eth_getLogs request returns an error failed to get logs for block.

[brg8]

System information

Geth version: v1.10.11
OS & Version: Linux
Commit hash : N/A

Expected behaviour

I expect the following request to return a non-error result.

curl localhost:8545 -X POST -H "Content-Type: application/json" -d '{"id": "someid","method": "eth_getLogs","jsonrpc": "2.0","params": [ { "fromBlock": "0xb45a", "toBlock": "0xb45a" } ]}'

Actual behaviour

It returns an error.

{"jsonrpc":"2.0","id":"someid","error":{"code":-32000,"message":"failed to get logs for block"}}

Steps to reproduce the behaviour

Run the provided curl against a geth node. Some bullets that might help:

• I have reproduced this on 10 separate archive nodes.
• Full nodes are hit or miss, some of them respond correctly.
• Parity/OE and Erigon nodes respond correctly.
• When I get the error, this appears in the geth logs:

ERROR[11-08|20:15:32.914] Invalid receipt array RLP                hash=4e3a37..5a1bdd err="rlp: expected input list for []*types.LogForStorage, decoding into ([]*rawdb.receiptLogs)[0](rawdb.storedReceiptRLP).Logs"

• If I add an address to the filters then I get a valid response:

curl localhost:8545 -X POST -H "Content-Type: application/json" -d '{"id": "someid","method": "eth_getLogs","jsonrpc": "2.0","params": [ { "fromBlock": "0xb45a", "toBlock": "0xb45a", "address": "0x1ad1C9CE861a072f843Eab9c070A52e295ad94B5" } ]}'

{"jsonrpc":"2.0","id":"someid","result":[]}

Thank you in advance for your help!

[noam-alchemy]

just tested on 1.10.12 as well and the issue is reproducible there as well.

[s1na]

Thanks for reporting the issue and the details. If I'm not mistaken this is caused by #23147 and affects really long-running nodes (2-3 years with no db reset) and specifically when querying logs of older blocks. The cause is that when implementing the optimization in #23147 I hadn't considered nodes that might have a legacy format for receipts in their freezer database. I'm working on a fix now.

[s1na]

Can you please try this branch and see if that solve the issue? https://github.com/s1na/go-ethereum/tree/fix/getlogs-legacy

[brg8]

@s1na Yes! That fixes it :)

[miohtama]

I just encountered this issue with ethereum/client-go:v1.10.19.

I received one of the:

{'code': -32000, 'message': 'failed to get logs for block #14682553 (0x6d0a1e..590911)

The geth log entry is the same message without further context:

ethereum     | WARN [08-03|18:50:13.633] Served eth_getLogs                       conn=127.0.0.1:56250 reqid=19  duration=595.788162ms err="failed to get logs for block #11802450 (0x638c92..2bad0b)"

I manage to repeat this at least two different blocks on my database.

Though could not repeat this yet. I wonder if there are other reasons why this get logs could fail.

[miohtama]

Upgraded to ethereum/client-go:v1.10.21, still getting the same error.

[symbolpunk]

A Prysm user is encountering this error here: https://discord.com/channels/476244492043812875/667761857209040916/1019616603542921236

Will update this with Geth version if/when provided.

[shibug]

A Prysm user is encountering this error here: https://discord.com/channels/476244492043812875/667761857209040916/1019616603542921236

Will update this with Geth version if/when provided.

I am the one who is facing this issue. My database was fast synced in the past 2 weeks using Geth 1.10.23. So, no old parts here

[shibug]

I went through the prysm checklist only step #9 fails. Here is the output:
curl http://localhost:3500/eth/v1alpha1/node/eth1/connections {"currentAddress":"http://localhost:8551","currentConnectionError":"failed to get logs for block #15510142 (0x367e14..8e5b43)","addresses":["http://localhost:8551"],"connectionErrors":[]}%

[TheCookieLab]

I'm another Prysm user seeing this error message on repeat:

Geth (1.10.25+build28116+focal) is fully synced / not optimistic.

[shibug]

I'm another Prysm user seeing this error message on repeat:

Geth (1.10.25+build28116+focal) is fully synced / not optimistic.

Yep, just like mine but on a different block

[shibug]

Someone should open this issue

[TheCookieLab]

This kept causing my Geth to eventually unsync. I deleted my chaindata and synced from scratch which resolved things permanently. If you are still experiencing this I would suggest you do the same.

[lymweb]

Yep, just like mine but on a different block

[holiman]

If this issue persists, please open a new ticket, and provide logs from your systems + version information.