Closed adizere closed 2 years ago
Here is the logs around the event on Cros-nest's relayer. We are running Hermes with INFO log level
This is not a bug. It's the same as JUNO chain. One of the transactions is a "no-op". In mintscan it shows both as "success", but only one of them actually executed.
This is not a bug. It's the same as JUNO chain. One of the transactions is a "no-op". In mintscan it shows both as "success", but only one of them actually executed.
One instance of hermes send two identical txn in one block. Signer is same. In some part it is a bug tendermint in CheckTX. But another side why hermes send two txn? I'll make a log and attach to the topic.
Problem summary
This may be a bug in the Hermes logic for
process_pending
.These are logs reported in this comment below. The following happens:
comdex-1
, hashA5129724DBE2BCFD87BF02CFB1ED01BE21D1BD67B73DD4E640881C34D50AC14E
comdex-1
, this time with hash5315E04F87880471CDFE49FF1071431D48CCC3232BEF4648F2AEE655BE62BFE0
Acceptance criteria
The problem in the present issue is related to #1792. If that issue is fixed, that will help with fixing this issue, but we should nevertheless increase the timeout.
Original issue description
Two separate relayer operators reported that Hermes submits duplicate transactions.
Originally posted by @dualsystems in https://github.com/informalsystems/ibc-rs/issues/1608#issuecomment-987389966
From a Telegram group:
It's unclear how to replicate this, so some Hermes logs would be useful in debugging. If any relayer operator is open to sharing their logs with us, we would be grateful!