Skip to content
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.

Sign up for GitHub

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

Jump to bottom

Test false negative dispute spam #907

Open
eskimor opened this issue Nov 23, 2021 · 0 comments
Open

Test false negative dispute spam #907

eskimor opened this issue Nov 23, 2021 · 0 comments
Assignees
@eskimor @Lldenaurois

Comments

@eskimor
Copy link
Member

eskimor commented Nov 23, 2021

If there are lots of false negative disputes (disputes where validators vote invalid, although it is valid), things could go wrong:

  1. If participation does not happen in a proper order, disputes might never conclude.
  2. If there are lots of recent disputes, we rely on filtering to keep block sizes sane
  3. If the filtering is wrong, disputes might never conclude on chain again (e.g. same disputes are imported all over again)
  4. If filtering is taking too long - block author might skip its block.
  5. If weights in filtering are wrong, disputes might not get included in block due to weight/ block might get skipped.
  6. Some random bug triggers, I have not thought about

To test if things are working properly we want to create a high volume of those disputes and make sure that:

  1. Nothing breaks
  2. Disputes keep getting resolved no matter what

We don't want to break Rococo, so what could go wrong and would it cause breakage?

ad 1) no - disputes just don't conclude.
ad 2) if blocks get large - block times rise, but that should not cause breakage
ad 3) Disputes would not work, but again no breakage
ad 4) If a block gets skipped, after every two blocks the timeout gets doubled - so eventually a block producer will be able to build a block, (if we don't go completely crazy)
ad 5) If disputes won't get included - again no breakage. If block creation and block import code are out of sync and we would really discard the block, then yes we would break the chain.
ad 6) might break chain

5 should not happen and if it does, finding that out, would arguably be worth breaking Rococo (better than breaking Kusama/Polkadot later)

For 6 the same argument holds.

Results

Ideally we should be seeing that disputes are being resolved on chain no matter what and also block times should stay within reason.

Execution

  1. Prepare validator nodes with logging for the following targets:

parachain::dispute-coordinator - trace
parachain::dispute-distribution - debug
babe - debug

  1. Launch malicious validators on Rococo, slowly increasing until block times increase/we see we are pushing the limits.
  2. Check metrics - see how long creating the inherent takes, find out what is actually pushing the limits
  3. Push a bit further to confirm any theories
  4. Continue until we are reasonably sure we know that happens and see that things work within expectations.
@Sophia-Gold Sophia-Gold transferred this issue from paritytech/polkadot Aug 24, 2023
claravanstaden added a commit to Snowfork/polkadot-sdk that referenced this issue Dec 8, 2023
@claravanstaden @vgeddes
@doubledup @alistair-singh @yrong
Beacon Fuzz testing (paritytech#907)
60a82e0 
* Improve contracts

* More improvements to contracts

* Fix parachain build in proxy-contracts branch (paritytech#889)

* Upgrade ssz_rs crate. (paritytech#880)

* Upgrade ssz_rs crate.

* Upgrade ssz_rs crate.

---------

Co-authored-by: claravanstaden <Cats 4 life!>

* Bump nixpkgs to use its geth package (paritytech#885)

* Bump nixpkgs to use its geth package

* Use source config in PolkadotListener

* Match field order in struct def

* Move info log to relay creation

* Remove unused channel

* Whitespace

* Add context to errors

* Add logs

* Bump node & pnpm in workflow

* Rename locals

* Remove unused variable

* Add Troubleshooting README section

* Fix up .envrc-example files

* Add note about pure shells

* Update cumulus submodule (paritytech#886)

* Inbound queue benchmarks (paritytech#876)

* Start with inbound channel benchmarks.

* Add method to set execution header storage for benchmark test.

* Working on benchmarks

* Basic working version

* Cleanup

* Removes cleanup.

* Adds some comments for Alistair.

* Adds branch name.

* Makes note

* Test transactions

* Cleaning up beacon client deps.

* Clean up comments.

* Tests cleanup.

* Fixes non-benchmark test runs.

* Cleanup.

* Update fixtures and generates benchmarks.

* Revert relayer logs.

* Cleanup BenchmarkHelper impl and inbound queue dependencies.

* fmt

* Cleanup imports.

* Cleanup imports.

* Touch

* Adds weights in inbound queue pallet.

* Fix tests.

* Update cumulus.

---------

Co-authored-by: claravanstaden <Cats 4 life!>

* Fix parachain build

* Move BalanceOf outside of pallet

* remove benchmark for non existing method

* downgrade cargo.lock to match cumulus

* fix benchmarks

---------

Co-authored-by: Clara van Staden <[email protected]>
Co-authored-by: Alistair Singh <[email protected]>

* forge install: openzeppelin-contracts

v4.9.2

* sno-472 rebased (paritytech#890)

* Halting & resuming bridge pallets

* Ignore .env

* Remove .env

* Some polish

* Set owner of bridge pallets

* Upgrade ssz_rs crate. (paritytech#880)

* Upgrade ssz_rs crate.

* Upgrade ssz_rs crate.

---------

Co-authored-by: claravanstaden <Cats 4 life!>

* Update cumulus

* Bump nixpkgs to use its geth package (paritytech#885)

* Bump nixpkgs to use its geth package

* Use source config in PolkadotListener

* Match field order in struct def

* Move info log to relay creation

* Remove unused channel

* Whitespace

* Add context to errors

* Add logs

* Bump node & pnpm in workflow

* Rename locals

* Remove unused variable

* Add Troubleshooting README section

* Fix up .envrc-example files

* Add note about pure shells

* Relax RANDAO_COMMIT_DELAY for local setup

* Update cumulus

* Update cumulus submodule (paritytech#886)

* Update cumulus

* Inbound queue benchmarks (paritytech#876)

* Start with inbound channel benchmarks.

* Add method to set execution header storage for benchmark test.

* Working on benchmarks

* Basic working version

* Cleanup

* Removes cleanup.

* Adds some comments for Alistair.

* Adds branch name.

* Makes note

* Test transactions

* Cleaning up beacon client deps.

* Clean up comments.

* Tests cleanup.

* Fixes non-benchmark test runs.

* Cleanup.

* Update fixtures and generates benchmarks.

* Revert relayer logs.

* Cleanup BenchmarkHelper impl and inbound queue dependencies.

* fmt

* Cleanup imports.

* Cleanup imports.

* Touch

* Adds weights in inbound queue pallet.

* Fix tests.

* Update cumulus.

---------

Co-authored-by: claravanstaden <Cats 4 life!>

* Fix test

* Fix parachain build

* Move BalanceOf outside of pallet

* remove benchmark for non existing method

* downgrade cargo.lock to match cumulus

* fix benchmarks

* Clara/sno 552 (paritytech#887)

* Spacing

* Spacing

* Undo typo.

* Minor updates.

* Adds comment about IrrelevantUpdate.

* One more comment.

* Update error name.

---------

Co-authored-by: claravanstaden <Cats 4 life!>

* Halting & resuming bridge pallets (paritytech#883)

* Halting & resuming bridge pallets

* Ignore .env

* Remove .env

* Some polish

* Set owner of bridge pallets

* Update cumulus

* Relax RANDAO_COMMIT_DELAY for local setup

* Update cumulus

* Update cumulus

* Fix test

* Fix Warnings

* Fix test

* Fix build & format

* Fix benchmark test

* Check for duplicate versions of substrate and polkadot (paritytech#891)

* modified pre-commit

* fixes

* testing

* testing

* testing

* fixed tests

* Format

* Some fix

* Update cumulus

* Update cumulus

---------

Co-authored-by: Clara van Staden <[email protected]>
Co-authored-by: David Dunn <[email protected]>
Co-authored-by: Alistair Singh <[email protected]>

* Proxy contracts Tests fixes (paritytech#892)

* fixed tests

* warnings and imports

* rustfmt

* updated cumulus

* Revert rustfmt (paritytech#896)

* Revert "rustfmt"

This reverts commit b83cec7929cdcc6ac972a2b1cad3a0e1fde81870.

* reverted xcm-builder

* Create agent (paritytech#895)

* base

* removed location conversion

* completed implementation

* remove xcm-builder

* update cumulus

* update cumulus

* use contains_key

* Fix openzeppelin-contracts submodule

---------

Co-authored-by: David Dunn <[email protected]>

* improve API for sending tokens

* Messy working version.

* Cleanup.

* More cleanup.

* Rollback unnecessary changes.

* Cleanup whitespace.

* Fix tests.

* Fuzz the submit method.

* Revert rebase oopsies.

* More fuzzing.

* More fuzzing.

* Last extrinsic.

* Remove unnecessary dependency.

* Revert readme.

* Fix tests and feature issues.

* Cleanup types.

* More cleanup.

* Call extrinsics directly. Adds readme.

* Adds CI.

* cd to correct dir

* Update CI.

* Correct nightly param.

* Remove runs.

* Own impl for SyncCommittee.

* Remove rng.
Cleans up workflow.

* Last cleanup.

* Update parachain/pallets/ethereum-beacon-client/fuzz/Cargo.toml

Co-authored-by: David Dunn <[email protected]>

* Adds cargo-fuzz to init script.

* PR comment changes.

* Update parachain/pallets/ethereum-beacon-client/fuzz/src/impls.rs

Co-authored-by: David Dunn <[email protected]>

* Update rust-toolchain.toml

Co-authored-by: David Dunn <[email protected]>

* Update .github/workflows/parachain.yml

Co-authored-by: David Dunn <[email protected]>

* Less runs for shorter Github actions.

---------

Co-authored-by: Vincent Geddes <[email protected]>
Co-authored-by: David Dunn <[email protected]>
Co-authored-by: Alistair Singh <[email protected]>
Co-authored-by: Ron <[email protected]>
Co-authored-by: claravanstaden <Cats 4 life!>
bkchr pushed a commit that referenced this issue Apr 10, 2024
@HCastano @bkchr
Prettify test output when encoding doesn't match (#907)
b8e021c
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@eskimor eskimor

@Lldenaurois Lldenaurois

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@eskimor @Lldenaurois