I share some content between my friends.
To be -the content- more fast for download, I am trying to upload it in one more computer behing different internet aDSL connection.
So, I run on second computer the following commands
ipfs swarm connect /ip4/10.0.0.2/tcp/4001/p2p/12D3Koo…
ipfs get /ipns/k51qzi5…
But always the download stucks.
Last time it stucks on 1.6GB of 28GB
(Both computers are in the same VPN layer 2, so the ip 10.0.0.2)
What diagnostic will be useful to understand the problem?
I’m pretty sure there are some config settings you can tweak in your instance to turn on additional logging. I’ve never done it or else I’d try to help more.
Does the download resume from the last point, or does it totally start over each time?
I found the problem.
It was a renamed file on my disk, pointed by filestore.
So, it is not error proof. It does not continue to next blocks. Just stalls.
Anyway it was my fault.
I’m working on fixing this : Make IPFS reflink aware, dedup file storage between IPFS and user downloaded files · Issue #8201 · ipfs/go-ipfs · GitHub
This will allows you to move and edits your filestored files and still keep the data intact in IPFS. (this uses reflinking of the linux kernel, instead of storing a plain file and offset, this will create a reflink, so any attempt to modify the filestored file would actually create a copy of it).
But I live on ext4 filesystem so reflink is not available.
Anyway, it is a good hack!