[BUG] Downloads Fail on hosts with IPv4/IPv6 dual stack

[jjschwarz]

Is there an existing issue for this?

• [X] I have searched the existing issues

Current Behavior

When you attempt to download the initial resources the system seems to try and connect to localhost:5001 , however on hosts with dual stack (IPv4/IPv6) networks the connection gives an error connecting to port 5001 on ::1 (IPv6 localhost) because the API server is not binding to the IPv6 host.

Expected Behavior

API server binds to port 5001 on both IPv4 and IPv6

Steps To Reproduce

1. Deploy in dual stack environment
2. Attempt to download resources
3. Receive error due to failure to connect on port 5001 to host ::1

Environment

- OS: K8S v1.27.6

CPU architecture

x86-64

Docker creation

name: emulatorjs-deployment
  namespace: emulatorjs
  labels:
    app: emulatorjs
spec:
  selector:
    matchLabels:
      app: emulatorjs
  replicas: 1
  revisionHistoryLimit: 0 # Default to 10 if not specified
  strategy:
    type: Recreate
  template:
    metadata:
      labels:
        app: emulatorjs
    spec:
      containers:
      - name: emulatorjs
        image: linuxserver/emulatorjs
        ports:
          - name: web
            containerPort: 80
            protocol: TCP
          - name: config
            containerPort: 3000
            protocol: TCP
        env:
        - name: PUID
          value: "1000"
        - name: PGID
          value: "1000"
        - name: TZ
          value: "America/New_York"
        volumeMounts:
        - name: local-conf
          mountPath: /etc/sysctl.d/local.conf
          subPath: local.conf
        - name: emulatorjs-data
          mountPath: /data
        - name: emulatorjs-config
          mountPath: /config
        livenessProbe:
          httpGet:
            path: /
            port: 80
            scheme: HTTP
          initialDelaySeconds: 120
          periodSeconds: 90
      securityContext:
        fsGroup: 1000
      volumes:
      - name: emulatorjs-data
        persistentVolumeClaim:
          claimName: emulatorjs-data
      - name: emulatorjs-config
        persistentVolumeClaim:
          claimName: emulatorjs-config
      - name: local-conf
        configMap:
          name: local-conf

Container logs

Gateway (readonly) server listening on /ip4/127.0.0.1/tcp/8080
Daemon is ready
[ls.io-init] done.
node:internal/deps/undici/undici:11372
    Error.captureStackTrace(err, this);
          ^

TypeError: fetch failed
    at Object.fetch (node:internal/deps/undici/undici:11372:11)
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5) {
  cause: Error: connect ECONNREFUSED ::1:5001
      at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1555:16) {
    errno: -111,
    code: 'ECONNREFUSED',
    syscall: 'connect',
    address: '::1',
    port: 5001
  }
}

Node.js v18.18.2

[jjschwarz]

Removing ::1 temporarily from /etc/hosts allows you to bypass the issue, however since Kubernetes controls that file it must be done manually each time you deploy the software.

[j0nnymoe]

Can this be replicated with docker? We've not experienced the same issue in testing. K8s isn't a supported platform by us : https://docs.linuxserver.io/misc/support-policy/#reasonable-endeavours-support

[jjschwarz]

I have not replicated it in docker, however I found a fix for the issue. If the API line in /data/.ipfs/config is updated to read: "API": ["/ip4/127.0.0.1/tcp/5001","/ip6/::1/tcp/5001"], then the issue is resolved.

At this point I have made that change manually which is persistent. Obviously it would be more ideal for that to be in the upstream if the change does not impact your Docker build.

[RedPine404]

I can confirm that this is an issue with docker running a ULA IPv6 address. As a workaround I just disabled IPv6 in my container with sysctls: - net.ipv6.conf.all.disable_ipv6=1

[jjschwarz]

Disabling IPv6 is not an option in our environment because it would impact other containers which use it.  I am using the patch I suggested for now.On Dec 2, 2023, at 2:17 PM, RedPine404 @.***> wrote: I can confirm that this is an issue with docker running a ULA IPv6 address. As a workaround I just disabled IPv6 in my container with sysctls: - net.ipv6.conf.all.disable_ipv6=1

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: @.***>

[LinuxServer-CI]

This issue has been automatically marked as stale because it has not had recent activity. This might be due to missing feedback from OP. It will be closed if no further activity occurs. Thank you for your contributions.

[jjschwarz]

Would you consider patching the API line in /data/.ipfs/config is updated to read: "API": ["/ip4/127.0.0.1/tcp/5001","/ip6/::1/tcp/5001"] to prevent this issue?

[LinuxServer-CI]

This issue has been automatically marked as stale because it has not had recent activity. This might be due to missing feedback from OP. It will be closed if no further activity occurs. Thank you for your contributions.