Call for testing: ipfs-cluster


After finding a workround for an issue with the home in go projects, we are ready to start testing the ipfs-cluster.

If you remember from my various posts about ipfs, this is basically a distributed file system that takes a new approach on addressing. Instead of addressing by name, things are addressed by a hash of their contents. The thing is that if you put a document in the network, it will only be replicated if it’s of interest to other people. If at some point in time no client with a copy of your document is online, it will not be accessible.

So, ipfs-cluster lets you control a group of ipfs nodes to give high availability to your content. You define a number of replicas for the document, and the cluster tries to keep that count when one of its nodes goes down.

To help us testing, you can set up the cluster with:

$ sudo snap install ipfs
$ ipfs init
$ ipfs daemon &
$ sudo snap install ipfs-cluster --edge
$ ipfs-cluster.service init
$ ipfs-cluster.service
$ ipfs-cluster.ctl id

If you find any errors with those commands, please reply here. If all looks correctly, you can start experimenting with the cluster commands. Here is a guide with detailed explanations:

Here is the snapcraft.yaml:

pura vida

1 Like

Hey! I just wanted to chime in saying I’m currently using the IPFS and IPFS-Cluster snap on two different Ubuntu Server, connected to an IPFS / IPFS-Cluster built from source on Ubuntu Desktop. Everything works fine!