Using juicefs mount on windows, using etcd database, the connection always fails. Auto-disconnect!

[masx200]

What happened:

Using juicefs mount on windows, using etcd database, the connection always fails. Auto-disconnect!

The error might be caused by the fact that etcd is running in single-machine mode and thus unable to obtain the cluster address.

2025-04-23 15:10:35.099: {"level":"warn","ts":"2025-04-23T23:10:35.09941+0800","logger":"etcd-client","caller":"[email protected]/retry_interceptor.go:62","msg":"retrying of unary invoker failed","target":"etcd-endpoints://0xc0015aa540/10.200.200.3:2379","attempt":0,"error":"rpc error: code = DeadlineExceeded desc = latest balancer error: last connection error: connection error: desc = \"transport: Error while dialing: dial tcp 0.0.0.0:2379: connectex: No connection could be made because the target machine actively refused it.\""}
2025-04-23 15:10:35.099: {"level":"info","ts":"2025-04-23T23:10:35.09941+0800","logger":"etcd-client","caller":"[email protected]/client.go:210","msg":"Auto sync endpoints failed.","error":"context deadline exceeded"}
2025-04-23 15:11:40.124: {"level":"warn","ts":"2025-04-23T23:11:40.124576+0800","logger":"etcd-client","caller":"[email protected]/retry_interceptor.go:62","msg":"retrying of unary invoker failed","target":"etcd-endpoints://0xc0015aa540/10.200.200.3:2379","attempt":0,"error":"rpc error: code = DeadlineExceeded desc = latest balancer error: last connection error: connection error: desc = \"transport: Error while dialing: dial tcp 0.0.0.0:2379: connectex: No connection could be made because the target machine actively refused it.\""}
2025-04-23 15:11:40.124: {"level":"info","ts":"2025-04-23T23:11:40.124576+0800","logger":"etcd-client","caller":"[email protected]/client.go:210","msg":"Auto sync endpoints failed.","error":"context deadline exceeded"}

What you expected to happen:

How to reproduce it (as minimally and precisely as possible):

juicefs.exe mount --max-uploads=50 --writeback --cache-size 204800 --bucket=http://<address>:19000/juicefs --attr-cache 600s    --entry-cache 600s        --dir-entry-cache 600s    "etcd://root:<password>@<address>:2379/juicefs" z:

Anything else we need to know?

Environment:

• JuiceFS version (use juicefs --version) or Hadoop Java SDK version:
  1.2.3+2025-01-22.4f2aba8
• Cloud provider or hardware configuration running JuiceFS:
• OS (e.g cat /etc/os-release):
  windows 11 26100.3775
• Kernel (e.g. uname -a):
• Object storage (cloud provider and region, or self maintained):
  minio
• Metadata engine info (version, cloud provider managed or self maintained):
  etcd standalone

ETCD_LISTEN_CLIENT_URLS=http://0.0.0.0:2379

ETCD_ADVERTISE_CLIENT_URLS=http://0.0.0.0:2379

• Network connectivity (JuiceFS to metadata engine, JuiceFS to object storage):
  not on same computer
• Others:

[chenjie4255]

Single-node mode works fine on my side.

And from the log you provided:

dial tcp 0.0.0.0:2379: connectex: No connection could be made because the target machine actively refused it.\""},

Maybe you should change ETCD_ADVERTISE_CLIENT_URLS=http://0.0.0.0:2379 to an real address(not 0.0.0.0)

[masx200]

Since my etcd service is running inside a Docker container and has undergone NAT (Network Address Translation), the IP address within the container is mapped to the host machine's address through a port. As a result, remote computers cannot access the IP address within the container.

[chenjie4255]

Since my etcd service is running inside a Docker container and has undergone NAT (Network Address Translation), the IP address within the container is mapped to the host machine's address through a port. As a result, remote computers cannot access the IP address within the container.

Could you try ping 0.0.0.0 on your Windows command prompt or PowerShell? On my side, it doesn't work either. And on a Linux system, it seems that the command of pinging 0.0.0.0 will finally result in the pinging of 127.0.0.1.

[masx200]

version: '3.8'
services:
  etcd:
    stdin_open: true

    tty: true
    privileged: true
    restart: always
    image: bitnami/etcd:3.5
    container_name: etcd-standalone
    environment:
      - ETCD_ROOT_PASSWORD=<password>  # 自定义 root 密码
      - ETCD_NAME=etcd-single                  # 节点名称
      - ETCD_LISTEN_CLIENT_URLS=http://0.0.0.0:2379
      - ETCD_ADVERTISE_CLIENT_URLS=http://0.0.0.0:2379
      - ETCD_INITIAL_CLUSTER_TOKEN=<password>
      - ETCD_INITIAL_CLUSTER_STATE=new
    ports:
      - "2379:2379"
      - "2380:2380"
    volumes:
      - /media/debian-msx/542c5887-1dbd-48d8-95a1-125a7b3d01c5/etcd-data:/bitnami/etcd/data
    networks:
      - etcd-net

networks:
  etcd-net:
    driver: bridge
    external: true
    name: "mount-verysync_default"
    
    enable_ipv6: true  # 启用IPv6支持 [[5]][[9]]
    ipam:
      config:
        - subnet: 172.20.0.0/24       # IPv4子网 [[2]]
          gateway: 172.20.0.1
        - subnet: fc01:db8:21:172::/64     # IPv6子网（示例，需符合实际网络规划）
          gateway: fc01:db8:21:172::1

juicefs format --storage minio     --bucket http://10.200.200.3:19000/juicefs     --access-key <password>     --secret-key <password>    --trash-days=30 "etcd://root:<password>@100.122.154.114:2379/juicefs" juicefs