trace_transaction randomly returns first run for txIndex 0 error

[eldarbogdanov]

trace_transaction sometimes, randomly, returns an error likefirst run for txIndex 0 error: nonce too high: address 0x0000000000000000000000000000000000000000, tx: 12923888 state: 0 in response to trace_transaction requests. This persists for several minutes at a time for a given transaction hash, and autorecovers afterwards. The symptoms seem very similar to #8572. Due to the intermittent nature of the issue, I cannot provide repro steps.

The node in question is up-to-date during these occurrences. Here is the output from integration print_stages that shows its current state (caught up to public Sepolia explorer):

                                 stage_at        prune_at
Snapshots                        5727592         0
Headers                          5727593         0
BorHeimdall                      0               0
BlockHashes                      5727593         0
Bodies                           5727593         0
Senders                          5727593         0
Execution                        5727593         5727593
Translation                      0               0
HashState                        5727593         0
IntermediateHashes               5727593         5727593
AccountHistoryIndex              5727593         0
StorageHistoryIndex              5727593         0
LogIndex                         5727593         0
CallTraces                       5727593         5727593
TxLookup                         5727593         0
Finish                           5727593         0
--
prune distance: 

blocks.v2: false, blocks=0, segments=0, indices=0
blocks.bor.v2: segments=0, indices=0

history.v3: false,  idx steps: 0.00, lastBlockInSnap=0, TxNums_Index(0,1)

sequence: EthTx=209707022, NonCanonicalTx=9927

in db: first header 1, last header 5727593, first body 1, last body 5727593

System information

Erigon version: erigon version 2.59.2-a013ec25 (note that this version includes the #9641 fix)

OS & Version: Alpine Linux v3.17

Commit hash:

Erigon Command (with flags/config):

Consensus Layer: Prysm

Consensus Layer Command (with flags/config):

Chain/Network: Eth Mainnet, Eth Sepolia

[eldarbogdanov]

As an additional note, I have recently observed the same behavior on eth_traceFilter method as well.

[cshintov]

Facing same issue for block 43827539 and trace_block!

curl -s -S --location --request POST http://localhost:8945 --header 'Content-Type: application/json' --data-raw '{
  "jsonrpc": "2.0",
  "method": "trace_block",
  "params": [
    "43827539"
  ],
  "id": 1
}'
{
  "jsonrpc": "2.0",
  "id": 1,
  "error": {
    "code": -32000,
    "message": "first run for txIndex 0 error: nonce too high: address 0x0000000000000000000000000000000000000000, tx: 78295 state: 0"
  }
}

erigon version 2.59.3 , polygon mainnet archive!

[eldarbogdanov]

@cshintov You might want to subscribe to #8572

[yperbasis]

Probably duplicate of #8246 & #8572

[somnergy]

Cannot reproduce at the moment. Will keep trying for a little while and close it as such. Meanwhile, more details and logs are appreciated @eldarbogdanov and participants of #8246 #8572