-
Notifications
You must be signed in to change notification settings - Fork 721
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
panic: Scraped backing votes had invalid signature #3878
Comments
The backtrace shows that a "Valid" backing vote has got an invalid signature:
I'm not sure how this can happen. What were you testing? Anything off that can cause such behaviour? |
I was testing #2944 with the default LibP2P network backend (not the LiteP2P). There might be something off with the network format, but I doubt it is coming from that. We may have missed these logs over time. To discover it, I created a query that filters lines that contain |
Last time I remember this happening (and the only explanation I got for it) is when there was a dispute and a reversion to a block at least 2 sessions ago. The reason for this is that we cache SessionInfo which includes validator public keys per session index and if we do a past session reversion, technically, this caching is not sound: #633. I wonder if this was the case here or something else? If yes, this is a known issue that is very unlikely to happen in practice (as it requires commiting to an invalid parachain block and risking getting 100% * 2 of validator stake slashed AND stalling finality for a couple of days/many hours on Polkadot/Kusama) and no need to worry about it. |
We've discussed that problem lots, not sure if we ever had really great answers, but not participating in parachains when our session info differs maybe doable, assuming we can tell what session info others use. We then risk attacks where attackers try breaking validator's sesision info, so that must be treated as a soundness risk, but we must do similar anyways. It's probably also broken if attackers can make them flood the system with disputes, and then do attacks while everything is limping along with no validators. |
So we can close this issue? |
This issue was most probably related to our on-spot test instances. Closing this for now, thanks! |
The following was discovered during versi-network triage versi-net logs:
Version: #2944
Extra logs during the incident
cc @paritytech/networking @alexggh
The text was updated successfully, but these errors were encountered: