IPFS seemed like a worthy enterprise, so I got involved. A while ago, I ipfs routing provided a small directory and tried to retrieve it from another node, it was taking forever, and so, being an obsessive type, I went down the rabbit hole…Ended up with clusters (routinely go on follower type pinning rounds for ipfs websites) a personal ruby pinning service, various ipfs/kubo instances on different machines, it’s quite fascinating. I refuse to “cheat”, as it were, with this, small directory, recursive CID and its children (is it?) I want the machine in question to genuinely search “everywhere”, in its fashion and retrieve the goods, one way or another. The 5 clusters on the machine assigned the task are on attempt 1450 (while also successfully retrieving, pinning, unpinning, other CIDs elsewhere, so it’s not like they “don’t know how”) You’ve got to laugh…
Perhaps I should add that the array of 4 clusters on the machine from whence the CID originated, and where the directory is located, did manage to retrieve the CIDs, in a relatively short time, but I don’t regard that as much of an achievement, somehow (though I experienced an unreasonable amount of elation, and some relief) and I’m perplexed by the others’ inability to find…It’s not that the cluster arrays, or the kubo nodes, or the pinning service, are unaffiliated to each other.