Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Failing to receive blinded blocks on sepolia #12103

Open
mcdee opened this issue Mar 9, 2023 · 14 comments
Open

Failing to receive blinded blocks on sepolia #12103

mcdee opened this issue Mar 9, 2023 · 14 comments
Assignees
Labels
Bug Something isn't working

Comments

@mcdee
Copy link
Contributor

mcdee commented Mar 9, 2023

🐞 Bug Report

Description

A call to obtain a blinded block from prysm on sepolia is failing.

In addition, prysm overrides the request to obtain a blinded block if it thinks the payload is a higher value. It would be handy to have a flag to disable this behavior ,and always return the blinded block with payload from the MEV relay.

Has this worked before in a previous version?

Yes.

🔬 Minimal Reproduction

Sent to prysm:

/eth/v1/validator/blinded_blocks/1882275?randao_reveal=0x97d2a7e4fbe3a21e79ab310a8e0a33104f0c650949d374da9f540756187e37a64b54ac147100edf453b297a897c36bc3119e1f86f1def8382ffab25620e8cbd8a4b9b659e76c11fcc648368b18bc5244989d81280966348345d90bd7231eb4a5&graffiti=0x417474657374616e742f566f7563680000000000000000000000000000000000

🔥 Error

Received from prysm:

{"message":"Unsupported block type","code":400}

prysm logs show:

{"blockHash":"0xd98081ea8366e15d409ce5f7e288258ad938936c6a2489ef9ee72619d2204faa","builderPubKey":"0x845bd072b7cd566f02faeb0a4033ce9399e42839ced64e8b2adcfc859ed1e8e1a5a293336a49feac6d9a5edb779be53a","level":"info","msg":"Received header with bid","prefix":"rpc/validator","time":"2023-03-09T00:15:00Z","value":"957081901767780"}
{"builderValue":957081901767780,"level":"warning","localValue":957081901914780,"msg":"Proposer: using local execution payload because higher value","prefix":"rpc/validator","time":"2023-03-09T00:15:00Z"}

🌍 Your Environment

prysm 3.2.2-rc.3

@nisdas
Copy link
Member

nisdas commented Mar 9, 2023

cc @terencechain

@nisdas nisdas added the Bug Something isn't working label Mar 9, 2023
@potuz
Copy link
Contributor

potuz commented Mar 10, 2023

I would be very strongly against such a flag. One thing is facilitating access to MEV to avoid forks, another very different is actively weakening anti censoring mechanisms

@moshe-blox
Copy link
Contributor

moshe-blox commented Apr 19, 2023

We get the same error on mainnet at the moment, reported in this Discord message:

hey guys,
we have a Prysm v4.0.0 node with --http-mev-relay set, and when we request to produce a blinded block, like so:

GET /eth/v1/validator/blinded_blocks/<slot>?randao_reveal=<randao>&graffiti=<graffiti>

we get this error:

{"message":"Unsupported block type","code":400}

i believe it's most likely an error on our side, because i recall this working a few months ago, i'm just asking for your guidance to figure out why it may be

(this doesn't happen with Lighthouse)


I'm curious, how does Prysm's VC produce blinded blocks? Assuming this error isn't due to a misconfiguration, then I'd guess the VC isn't using this endpoint?

@rkapka
Copy link
Contributor

rkapka commented Apr 19, 2023

Our VC doesn't use the Beacon API

@moshe-blox
Copy link
Contributor

moshe-blox commented Apr 19, 2023

Our VC doesn't use the Beacon API

That's right, I was just curious whether it's doing something entirely different like producing the block by itself.

Back in September, we had an issue where the BN wouldn't unblind submitted blinded blocks (wasn't implemented in neither gRPC or Beacon API), which wasn't an issue for Prysm's VC because it was unblinding blocks by itself, but our VC isn't capable of that (yet 🤔) and relies on this endpoint.

By the way, regarding that (blinded block submission), do you know whether it's now implemented in the Beacon API?

@rkapka
Copy link
Contributor

rkapka commented Apr 19, 2023

Yes, for Capella it is implemented, but I am looking now if there is a bug in the code.

@zheli
Copy link

zheli commented May 25, 2023

Can someone follow up on this issue? Now my validator client (vouch) says "Prepared beacon block is not blinded" and failed the proposal.

@rkapka
Copy link
Contributor

rkapka commented May 25, 2023

Hi @zheli , did you observe a Proposer: using local execution payload because higher value log before the failed proposal?

@zheli
Copy link

zheli commented May 29, 2023

Unfortunately I have removed the the machine together with the log so it's not possible to find out now 😞 I will check next time when we have a missing proposal.

@mcdee
Copy link
Contributor Author

mcdee commented Jun 2, 2023

@rkapka yes we see that message prior to the error.

@hwhe
Copy link

hwhe commented Dec 19, 2023

Has this bug fixed in the latest prysm version?

@rkapka
Copy link
Contributor

rkapka commented Dec 22, 2023

Has this bug fixed in the latest prysm version?

It will be resolved once we implement ethereum/beacon-APIs#386

@mcdee
Copy link
Contributor Author

mcdee commented Jan 12, 2024

This error is still occurring with prysm 4.2.0. Recent call on sepolia:

/eth/v1/validator/blinded_blocks/4111200?randao_reveal=0xb13a6dcebaafa9ad8c429fe90fabf0f6d43cbf12d8f45ed216cd17ed3cb84b12bf6de430552395fc63abef8644f722d0148354d419a33e1f8007bddd0725821d2aa9dc230456c99e575619629fd253a38c2bfbc7ac042dfd5172f1524b74d2cb&graffiti=0x417474657374616e742f566f7563680000000000000000000000000000000000"

With response:

{"message":"Prepared block is not blinded","code":500}
beacon-chain --version
beacon-chain version Prysm/v4.2.0/3a2b486bde28f9b2a02c716782a738c64d62a8e8. Built at: 2024-01-11 01:35:41+00:00

@terencechain
Copy link
Member

Passing it to @rkapka

@linear linear bot closed this as not planned Won't fix, can't repro, duplicate, stale Nov 18, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Bug Something isn't working
Projects
None yet
Development

Successfully merging a pull request may close this issue.

9 participants