filecoin-project / lotus

Reference implementation of the Filecoin protocol, written in Go
https://lotus.filecoin.io/
Other
2.83k stars 1.25k forks source link

Placeholder for market issues #1538

Closed arajasek closed 3 years ago

arajasek commented 4 years ago

Collect market-related issues here, ideally spin them into their own issues before putting time into investigating.

arajasek commented 4 years ago

1) Follow up with partners regarding offline deal workflow / commP issues

arajasek commented 4 years ago

2)

Reported here: https://filecoinproject.slack.com/archives/CPFTWMY7N/p1585945303257400

laser Apr 3rd at 4:21 PM I am running a 10 daemon dev network on EC2 + one daemon running on my laptop. All daemons and storage miners have been peered successfully. When I propose a storage deal using my local daemon to the miners running in EC2, the lotus client list-deals command shows the deals progressing from StorageDealUnknown to StorageDealValidating and finally to StorageDealSealing. 9 of the 10 storage deals which I have proposed are stuck in the StorageDealSealing state, and one has made it to StorageDealActive. Using the lotus-storage-miner deals list command (on each of the storage miners which have accepted my deal proposals), I see no problems. When I use lotus-storage-miner sectors status 0, I see that each sector has Status: Proving, and that its Deals output includes the storage deal proposed by my client. Unfortunately, the client never sees 9 of the 10 deals progress from StorageDealSealing to StorageDealActive. Is this a known bug? Note: I'm using CloudFormation to provision this dev network, so this issue can be reproduced somewhat trivially using a few CloudFormation templates and Bash scripts. 9 replies

laser 10 days ago Here's what the client sees: 13:14 $ lotus client list-deals DealCid Provider State PieceCID Size Price Duration Message bafyreia7lld42jej4pyreig7piqxmkbfcidhzpyh5icj5sxd7hnpkl36oq t01016 StorageDealSealing bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna 2032 1000000 5 bafyreiamz5turke2vhgyfj3bhj4x3k2w4lp2hwieewhiaktqh3awbzeggm t01000 StorageDealSealing bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna 2032 1000000 5 bafyreidbydlapszgvkm2cfjjgfbzimypacleaq5jsu4cxidp6lw7j524cy t01012 StorageDealSealing bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna 2032 1000000 5 bafyreidsf67fuik6ilxcnjhuhzsxu7ywcbn67wmcfyynw7y4zz4zcrw72m t01018 StorageDealSealing bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna 2032 1000000 5 bafyreiemhx4sfhowr4bmmet5e3ie3qhdnzgosppxncfbyhsi7b2fj6s3fm t01004 StorageDealActive bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna 2032 1000000 5 bafyreiftfjgptfm6laeq4nlkgdj6xnmc4a3mukx6vpfr6qoyrba4sqwrkq t01002 StorageDealSealing bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna 2032 1000000 5 bafyreigg477zj4iyck2mhbxql2245obbfotetdjtuxniqsr7brhd2s7qd4 t01014 StorageDealSealing bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna 2032 1000000 5 bafyreigydvy2nbs57335qphsgjiqnnptniheadyfucvcpjrzzcr6rfjgpu t01006 StorageDealSealing bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna 2032 1000000 5 bafyreihgy6sbx5t4pmn6mhn4glxb2e3i25cllwmw2fhxs4w5xi24elesfa t01010 StorageDealSealing bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna 2032 1000000 5 bafyreihtzaoa5m6snmihil7a4vfx6dt7u4gxumfnma6g25ramdeixt7f5i t01008 StorageDealSealing bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna 2032 1000000 5

laser 10 days ago Here's what a miner sees: root@ip-10-0-10-11:/# lotus-storage-miner info Miner: t01006 Sector Size: 2 KiB Power: 2 KiB / 24 KiB (8.3333%) Committed: 2 KiB Proving: 2 KiB PoSt Submissions: Fallback: Epoch 407 (in 134 blocks, ~13m 24s) Deadline: Epoch 427 (in 374 blocks, ~37m 24s) Consecutive Failures: 0 Sectors: map[Proving:1 Total:1] root@ip-10-0-10-11:/# lotus-storage-miner deals list 2020-04-03T20:15:42.648Z WARN rpc jsonrpc/client.go:385 unmarshaling failed {"message": "[{\"Proposal\":{\"PieceCID\":{\"/\":\"bafk4chza7qwcfdo2s3t5rzh3fcvylobgoujy42bjixs3mk4dt3n7ilpgbqna\"},\"PieceSize\":2048,\"Client\":\"t3uqqtf7acuyw7masdmf2s6ja7p365f3qavvsmxpt5yuuvugj7a5gghtqijd32heshk6a2bojuwegfidccmava\",\"Provider\":\"t01006\",\"StartEpoch\":10063,\"EndEpoch\":10068,\"StoragePricePerEpoch\":\"1000000\",\"ProviderCollateral\":\"2032\",\"ClientCollateral\":\"0\"},\"ClientSignature\":{\"Type\":2,\"Data\":\"ocTiHQQtjt6vsO3x1wG/ggTL2WzZwR+zTZ0Jn96vdIs+5dKaByLb3EUTkJwVKqy1Aa8DDedz1VJGg6bOgWWxOs8EWKnZA8xPFHNVR2Dik5bmM5diwFuTzzkG7RGoCIbD\"},\"ProposalCid\":{\"/\":\"bafyreigydvy2nbs57335qphsgjiqnnptniheadyfucvcpjrzzcr6rfjgpu\"},\"Miner\":\"\",\"Client\":\"12D3KooWQHcEGrqGTNHBiYFrknM6KGKCT8kgp5Jedr9Ce8oRjy11\",\"State\":5,\"PiecePath\":\"fstmp473842281\",\"Message\":\"\",\"Ref\":{\"TransferType\":\"graphsync\",\"Root\":{\"/\":\"bafkreictaj2jkgz7pkyqtm3x3fdzos5pr7xcwxgdwaotdzictvpa4njjre\"},\"PieceCid\":null,\"PieceSize\":0},\"DealID\":5}]"} 2020-04-03T20:15:42.648Z WARN main lotus-storage-miner/main.go:80 RPC client error: unmarshaling result: failed to parse peer ID: cid too short

laser 10 days ago root@ip-10-0-10-11:/# lotus state power 24576(24 KiB)

laser 10 days ago root@ip-10-0-10-11:/# lotus state proving t01006 0: 6261666b3465687a6173726934797737736735617561777967636136776b636d6c34726179666275336a6932376565356a6864707764337168726e7361 I do have some software which reproduces this issue... but it's pretty complicated (probably doesn't fit the definition of "a script"). That said, if you have AWS credentials that are authorized to provision loads of virtualized network hardware and EC2 instances, then you can use it with minimal modification. It's just a couple of commands: something to bring up the network, something to run the test, and something to destroy the network. https://github.com/laser/aws-filecoin This set of scripts and CloudFormation templates provisions 150 EC2 instances, each running lotus storage miners. They all connect to each other, and then a storage client (running on your local host) attempts to make storage deals with each of them. Then the network is torn down, EC2 instances and VPC destroyed, yadda yadda. (edited)

arajasek commented 4 years ago

3)

Reported here: https://filecoinproject.slack.com/archives/CPFTWMY7N/p1586590716487500

When I execute. / Lotus client deal bafkreigegv6sa6qm3n6eqzfqk7itv2xaijpfs6gc6hu5bfhi2mrcp24uu t01000.0000000000000001 10, except for the first success, the rest will inevitably fail 2020-04-11T15:37:05.516+0800 WARN providerstates providerstates/provider_states.go:292 deal bafyreigkcyiz7pyhn7wdleerrnkxn6xx3r44sl5ivofbvscvmvy75iod2y failed: error calling node: publishing deal: publishing deal failed: exit 1 r=&{Receipt:{ExitCode:SysErrSenderInvalid(1) Return:[] GasUsed:2368} TipSet:0xc01d981280}

long568 commented 4 years ago

I got the same proplem...

magik6k commented 3 years ago

If any of this is still relevant I'd open this as new issues