libp2p/hydra-booster

Hydra nodes forward non bitswap provider records to Kubo nodes

Jorropo opened this issue · 4 comments

Jorropo commented 
$ ipfs dht findprovs bafybeibml5uieyxa5tufngvg7fgwbkwvlsuntwbxgtskoqynbt7wlchmfm | parallel -j32 ipfs id {}
{
	"ID": "12D3KooWSH5uLrYe7XSFpmnQj1NCsoiGeKSRCV7T5xijpX2Po2aT",
	"PublicKey": "CAESIPSTR/o7SFKWn1Xuj6Cic9hjUbHjxPcXHu/eXSoiL5Xo",
	"Addresses": [
		"/ip4/145.40.90.155/tcp/4001/p2p/12D3KooWSH5uLrYe7XSFpmnQj1NCsoiGeKSRCV7T5xijpX2Po2aT",
		"/ip4/145.40.90.155/tcp/4002/ws/p2p/12D3KooWSH5uLrYe7XSFpmnQj1NCsoiGeKSRCV7T5xijpX2Po2aT",
		"/ip6/2604:1380:45e1:2700::d/tcp/4001/p2p/12D3KooWSH5uLrYe7XSFpmnQj1NCsoiGeKSRCV7T5xijpX2Po2aT",
		"/ip6/2604:1380:45e1:2700::d/tcp/4002/ws/p2p/12D3KooWSH5uLrYe7XSFpmnQj1NCsoiGeKSRCV7T5xijpX2Po2aT"
	],
	"AgentVersion": "go-ipfs/0.13.0-dev/aa5336f",
	"ProtocolVersion": "ipfs/0.1.0",
	"Protocols": [
		"/floodsub/1.0.0",
		"/ipfs/bitswap",
		"/ipfs/bitswap/1.0.0",
		"/ipfs/bitswap/1.1.0",
		"/ipfs/bitswap/1.2.0",
		"/ipfs/id/1.0.0",
		"/ipfs/id/push/1.0.0",
		"/ipfs/kad/1.0.0",
		"/ipfs/lan/kad/1.0.0",
		"/ipfs/ping/1.0.0",
		"/libp2p/autonat/1.0.0",
		"/libp2p/circuit/relay/0.1.0",
		"/libp2p/circuit/relay/0.2.0/hop",
		"/libp2p/circuit/relay/0.2.0/stop",
		"/libp2p/fetch/0.0.1",
		"/meshsub/1.0.0",
		"/meshsub/1.1.0",
		"/p2p/id/delta/1.0.0",
		"/x/"
	]
}

Error: no good addresses
{
	"ID": "12D3KooWLDf6KCzeMv16qPRaJsTLKJ5fR523h65iaYSRNfrQy7eU",
	"PublicKey": "CAESIJqMXQzM/L0Gprj3KmrLCFy5wshdwL2+XOtjC1Bs+uAl",
	"Addresses": [
		"/ip4/141.138.64.21/tcp/11337/p2p/12D3KooWLDf6KCzeMv16qPRaJsTLKJ5fR523h65iaYSRNfrQy7eU"
	],
	"AgentVersion": "lotus-1.15.3-rc2+mainnet+git.ddbcf1fc6",
	"ProtocolVersion": "ipfs/0.1.0",
	"Protocols": [
		"/fil/datatransfer/1.2.0",
		"/fil/kad/testnetnet/kad/1.0.0",
		"/fil/retrieval/qry/0.0.1",
		"/fil/retrieval/qry/1.0.0",
		"/fil/storage/ask/1.0.1",
		"/fil/storage/ask/1.1.0",
		"/fil/storage/mk/1.0.1",
		"/fil/storage/mk/1.1.0",
		"/fil/storage/status/1.0.1",
		"/fil/storage/status/1.1.0",
		"/floodsub/1.0.0",
		"/ipfs/graphsync/1.0.0",
		"/ipfs/graphsync/2.0.0",
		"/ipfs/id/1.0.0",
		"/ipfs/id/push/1.0.0",
		"/ipfs/ping/1.0.0",
		"/legs/head//indexer/ingest/mainnet/0.0.1",
		"/libp2p/autonat/1.0.0",
		"/meshsub/1.0.0",
		"/meshsub/1.1.0",
		"/p2p/id/delta/1.0.0"
	]
}

{
	"ID": "12D3KooWRBy97UB99e3J6hiPesre1MZeuNQvfan4gBziswrRJsNK",
	"PublicKey": "CAESIORoCy+MjSEJDmqjJ/G7NCq459kjjx41gxpU1qj1yREk",
	"Addresses": [
		"/ip4/127.0.0.1/tcp/4001/p2p/12D3KooWRBy97UB99e3J6hiPesre1MZeuNQvfan4gBziswrRJsNK",
		"/ip4/127.0.0.1/udp/4001/quic/p2p/12D3KooWRBy97UB99e3J6hiPesre1MZeuNQvfan4gBziswrRJsNK",
		"/ip4/192.168.178.63/tcp/4001/p2p/12D3KooWRBy97UB99e3J6hiPesre1MZeuNQvfan4gBziswrRJsNK",
		"/ip4/192.168.178.63/udp/4001/quic/p2p/12D3KooWRBy97UB99e3J6hiPesre1MZeuNQvfan4gBziswrRJsNK",
		"/ip4/89.14.204.222/tcp/38789/p2p/12D3KooWRBy97UB99e3J6hiPesre1MZeuNQvfan4gBziswrRJsNK",
		"/ip4/89.14.204.222/udp/38789/quic/p2p/12D3KooWRBy97UB99e3J6hiPesre1MZeuNQvfan4gBziswrRJsNK",
		"/ip6/::1/tcp/4001/p2p/12D3KooWRBy97UB99e3J6hiPesre1MZeuNQvfan4gBziswrRJsNK",
		"/ip6/::1/udp/4001/quic/p2p/12D3KooWRBy97UB99e3J6hiPesre1MZeuNQvfan4gBziswrRJsNK"
	],
	"AgentVersion": "go-ipfs/0.13.0/c9d51bbe0",
	"ProtocolVersion": "ipfs/0.1.0",
	"Protocols": [
		"/ipfs/bitswap",
		"/ipfs/bitswap/1.0.0",
		"/ipfs/bitswap/1.1.0",
		"/ipfs/bitswap/1.2.0",
		"/ipfs/id/1.0.0",
		"/ipfs/id/push/1.0.0",
		"/ipfs/kad/1.0.0",
		"/ipfs/lan/kad/1.0.0",
		"/ipfs/ping/1.0.0",
		"/libp2p/autonat/1.0.0",
		"/libp2p/circuit/relay/0.1.0",
		"/libp2p/circuit/relay/0.2.0/hop",
		"/libp2p/circuit/relay/0.2.0/stop",
		"/libp2p/dcutr",
		"/p2p/id/delta/1.0.0",
		"/x/"
	]
}

{
	"ID": "12D3KooWChhhfGdB9GJy1GbhghAAKCUR99oCymMEVS4eUcEy67nt",
	"PublicKey": "CAESICreFMj5TRvD7BcmouPoF8l/m4+OY9mxOnRBLM5o+G3l",
	"Addresses": [
		"/ip4/127.0.0.1/tcp/4001/p2p/12D3KooWChhhfGdB9GJy1GbhghAAKCUR99oCymMEVS4eUcEy67nt",
		"/ip4/127.0.0.1/udp/4001/quic/p2p/12D3KooWChhhfGdB9GJy1GbhghAAKCUR99oCymMEVS4eUcEy67nt",
		"/ip4/172.105.8.48/tcp/4001/p2p/12D3KooWChhhfGdB9GJy1GbhghAAKCUR99oCymMEVS4eUcEy67nt",
		"/ip4/172.105.8.48/udp/4001/quic/p2p/12D3KooWChhhfGdB9GJy1GbhghAAKCUR99oCymMEVS4eUcEy67nt",
		"/ip6/2600:3c04::f03c:92ff:feed:761d/tcp/4001/p2p/12D3KooWChhhfGdB9GJy1GbhghAAKCUR99oCymMEVS4eUcEy67nt",
		"/ip6/2600:3c04::f03c:92ff:feed:761d/udp/4001/quic/p2p/12D3KooWChhhfGdB9GJy1GbhghAAKCUR99oCymMEVS4eUcEy67nt",
		"/ip6/::1/tcp/4001/p2p/12D3KooWChhhfGdB9GJy1GbhghAAKCUR99oCymMEVS4eUcEy67nt",
		"/ip6/::1/udp/4001/quic/p2p/12D3KooWChhhfGdB9GJy1GbhghAAKCUR99oCymMEVS4eUcEy67nt"
	],
	"AgentVersion": "go-ipfs/0.11.0/",
	"ProtocolVersion": "ipfs/0.1.0",
	"Protocols": [
		"/floodsub/1.0.0",
		"/ipfs/bitswap",
		"/ipfs/bitswap/1.0.0",
		"/ipfs/bitswap/1.1.0",
		"/ipfs/bitswap/1.2.0",
		"/ipfs/id/1.0.0",
		"/ipfs/id/push/1.0.0",
		"/ipfs/kad/1.0.0",
		"/ipfs/lan/kad/1.0.0",
		"/ipfs/ping/1.0.0",
		"/libp2p/autonat/1.0.0",
		"/libp2p/circuit/relay/0.1.0",
		"/libp2p/circuit/relay/0.2.0/hop",
		"/libp2p/circuit/relay/0.2.0/stop",
		"/libp2p/fetch/0.0.1",
		"/meshsub/1.0.0",
		"/meshsub/1.1.0",
		"/p2p/id/delta/1.0.0",
		"/x/"
	]
}

{
	"ID": "12D3KooWAqjDEE3PTqp37YnjznYcTX5j2hu9oXwRa1kpuNAHWPfb",
	"PublicKey": "CAESIA81D5oEiNnPmed9IaxduBTan+ohcE+N0zSxMUtqDu4+",
	"Addresses": [
		"/ip4/145.40.80.135/tcp/10211/p2p/12D3KooWAqjDEE3PTqp37YnjznYcTX5j2hu9oXwRa1kpuNAHWPfb"
	],
	"AgentVersion": "lotus-1.15.2+mainnet+git.518dc962e",
	"ProtocolVersion": "ipfs/0.1.0",
	"Protocols": [
		"/fil/datatransfer/1.2.0",
		"/fil/kad/testnetnet/kad/1.0.0",
		"/fil/retrieval/qry/0.0.1",
		"/fil/retrieval/qry/1.0.0",
		"/fil/storage/ask/1.0.1",
		"/fil/storage/ask/1.1.0",
		"/fil/storage/mk/1.0.1",
		"/fil/storage/mk/1.1.0",
		"/fil/storage/status/1.0.1",
		"/fil/storage/status/1.1.0",
		"/floodsub/1.0.0",
		"/ipfs/graphsync/1.0.0",
		"/ipfs/graphsync/2.0.0",
		"/ipfs/id/1.0.0",
		"/ipfs/id/push/1.0.0",
		"/ipfs/ping/1.0.0",
		"/legs/head//indexer/ingest/mainnet/0.0.1",
		"/libp2p/autonat/1.0.0",
		"/meshsub/1.0.0",
		"/meshsub/1.1.0",
		"/p2p/id/delta/1.0.0"
	]
}

{
	"ID": "12D3KooWQUpdfNYyJNoJZyk6kqfrLePashLL15d6ti8ETdDjYy6d",
	"PublicKey": "CAESINndnFtqdhRyb31C8kbILp1RJmW5/sWY8sgi+QFUWFUe",
	"Addresses": [
		"/ip4/113.29.246.221/tcp/18375/p2p/12D3KooWQUpdfNYyJNoJZyk6kqfrLePashLL15d6ti8ETdDjYy6d"
	],
	"AgentVersion": "lotus-1.15.2+mainnet+git.518dc962e",
	"ProtocolVersion": "ipfs/0.1.0",
	"Protocols": [
		"/fil/datatransfer/1.2.0",
		"/fil/kad/testnetnet/kad/1.0.0",
		"/fil/retrieval/qry/0.0.1",
		"/fil/retrieval/qry/1.0.0",
		"/fil/storage/ask/1.0.1",
		"/fil/storage/ask/1.1.0",
		"/fil/storage/mk/1.0.1",
		"/fil/storage/mk/1.1.0",
		"/fil/storage/status/1.0.1",
		"/fil/storage/status/1.1.0",
		"/floodsub/1.0.0",
		"/ipfs/graphsync/1.0.0",
		"/ipfs/graphsync/2.0.0",
		"/ipfs/id/1.0.0",
		"/ipfs/id/push/1.0.0",
		"/ipfs/ping/1.0.0",
		"/legs/head//indexer/ingest/mainnet/0.0.1",
		"/libp2p/autonat/1.0.0",
		"/meshsub/1.0.0",
		"/meshsub/1.1.0",
		"/p2p/id/delta/1.0.0"
	]
}
guseggert commented

What is actionable here? What is the expected behavior?

Jorropo commented

@guseggert the indexer protocol allows to mark which protocols peers support.
We should ensure that peers supports bitswap and only return thoses. (as a parameter in reframe ? filter code done in hydras ? other ? idk).

ribasushi commented

@guseggert as @Jorropo pointed out - currently kubo and other IPFS implementations expect the DHT to exclusively contain bitswap-dialable peers. This is starting to become a problem in more and more places as the Filecoin SP community is pushing more and more data into the indexer. Effectively the hydra is now actively poisoning the DHT, filling it up with records that will not respond to requests.

We need to fix this ASAP, it's a structural network concern at this stage.

cc @BigLep @aschmahmann @mikeal

  • 👍4
BigLep commented

2022-08-12 triage conversation: this isn't as bad as it seems because:

  • We dial peers in parallel so as long as there is a bitswap peer in the list of providers.
  • STI with its reframe server implementation is actually filtering out the non-bitswap providers. (Per below, we do think we should change this though.)
  • We observe the behavior reported above right now because Hydras are still querying storetheindex using the pre-reframe API which doesn't filter and we don't have any filtering logic on the client/Hydra size.

What needs to happen:

  1. Hydras should be deployed using the latest go-delegated-routing. This is covered in #162
  2. Ensure we're filtering out non-bitswap providers in the Hydra/client-side
  3. storetheindex should update its reframe endpoint to send all records (graphsync and bitswap).

There's a potential parallel track where reframe findProviders will take hints of what provider records the client wants (e.g., bitswap only). That's a future optimization that we don't have a backlog item right now but also aren't planning to prioritize.