1

I am facing a hard time trying to deploy a multihost Hyperledger fabric network using a cjdns ipv6 tunnelling. I was able to successfully deploy a multihost network using docker swarm, but for my point of view docker swarm is not very practical in production mode. Does anyone know what kind of modifications should I do to Docker compose files, in order to make a successful deployment of a multihost consortium in Hyperledger fabric using ipv6 cjdns tunnelling?

George Misiakoulis
  • 11
  • 3

1 Answers1

1

Unless you actually plan to use Docker Swarm, the Docker Compose files are going to be essentially useless in terms of deploying a network across multiple hosts.

You may be able to break apart the Compose file into separate files if you wish to stand up different Fabric nodes on different hosts and/or simply use them as a reference for settings to pass to the various containers when starting them on different hosts.

But the Docker Compose examples provided by Fabric all rely on the fact that Compose provides some convenience mechanisms for communicating across a Docker network.

If you choose to set things up across different hosts without using something like Compose or Kubernetes, you will need to create the propose DNS or hosts file entries for the FQDNs you choose to use for the various Fabric nodes.

You might want to have a look at https://hyperledger-fabric.readthedocs.io/en/release-2.0/deployment_guide_overview.html# as it provides pretty good guidance on the steps required to get a set of Fabric nodes up and running.

Gari Singh
  • 11,418
  • 2
  • 18
  • 41
  • Thank you very much. The FQDNs is to register the peers like the peerN.orgN.example.com? – George Misiakoulis May 07 '20 at 13:51
  • `The FQDNs is to register the peers like the peerN.orgN.example.com` - exactly – Gari Singh May 07 '20 at 14:12
  • I am wondering, having a cjdns tunelling means that two or more hosts are in the same encrypted network. I manage to ping a running orderer which located in the Host A from the Host B (Both Hosts are in the same cjdns tunnel), but when i fire up the peer 0 which it is located in Host A, i cant communicate fto the peer 0 rom Host B. Is this because of FQDNs or i need something else? – George Misiakoulis May 07 '20 at 14:17