+      "max_context_length": 4096,
+      "memory_fraction": 0.8
+    },
+    "system": {
+      "min_ram_gb": 16,
+      "recommended_ram_gb": 32,
+      "min_storage_gb": 50
+    }
+  },
+  
+  "ollama": {
+    "host": "0.0.0.0",
+    "port": 11434,
+    "keep_alive": "24h",
+    "num_parallel": 2,
+    "max_loaded_models": 2,
+    "flash_attention": true,
+    "gpu_memory_fraction": 0.8,
+    "environment": {
+      "CUDA_VISIBLE_DEVICES": "0",
+      "OLLAMA_GPU_OVERHEAD": "0.2",
+      "OLLAMA_NUM_PARALLEL": "2",
+      "OLLAMA_MAX_LOADED_MODELS": "2",
+      "OLLAMA_FLASH_ATTENTION": "1"
+    }
+  },
+  
+  "webui": {
+    "port": 3000,
+    "default_locale": "ru-RU",
+    "enable_signup": true,
+    "enable_login_form": true,
+    "show_admin_details": false,
+    "enable_community_sharing": false,
+    "default_models": ["llama3.2:3b", "codellama:7b"],
+    "model_filter_enabled": true
+  },
+  
+  "models": {
+    "recommended_for_rtx2060": [
+      {
+        "name": "llama3.2:1b",
+        "size_gb": 1.3,
+        "vram_usage_gb": 1.5,
+        "description": "Быстрая модель для повседневных задач",
+        "use_case": "Общие вопросы, быстрые ответы",
+        "performance": "высокая",
+        "quality": "хорошая"
+      },
+      {
+        "name": "llama3.2:3b", 
+        "size_gb": 2.0,
+        "vram_usage_gb": 2.5,
+        "description": "Оптимальная модель для RTX 2060",
+        "use_case": "Сбалансированное качество и скорость",
+        "performance": "средняя",
+        "quality": "отличная"
+      },
+      {
+        "name": "codellama:7b",
+        "size_gb": 3.8,
+        "vram_usage_gb": 4.2,
+        "description": "Специализированная модель для программирования",
+        "use_case": "Написание и анализ кода",
+        "performance": "средняя",
+        "quality": "отличная"
+      },
+      {
+        "name": "mistral:7b",
+        "size_gb": 4.1,
+        "vram_usage_gb": 4.5,
+        "description": "Универсальная модель высокого качества",
+        "use_case": "Сложные задачи, анализ текста",
+        "performance": "низкая",
+        "quality": "превосходная"
+      }
+    ],
+    
+    "not_recommended": [
+      {
+        "name": "llama3:70b",
+        "reason": "Требует более 40GB VRAM, не поместится в RTX 2060"
+      },
+      {
+        "name": "codellama:34b", 
+        "reason": "Требует более 20GB VRAM"
+      }
+    ]
+  },
+  
+  "performance_tuning": {
+    "rtx_2060_optimizations": {
+      "batch_size": 8,
+      "sequence_length": 2048,
+      "attention_type": "flash_attention",
+      "precision": "fp16",
+      "gradient_checkpointing": true,
+      "memory_efficient_attention": true
+    },
+    
+    "system_optimizations": {
+      "docker_memory_limit": "6G",
+      "docker_cpu_limit": "4.0",
+      "swap_usage": "minimal",
+      "gpu_memory_growth": true
+    }
+  },
+  
+  "monitoring": {
+    "gpu_thresholds": {
+      "temperature_warning": 80,
+      "temperature_critical": 90,
+      "memory_warning": 90,
+      "utilization_high": 95
+    },
+    
+    "system_thresholds": {
+      "cpu_warning": 80,
+      "memory_warning": 85,
+      "disk_warning": 90
+    }
+  },
+  
+  "networking": {
+    "ports": {
+      "webui": 3000,
+      "ollama_api": 11434,
+      "portainer": 9000,
+      "grafana": 3001,
+      "prometheus": 9090
+    },
+    
+    "firewall_rules": [
+      "ufw allow 3000/tcp comment 'AI Server Web UI'",
+      "ufw allow 11434/tcp comment 'Ollama API'",
+      "ufw allow 9000/tcp comment 'Portainer'"
+    ]
+  },
+  
+  "backup": {
+    "important_paths": [
+      "/opt/ai-server/config",
+      "/opt/ai-server/data", 
+      "/var/lib/docker/volumes/api-ai-server_open-webui-data"
+    ],
+    
+    "backup_schedule": "daily",
+    "retention_days": 7
+  },
+  
+  "troubleshooting": {
+    "common_issues": [
+      {
+        "issue": "GPU не обнаружен",
+        "solution": "Проверить nvidia-smi, переустановить драйверы"
+      },
+      {
+        "issue": "Нехватка VRAM",
+        "solution": "Использовать меньшую модель или уменьшить batch_size"
+      },
+      {
+        "issue": "Медленная работа",
+        "solution": "Проверить температуру GPU, очистить систему"
+      }
+    ]
+  }
+}