You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The JSON RPC APIs for contract-internal transaction tracing are important for auditing block processing, and include debug_traceTransaction and debug_traceBlockByNumber. They should return correct values with reasonable latency a very high percentage of the time.
Actual behaviour
Post-London fork on Ropsten, these APIs return errors with significantly greater frequency (i.e. 2-3 blocks / transactions per day). These errors are often timeouts, and happen regardless of what you set the timeout value to (e.g. 60s, 2m, 3m, etc.). This means that retries after the timeouts don't work - i.e., some blocks/transactions don't return a value no matter how many times the client retries.
For clients for whom these APIs are on the critical path for block processing, this bug is a blocker for London.
0x5a3b240e6cec91a9a1670954092ee599dd1fc62f3bdf573564eb21706b225ea4 - 10,8 M gas. That might hit some timeout, I can imagine.
0xd830832a19b9a7f68a1aeb8510f2986f1811cac4e68ff290abbcde873158d327 - only 34K gas, a legacy transaction. That should not error, and would be simpler to investigate than the first one
It seems that this is fixed by #23016 . Please reopen if you're still having problems. Also, for large-scale opcode-style dumping of transactions, it's better to use debug_standardTraceBlockToFile, which streams json-lines to disk, instead of sending a (several Gb) json blob over RPC
System information
Geth version: 1.10.4
OS & Version: Linux
Expected behaviour
The JSON RPC APIs for contract-internal transaction tracing are important for auditing block processing, and include
debug_traceTransaction
anddebug_traceBlockByNumber
. They should return correct values with reasonable latency a very high percentage of the time.Actual behaviour
Post-London fork on Ropsten, these APIs return errors with significantly greater frequency (i.e. 2-3 blocks / transactions per day). These errors are often timeouts, and happen regardless of what you set the timeout value to (e.g. 60s, 2m, 3m, etc.). This means that retries after the timeouts don't work - i.e., some blocks/transactions don't return a value no matter how many times the client retries.
For clients for whom these APIs are on the critical path for block processing, this bug is a blocker for London.
Steps to reproduce the behaviour
Run against a Ropsten Geth node, e.g.:
Also reproduceable for the following transactions:
For block tracing:
Also reproduceable for the following blocks:
Backtrace
When submitting logs: please submit them as text and not screenshots.
The text was updated successfully, but these errors were encountered: