Resetting connection to RPC node, crashes the node.

[nick-u410]

Tested on the SE chain Version of node: Namada 0.32.1

When running a program that fetches each block from the rpc node, if you close the program mid run, it resets the connection to the RPC node which apparently crashes it. I can consistently reproduce this.

A third party should not be able to crash my node simply due to resetting a connection, a user could bot this and take down all public RPC nodes.

Other users in discord are also experiencing this https://discord.com/channels/833618405537218590/1192974425423892550/1209072443243888720 https://discord.com/channels/833618405537218590/1192974425423892550/1206927791388626945 https://discord.com/channels/833618405537218590/1186723196527980687/1186732226478624919 https://discord.com/channels/833618405537218590/996760681892290611/1186368562244358164

[Fraccaman]

This is probably a problem in tower-abci (one of our dependency) as @sug0 was pointing out. We need to investigate. Thanks for reporting!

[brentstone]

@tzemanovic is this still relevant or have we deemed it solved in one of our latest releases somehow?

[tzemanovic]

@tzemanovic is this still relevant or have we deemed it solved in one of our latest releases somehow?

@brentstone From my testing this should no longer be an issue since v0.33 - resetting a connection is handled gracefully and reported by CometBFT as e.g. failed to write responses module=rpc-server err="write tcp ...: i/o timeout.

@nick-u410 Could you please check this from your side and let us know if that's not the case?

[cwgoes]

Closing this for now, but feel free to reopen if this issue recurs.