I’m using “ipfs-cluster-follow cord19-datasets run --init cord19.temporal.cloud” to help with the pandemic. Started the ipfs daemon beforehand.
I get this error:
22:42:07.808 INFO cluster: ** IPFS Cluster is READY ** cluster.go:655
22:42:42.807 ERROR p2p-gorpc: failed to dial : all dials failed
- [/ip4/127.0.0.1/udp/9096/quic] CRYPTO_ERROR: peer IDs don’t match
- [/ip4/192.168.2.40/tcp/9096] dial tcp4 0.0.0.0:9096->192.168.2.40:9096: i/o timeout
- [/ip4/192.168.2.41/tcp/9096] dial tcp4 0.0.0.0:9096->192.168.2.41:9096: i/o timeout
- [/ip4/192.168.2.41/udp/9096/quic] context deadline exceeded
- [/ip4/192.168.2.40/udp/9096/quic] context deadline exceeded call.go:64
What does “all dials failed” mean?
it means there is a peer in there which is not dialable, plus a peer advertising 127.0.0.1 as well which obviously cannot be dialed to.
libp2p has made big changes and I think peers are going to be more clever about what addresses are advertised etc, and we should get this in the next release.
But other than that, I think you can safely ignore these errors (as long the trusted peer is working, your peer should sync fine).
(It is shitty that these things appear as error too. For the library that is logging it, they are effectively errors, but upstream they should be just warnings or debug. I’m thinking of ways to improve this too, particularly in the context of ipfs-cluster-follow).
I think it would be easier if the cluster follower would also print a status line at the bottom, if it’s run in a console. Like
10 pinning actions active, 500 pins in queue, delay: 5h 3m, cluster size: 10, open connections: 8, last pinset update: 3h 1m ago
or something like this
Yeah I think
ipfs-cluster-follow run could use a more user friendly output