-
Notifications
You must be signed in to change notification settings - Fork 3.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Crisis Invariant check per block height #3967
Comments
That would be nice, but wouldn't it require that all nodes store all past state versions? I think the most concerning invariant violations are likely to break an invariant that will remain broken - admittedly, it's hard to formalize that intuition, there could be exceptional cases. |
Does cosmos store a certain amount of previous state? but this seems like way less of a priority given #3969 |
There is a default pruning strategy, which IIRC is fairly aggressive (prunes all but the most recent states). I don't think we can reasonably expect node operators to store all past states (which we would require if it were part of the state machine like this). Agreed that #3969 will reduce the need for this. |
@cwgoes as long as there are a few blocks worth of state stored then we can at least be reassured that the correct circuit breaker occurs.. like it could be a part of this transaction that if you want to include a block height it can only occur within the past say 5 blocks from the current height (or else reject the transaction) - this would be sufficient to ensure that auto-sent transactions (per #3969) would definitely circuit break correctly. Of course I don't think it's reasonable for any more than a few blocks to be stored by the node operators. but yeah this is low low priority AFAICT haha - for the long game |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. |
closing for now since we will end up rewriting how crisis works |
Offshoot of #3656
Currently, the invariant check messages run at the latest height of the blockchain. This could be a problem if the invariant was temporarily broken (and thus should have halted the blockchain) but was then somehow corrected at a later height. Thus we ought to implement the ability to run an invariant check at a specific height of the blockchain.
The text was updated successfully, but these errors were encountered: