docker部署的ragflow服务迁移 数据卷迁移

背景

通过 docker compose -f docker-compose.yml -p ragflow up -d 部署了 ragflow本地服务，现在想迁移到另一台服务器上，服务可以通过github 拉取最新的 https://github.com/infiniflow/ragflow 代码，重新配置启动。
但是原服务器上添加过的数据，比如知识库，怎么迁移到新服务器，避免重复添加，重复操作呢？

一、我的配置文件

启动 ragflow 基础配置文件如下 ： docker-compose-base.yml
可以看到，各个基础服务的 volumes 数据卷名称和挂载信息

(base) root@hostname:/usr/local/soft/ai/rag/v0.19.0/ragflow/docker# vim docker-compose-base.ymlservices:es01:container_name: ragflow-es-01profiles:- elasticsearchimage: elasticsearch:${STACK_VERSION}volumes:- esdata01:/usr/share/elasticsearch/dataports:- ${ES_PORT}:9200env_file: .envenvironment:- node.name=es01- ELASTIC_PASSWORD=${ELASTIC_PASSWORD}- bootstrap.memory_lock=false- discovery.type=single-node- xpack.security.enabled=true- xpack.security.http.ssl.enabled=false- xpack.security.transport.ssl.enabled=false- cluster.routing.allocation.disk.watermark.low=5gb- cluster.routing.allocation.disk.watermark.high=3gb- cluster.routing.allocation.disk.watermark.flood_stage=2gb- TZ=${TIMEZONE}mem_limit: ${MEM_LIMIT}ulimits:memlock:soft: -1hard: -1healthcheck:test: ["CMD-SHELL", "curl http://localhost:9200"]interval: 10stimeout: 10sretries: 120networks:- ragflowrestart: on-failureopensearch01:container_name: ragflow-opensearch-01profiles:- opensearchimage: hub.icert.top/opensearchproject/opensearch:2.19.1volumes:- osdata01:/usr/share/opensearch/dataports:- ${OS_PORT}:9201env_file: .envenvironment:- node.name=opensearch01- OPENSEARCH_PASSWORD=${OPENSEARCH_PASSWORD}- OPENSEARCH_INITIAL_ADMIN_PASSWORD=${OPENSEARCH_PASSWORD}- bootstrap.memory_lock=false- discovery.type=single-node- plugins.security.disabled=false- plugins.security.ssl.http.enabled=false- plugins.security.ssl.transport.enabled=true- cluster.routing.allocation.disk.watermark.low=5gb- cluster.routing.allocation.disk.watermark.high=3gb- cluster.routing.allocation.disk.watermark.flood_stage=2gb- TZ=${TIMEZONE}- http.port=9201mem_limit: ${MEM_LIMIT}ulimits:memlock:soft: -1hard: -1healthcheck:test: ["CMD-SHELL", "curl http://localhost:9201"]interval: 10stimeout: 10sretries: 120networks:- ragflowrestart: on-failureinfinity:container_name: ragflow-infinityprofiles:- infinityimage: infiniflow/infinity:v0.6.0-dev3volumes:- infinity_data:/var/infinity- ./infinity_conf.toml:/infinity_conf.tomlcommand: ["-f", "/infinity_conf.toml"]ports:- ${INFINITY_THRIFT_PORT}:23817- ${INFINITY_HTTP_PORT}:23820- ${INFINITY_PSQL_PORT}:5432env_file: .envenvironment:- TZ=${TIMEZONE}mem_limit: ${MEM_LIMIT}ulimits:nofile:soft: 500000hard: 500000networks:- ragflowhealthcheck:test: ["CMD", "curl", "http://localhost:23820/admin/node/current"]interval: 10stimeout: 10sretries: 120restart: on-failuresandbox-executor-manager:container_name: ragflow-sandbox-executor-managerprofiles:- sandboximage: ${SANDBOX_EXECUTOR_MANAGER_IMAGE-infiniflow/sandbox-executor-manager:latest}privileged: trueports:- ${SANDBOX_EXECUTOR_MANAGER_PORT-9385}:9385env_file: .envvolumes:- /var/run/docker.sock:/var/run/docker.socknetworks:- ragflowsecurity_opt:- no-new-privileges:trueenvironment:- TZ=${TIMEZONE}- SANDBOX_EXECUTOR_MANAGER_POOL_SIZE=${SANDBOX_EXECUTOR_MANAGER_POOL_SIZE:-3}- SANDBOX_BASE_PYTHON_IMAGE=${SANDBOX_BASE_PYTHON_IMAGE:-infiniflow/sandbox-base-python:latest}- SANDBOX_BASE_NODEJS_IMAGE=${SANDBOX_BASE_NODEJS_IMAGE:-infiniflow/sandbox-base-nodejs:latest}- SANDBOX_ENABLE_SECCOMP=${SANDBOX_ENABLE_SECCOMP:-false}- SANDBOX_MAX_MEMORY=${SANDBOX_MAX_MEMORY:-256m}- SANDBOX_TIMEOUT=${SANDBOX_TIMEOUT:-10s}healthcheck:test: ["CMD", "curl", "http://localhost:9385/healthz"]interval: 10stimeout: 5sretries: 5restart: on-failuremysql:# mysql:5.7 linux/arm64 image is unavailable.image: mysql:8.0.39container_name: ragflow-mysqlenv_file: .envenvironment:- MYSQL_ROOT_PASSWORD=${MYSQL_PASSWORD}- TZ=${TIMEZONE}command:--max_connections=1000--character-set-server=utf8mb4--collation-server=utf8mb4_unicode_ci--default-authentication-plugin=mysql_native_password--tls_version="TLSv1.2,TLSv1.3"--init-file /data/application/init.sql--binlog_expire_logs_seconds=604800ports:- ${MYSQL_PORT}:3306volumes:- mysql_data:/var/lib/mysql- ./init.sql:/data/application/init.sqlnetworks:- ragflowhealthcheck:test: ["CMD", "mysqladmin" ,"ping", "-uroot", "-p${MYSQL_PASSWORD}"]interval: 10stimeout: 10sretries: 3restart: on-failureminio:image: quay.io/minio/minio:RELEASE.2025-06-13T11-33-47Zcontainer_name: ragflow-miniocommand: server --console-address ":9001" /dataports:- ${MINIO_PORT}:9000- ${MINIO_CONSOLE_PORT}:9001env_file: .envenvironment:- MINIO_ROOT_USER=${MINIO_USER}- MINIO_ROOT_PASSWORD=${MINIO_PASSWORD}- TZ=${TIMEZONE}volumes:- minio_data:/datanetworks:- ragflowrestart: on-failurehealthcheck:test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]interval: 30stimeout: 20sretries: 3redis:# swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/valkey/valkey:8image: valkey/valkey:8container_name: ragflow-rediscommand: redis-server --requirepass ${REDIS_PASSWORD} --maxmemory 128mb --maxmemory-policy allkeys-lruenv_file: .envports:- ${REDIS_PORT}:6379volumes:- redis_data:/datanetworks:- ragflowrestart: on-failurehealthcheck:test: ["CMD", "redis-cli", "-a", "${REDIS_PASSWORD}", "ping"]interval: 5stimeout: 3sretries: 3start_period: 10svolumes:esdata01:driver: localosdata01:driver: localinfinity_data:driver: localmysql_data:driver: localminio_data:driver: localredis_data:driver: localnetworks:ragflow:driver: bridg

根据上面提供的 docker-compose-base.yml 配置文件内容，下面是 RAGFlow 各个组件在本地的数据存储路径说明：

---

✅ 数据存储路径一览（基于你的配置文件）

组件	容器内路径	本地路径（默认）	用途说明
MySQL	/var/lib/mysql	mysql_data（默认在 Docker 管理的 volume 中）	存储用户、文档元数据等结构化信息
MinIO	/data	minio_data（Docker volume）	存储上传的原始文档、处理后的文本等
Elasticsearch (es01)	/usr/share/elasticsearch/data	esdata01（Docker volume）	存储关键词搜索索引数据
OpenSearch (opensearch01)	/usr/share/opensearch/data	osdata01（Docker volume）	可选搜索引擎，用于检索
Infinity（向量数据库）	/var/infinity	infinity_data（Docker volume）	存储向量索引
Redis	/data	redis_data（Docker volume）	缓存数据、任务状态等
Sandbox Executor Manager	无持久化数据	无	用于运行沙箱任务，不持久化

---

关于这些 volumes 的说明

你在 docker-compose-base.yml 文件末尾定义了如下 volume：

volumes:esdata01:driver: localosdata01:driver: localinfinity_data:driver: localmysql_data:driver: localminio_data:driver: localredis_data:driver: local

这些 volumes 是 Docker 管理的命名卷（named volumes），默认情况下它们会存储在 Docker 的默认数据目录中：

• Linux 系统下通常为：/var/lib/docker/volumes/<volume_name>/_data

例如：

• MySQL 数据实际路径为：/var/lib/docker/volumes/mysql_data/_data
• MinIO 数据实际路径为：/var/lib/docker/volumes/minio_data/_data

---

📁 如何查看这些卷在本地的具体位置？

可能用到的docker 命令如下。

📦 总结建议

目标	建议
查看数据存储位置	使用 docker volume inspect <volume_name>
备份数据	备份对应路径，如 /var/lib/docker/volumes/mysql_data/_data
自定义路径	修改 docker-compose-base.yml 中的 volumes 配置
清理数据	删除对应 volume，如 docker volume rm mysql_data

---

所以，我们想查看某个 volume 的实际路径，可以有两种方法：

• find / -name ‘esdata01’
• docker volume inspect xxxx

本次我们使用如下命令，esdata01 是从上面配置文件中获悉：

(base) root@hostname:/home/ltkj# docker volume inspect esdata01
[]
Error response from daemon: get esdata01: no such volume(base) root@hostname:/home/ltkj# docker volume inspect ragflow_esdata01

输出示例：

[{"CreatedAt": "2025-06-13T10:52:28Z","Driver": "local","Labels": {"com.docker.compose.config-hash": "3bcef595c5f477c290ccaa07cbf05671d287fef95c1fa4b67fad841e66481794","com.docker.compose.project": "ragflow","com.docker.compose.version": "2.34.0","com.docker.compose.volume": "esdata01"},"Mountpoint": "/var/lib/docker/volumes/ragflow_esdata01/_data","Name": "ragflow_esdata01","Options": null,"Scope": "local"}
]

这样就能知道 esdata01，minio，MySQL 等基础服务的数据存在哪里了。

---

🛠️ 如果想自定义这些路径

可以在 volumes 配置中指定本地路径，例如：

volumes:mysql_data:driver: localdriver_opts:type: noneo: binddevice: /opt/ragflow/mysql_data

这样 MySQL 数据就会存在 /opt/ragflow/mysql_data。

---

TODO 未完 待续 。。。