Closed ebma closed 8 months ago
@pendulum-chain/product this should have a rather high priority because it affects the testing service quite often.
Hey team! Please add your planning poker estimate with Zenhub @b-yap @bogdanS98 @ebma @gianfra-t @TorstenStueber
Quite strange that the sequence number is higher, I don't understand the scenario in which we arrive at that. But I agree that by just removing the cache will solve this since loadAccount
will already fetch the the current sequence number.
@ebma does this require a runtime upgrade?
No it does not.
Sometimes the service gets stuck when retrying after receiving a bad sequence error. It will count up the sequence number indefinitely but never reach a valid one since the number is already too high.
TODO
We should change the retry logic to fetch the latest valid sequence number from the network in case of failure. For this, we probably just need to remove the caching logic here.