Connect to a remote IPFS server with multiaddr /dns4/name — required for docker-compose v3+ usage

[larskluge]

Pre-check

• This is not a IPFS Cluster website content issue (file those here)
• I read the troubleshooting section of the website and it did not help
• I searched for similar issues in the repo without luck
• All my peers are running the same cluster version
• All my peers are configured using the same cluster secret

Basic information

• Version information (mark as appropiate):
  • Master
  • Release candidate for next version
  • Latest stable version
  • An older version I should not be using
• Type (mark as appropiate):
  • Bug
  • Feature request
  • Enhancement
• Operating system (mark as appropiate):
  • Linux
  • macOS
  • Windows
  • Other: which?
• Installation method (mark as appropiate):
  • Binaries from dist.ipfs.io
  • Built from sources
  • Docker
  • Snap
  • Other: which?

Description

I'm trying to get ipfs-cluster connect to a remote ipfs server instance via IPFS_API environment var & a docker compose setup v3+ and I'm running into this error:

error creating IPFS Connector component: /dns4/ipfs/tcp/5001 is not a 'thin waist' address

my docker-compose.yml:

version: '3'
services:
  ipfs:
    image: jbenet/go-ipfs
  ipfs-cluster:
    image: ipfs/ipfs-cluster:v0.4.0
    environment:
      IPFS_API: "/dns4/ipfs/tcp/5001"
      # od  -vN 32 -An -tx1 /dev/urandom | tr -d ' \n'
      CLUSTER_SECRET: ac8acc267a3556c6cde27c881e4a311db210b7242abb36ad14f77e6827a690c6

Note: I'd like to use the recent docker compose v3+ format, meaning its not allowed to specify ip4 addresses for services/containers anymore.

A proposed version with fixed IPs does not work for me due to the docker-compose v2 vs. v3 #376 and https://docs.docker.com/compose/compose-file/#ipv4_address-ipv6_address

Is there a way to run ipfs-cluster with docker-compose v3+ ? Thanks!

[lanzafame]

@larskluge Just confirming you got the following error from running docker-compose up:

ipfs-cluster_1  | Using $IPFS_API as value for ipfs_connector.ipfshttp.node_multiaddress in configuration
ipfs-cluster_1  | Warning: By default, the API and Proxy endpoints will listen on 0.0.0.0!
ipfs-cluster_1  | 00:23:51.096  INFO    service: Initializing. For verbose output run with "-l debug". Please wait... daemon.go:42
ipfs-cluster_1  | error creating IPFS Connector component: /dns4/ipfs/tcp/5001 is not a 'thin waist' address

[larskluge]

Yes, that's it. Thank you @lanzafame

[hsanjuan]

Hey @larskluge , can you test with docker image ipfs/ipfs-cluster:fix462 and let me know if things are working better for you?

(also mind you pasted your secret publicly)

[larskluge]

Beautiful, that fixes it—thanks for addressing and fixing it (well, adding this feature :) so quickly, much appreciated!

PS: And thanks for the hint, but don't worry, the posted 'secret' was modified to be more a 'public secret'.. ;-)

[hsanjuan]

@larskluge it's a bug that this feature was not there yet :)

[dill0wn]

I know this isn't quite the same issue, but I'm unable to get my cluster to recognize the $IPFS_API environment variable.

version: '3'

networks:
  ipfs_net:

services:
  ipfs:
    image: ipfs/go-ipfs
    command: 
      - daemon
      - --writable
      - --enable-pubsub-experiment
    # expose:
    #   - 4001
    #   - 5001
    #   - 8080
    ports:
      - "4001:4001"
      - "5001:5001"
      - "8080:8080"
    networks:
      - ipfs_net

  ipfs-cluster:
    image: ipfs/ipfs-cluster
    networks:
      - ipfs_net
    depends_on:
      - ipfs
    ports:
      - 9095:9095
      - 9096:9096
    environment:
      IPFS_API: "/dns4/ipfs/tcp/5001"

ipfs_1          | Changing user to ipfs
ipfs-cluster_1  | Changing user to ipfs
ipfs_1          | ipfs version 0.4.15
ipfs_1          | Found IPFS fs-repo at /data/ipfs
ipfs_1          | Initializing daemon...
ipfs-cluster_1  | ipfs-cluster-service version 0.4.0
ipfs-cluster_1  | Found IPFS cluster configuration at /data/ipfs-cluster
ipfs-cluster_1  | 20:53:36.726  INFO    service: Initializing. For verbose output run with "-l debug". Please wait... daemon.go:42
ipfs-cluster_1  | 20:53:36.729 ERROR     config: error parsing ipfs_node_multiaddress: no protocol with name ipv4 config.go:288
ipfs-cluster_1  | 20:53:36.737 ERROR     config: error parsing ipfs_node_multiaddress: no protocol with name ipv4 config.go:288
ipfs-cluster_1  | 20:53:36.752  INFO    cluster: IPFS Cluster v0.4.0-ac082929 listening on:
ipfs-cluster_1  |         /ip4/127.0.0.1/tcp/9096/ipfs/SOMEHASH
ipfs-cluster_1  |         /ip4/192.168.96.3/tcp/9096/ipfs/SOMEHASH
ipfs-cluster_1  | 
ipfs-cluster_1  |  cluster.go:95
ipfs-cluster_1  | 20:53:36.753  INFO  consensus: existing Raft state found! raft.InitPeerset will be ignored raft.go:203
ipfs-cluster_1  | 20:53:36.753  INFO    restapi: REST API (HTTP): /ip4/0.0.0.0/tcp/9094 restapi.go:386
ipfs-cluster_1  | 20:53:36.753  INFO   ipfshttp: IPFS Proxy: /ip4/127.0.0.1/tcp/9095 -> /ip4/127.0.0.1/tcp/5001 ipfshttp.go:201
ipfs-cluster_1  | 20:53:36.753  INFO    restapi: REST API (libp2p-http): ENABLED. Listening on:
ipfs-cluster_1  |         /ip4/127.0.0.1/tcp/9096/ipfs/SOMEHASH
ipfs-cluster_1  |         /ip4/192.168.96.3/tcp/9096/ipfs/SOMEHASH
ipfs-cluster_1  | 
ipfs-cluster_1  |  restapi.go:403
ipfs_1          | Swarm listening on /ip4/127.0.0.1/tcp/4001
ipfs_1          | Swarm listening on /ip4/192.168.96.2/tcp/4001
ipfs_1          | Swarm listening on /p2p-circuit/ipfs/SOMEOTHERHASH
ipfs_1          | Swarm announcing /ip4/127.0.0.1/tcp/4001
ipfs_1          | Swarm announcing /ip4/192.168.96.2/tcp/4001
ipfs_1          | API server listening on /ip4/0.0.0.0/tcp/5001
ipfs_1          | Gateway (writable) server listening on /ip4/0.0.0.0/tcp/8080
ipfs_1          | Daemon is ready
ipfs-cluster_1  | 20:53:38.754  INFO  consensus: Current Raft Leader: SOMEHASH raft.go:293
ipfs-cluster_1  | 20:53:38.755  INFO    cluster: Cluster Peers (without including ourselves): cluster.go:388
ipfs-cluster_1  | 20:53:38.755  INFO    cluster:     - No other peers cluster.go:390
ipfs-cluster_1  | 20:53:38.755  INFO    cluster: ** IPFS Cluster is READY ** cluster.go:401
ipfs-cluster_1  | 20:53:38.757 ERROR   ipfshttp: error posting to IPFS:Post http://127.0.0.1:5001/api/v0/repo/stat: dial tcp 127.0.0.1:5001: connect: connection refused ipfshttp.go:750
ipfs-cluster_1  | 20:53:38.757 ERROR   ipfshttp: Post http://127.0.0.1:5001/api/v0/repo/stat: dial tcp 127.0.0.1:5001: connect: connection refused ipfshttp.go:900
ipfs-cluster_1  | 20:53:38.757 ERROR  p2p-gorpc: Post http://127.0.0.1:5001/api/v0/repo/stat: dial tcp 127.0.0.1:5001: connect: connection refused call.go:64
ipfs-cluster_1  | 20:53:38.757 ERROR   diskinfo: Post http://127.0.0.1:5001/api/v0/repo/stat: dial tcp 127.0.0.1:5001: connect: connection refused disk.go:87

Most notably, i'm not seeing the line mentioned above:

ipfs-cluster_1  | Using $IPFS_API as value for ipfs_connector.ipfshttp.node_multiaddress in configuration

[hsanjuan]

@dill0wn this is because there is already a service.json in place:

Found IPFS cluster configuration at /data/ipfs-cluster

The node_multiaddress in it will apply. $IPFS_API is only used when no service.json exists and the entrypoint initializes a default one.