Attempting to improve our CI speeds as its recently been a pain point.
Major changes:
- Use a github action to pull stable/nightly rust rather than building it each run
- Shift test suite to `nexttest` https://github.com/nextest-rs/nextest for CI
UPDATE:
So I've iterated on some changes, and although I think its still not optimal I think this is a good base to start from. Some extra things in this PR:
- Shifted where we pull rust from. We're now using this thing: https://github.com/moonrepo/setup-rust . It's got some interesting cache's built in, but was not seeing the gains that Jimmy managed to get. In either case tho, it can pull rust, cargofmt, clippy, cargo nexttest all in < 5s. So I think it's worthwhile.
- I've grouped a few of the check-like tests into a single test called `code-test`. Although we were using github runners in parallel which may be faster, it just seems wasteful. There were like 4-5 tests, where we would pull lighthouse, compile it, then run an action, like clippy, cargo-audit or fmt. I've grouped these into a single action, so we only compile lighthouse once, then in each step we run the checks. This avoids compiling lighthouse like 5 times.
- Ive made doppelganger tests run on our local machines to avoid pulling foundry, building and making lcli which are all now baked into the images.
- We have sccache and do not incremental compile lighthouse
Misc bonus things:
- Cargo update
- Fix web3 signer openssl keys which is required after a cargo update
- Use mock_instant in an LRU cache test to avoid non-deterministic test
- Remove race condition in building web3signer tests
There's still some things we could improve on. Such as downloading the EF tests every run and the web3-signer binary, but I've left these to be out of scope of this PR. I think the above are meaningful improvements.
Co-authored-by: Paul Hauner <paul@paulhauner.com>
Co-authored-by: realbigsean <seananderson33@gmail.com>
Co-authored-by: antondlr <anton@delaruelle.net>
## Issue Addressed
We're OOM'ing on Docker builds on the Deneb branch https://github.com/sigp/lighthouse/issues/3929
Are we ok to self host automated docker builds?
Co-authored-by: realbigsean <seananderson33@gmail.com>
Co-authored-by: realbigsean <sean@sigmaprime.io>
Co-authored-by: antondlr <anton@delaruelle.net>
## Issue Addressed
I noticed a node.js version warning on our CI, and thought about updating the version to get rid of this warning, but then realized we may not need node.js anymore now we're using `anvil` instead of `ganache`.
> The following actions uses node12 which is deprecated and will be forced to run on node16: actions/setup-node@v2. For more info: https://github.blog/changelog/2023-06-13-github-actions-all-actions-will-run-on-node16-instead-of-node12-by-default/
## Issue Addressed
Upgrade libp2p to v0.52
## Proposed Changes
- **Workflows**: remove installation of `protoc`
- **Book**: remove installation of `protoc`
- **`Dockerfile`s and `cross`**: remove custom base `Dockerfile` for cross since it's no longer needed. Remove `protoc` from remaining `Dockerfiles`s
- **Upgrade `discv5` to `v0.3.1`:** we have some cool stuff in there: no longer needs `protoc` and faster ip updates on cold start
- **Upgrade `prometheus` to `0.21.0`**, now it no longer needs encoding checks
- **things that look like refactors:** bunch of api types were renamed and need to be accessed in a different (clearer) way
- **Lighthouse network**
- connection limits is now a behaviour
- banned peers no longer exist on the swarm level, but at the behaviour level
- `connection_event_buffer_size` now is handled per connection with a buffer size of 4
- `mplex` is deprecated and was removed
- rpc handler now logs the peer to which it belongs
## Additional Info
Tried to keep as much behaviour unchanged as possible. However, there is a great deal of improvements we can do _after_ this upgrade:
- Smart connection limits: Connection limits have been checked only based on numbers, we can now use information about the incoming peer to decide if we want it
- More powerful peer management: Dial attempts from other behaviours can be rejected early
- Incoming connections can be rejected early
- Banning can be returned exclusively to the peer management: We should not get connections to banned peers anymore making use of this
- TCP Nat updates: We might be able to take advantage of confirmed external addresses to check out tcp ports/ips
Co-authored-by: Age Manning <Age@AgeManning.com>
Co-authored-by: Akihito Nakano <sora.akatsuki@gmail.com>
## Issue Addressed
NA
## Proposed Changes
Carries on from #4115, with the following modifications:
1. Self-hosted runners are only enabled if `github.repository == sigp/lighthouse`.
- This allows forks to still have Github-hosted CI.
- This gives us a method to switch back to Github-runners if we have extended downtime on self-hosted.
1. Does not remove any existing dependency builds for Github-hosted runners (e.g., installing the latest Rust).
1. Adds the `WATCH_HOST` environment variable which defines where we expect to find the postgres db in the `watch` tests. This should be set to `host.docker.internal` for the tests to pass on self-hosted runners.
## Additional Info
NA
Co-authored-by: antondlr <anton@delaruelle.net>
## Issue Addressed
Speed up CI by installing foundry with Github action instead of building Anvil from source.
Building anvil from source on GItHub hosted runners currently takes about 10 mins. Using the `foundry-toolchain` action to install only takes about 2 seconds.
## Issue Addressed
Closes#3964
## Proposed Changes
Use the `maxperf` profile to build Windows binaries, now that Rust 1.70.0 fixed the underlying issue.
## Issue Addressed
Workaround for https://github.com/foundry-rs/foundry/issues/5115.
## Proposed Changes
Allow Anvil to be installed on Windows without errors by enabling the IPC features (which we don't use, but Anvil expects to exist).
## Issue Addressed
N/A
## Proposed Changes
Modifies the local testnet scripts to start a network with genesis validators embedded into the genesis state. This allows us to start a local testnet without the need for deploying a deposit contract or depositing validators pre-genesis.
This also enables us to start a local test network at any fork we want without going through fork transitions. Also adds scripts to start multiple geth clients and peer them with each other and peer the geth clients with beacon nodes to start a post merge local testnet.
## Additional info
Adds a new lcli command `mnemonics-validators` that generates validator directories derived from a given mnemonic.
Adds a new `derived-genesis-state` option to the `lcli new-testnet` command to generate a genesis state populated with validators derived from a mnemonic.
## Issue Addressed
N/A
## Proposed Changes
Replace ganache-cli with anvil https://github.com/foundry-rs/foundry/blob/master/anvil/README.md
We can lose all js dependencies in CI as a consequence.
## Additional info
Also changes the ethers-rs version used in the execution layer (for the transaction reconstruction) to a newer one. This was necessary to get use the ethers utils for anvil. The fixed execution engine integration tests should catch any potential issues with the payload reconstruction after #3592
Co-authored-by: Michael Sproul <michael@sigmaprime.io>
## Issue Addressed
The latest stable version (1.69.0) of Rust was released on 20 April and contains this change:
- [Update the minimum external LLVM to 14.](https://github.com/rust-lang/rust/pull/107573/)
This impacts some of our CI workflows (build and release-test-windows) that uses LLVM 13.0. This PR updates the workflows to install LLVM 15.0.
**UPDATE**: Also updated `h2` to address [this issue](https://github.com/advisories/GHSA-f8vr-r385-rh5r)
## Issue Addressed
There was a [`VecDeque` bug](https://github.com/rust-lang/rust/issues/108453) in some recent versions of the Rust standard library (1.67.0 & 1.67.1) that could cause Lighthouse to panic (reported by `@Sea Monkey` on discord). See full logs below.
The issue was likely introduced in Rust 1.67.0 and [fixed](https://github.com/rust-lang/rust/pull/108475) in 1.68, and we were able to reproduce the panic ourselves using [@michaelsproul's fuzz tests](https://github.com/michaelsproul/lighthouse/blob/fuzz-lru-time-cache/beacon_node/lighthouse_network/src/peer_manager/fuzz.rs#L111) on both Rust 1.67.0 and 1.67.1.
Users that uses our Docker images or binaries are unlikely affected, as our Docker images were built with `1.66`, and latest binaries were built with latest stable (`1.68.2`). It likely impacts user that builds from source using Rust versions 1.67.x.
## Proposed Changes
Bump Rust version (MSRV) to latest stable `1.68.2`.
## Additional Info
From `@Sea Monkey` on Lighthouse Discord:
> Crash on goerli using `unstable` `dd124b2d6804d02e4e221f29387a56775acccd08`
```
thread 'tokio-runtime-worker' panicked at 'Key must exist', /mnt/goerli/goerli/lighthouse/common/lru_cache/src/time.rs:68:28
stack backtrace:
Apr 15 09:37:36.993 WARN Peer sent invalid block in single block lookup, peer_id: 16Uiu2HAm6ZuyJpVpR6y51X4Enbp8EhRBqGycQsDMPX7e5XfPYznG, error: WouldRevertFinalizedSlot { block_slot: Slot(5420212), finalized_slot: Slot(5420224) }, root: 0x10f6…3165, service: sync
0: rust_begin_unwind
at /rustc/d5a82bbd26e1ad8b7401f6a718a9c57c96905483/library/std/src/panicking.rs:575:5
1: core::panicking::panic_fmt
at /rustc/d5a82bbd26e1ad8b7401f6a718a9c57c96905483/library/core/src/panicking.rs:64:14
2: core::panicking::panic_display
at /rustc/d5a82bbd26e1ad8b7401f6a718a9c57c96905483/library/core/src/panicking.rs:135:5
3: core::panicking::panic_str
at /rustc/d5a82bbd26e1ad8b7401f6a718a9c57c96905483/library/core/src/panicking.rs:119:5
4: core::option::expect_failed
at /rustc/d5a82bbd26e1ad8b7401f6a718a9c57c96905483/library/core/src/option.rs:1879:5
5: lru_cache::time::LRUTimeCache<Key>::raw_remove
6: lighthouse_network::peer_manager::PeerManager<TSpec>::handle_ban_operation
7: lighthouse_network::peer_manager::PeerManager<TSpec>::handle_score_action
8: lighthouse_network::peer_manager::PeerManager<TSpec>::report_peer
9: network::service::NetworkService<T>::spawn_service::{{closure}}
10: <futures_util::future::select::Select<A,B> as core::future::future::Future>::poll
11: <futures_util::future::future::map::Map<Fut,F> as core::future::future::Future>::poll
12: <futures_util::future::future::flatten::Flatten<Fut,<Fut as core::future::future::Future>::Output> as core::future::future::Future>::poll
13: tokio::loom::std::unsafe_cell::UnsafeCell<T>::with_mut
14: tokio::runtime::task::core::Core<T,S>::poll
15: tokio::runtime::task::harness::Harness<T,S>::poll
16: tokio::runtime::scheduler::multi_thread::worker::Context::run_task
17: tokio::runtime::scheduler::multi_thread::worker::Context::run
18: tokio::macros::scoped_tls::ScopedKey<T>::set
19: tokio::runtime::scheduler::multi_thread::worker::run
20: tokio::loom::std::unsafe_cell::UnsafeCell<T>::with_mut
21: tokio::runtime::task::core::Core<T,S>::poll
22: tokio::runtime::task::harness::Harness<T,S>::poll
23: tokio::runtime::blocking::pool::Inner::run
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
Apr 15 09:37:37.069 INFO Saved DHT state service: network
Apr 15 09:37:37.070 INFO Network service shutdown service: network
Apr 15 09:37:37.132 CRIT Task panic. This is a bug! advice: Please check above for a backtrace and notify the developers, message: <none>, task_name: network
Apr 15 09:37:37.132 INFO Internal shutdown received reason: Panic (fatal error)
Apr 15 09:37:37.133 INFO Shutting down.. reason: Failure("Panic (fatal error)")
Apr 15 09:37:37.135 WARN Unable to free worker error: channel closed, msg: did not free worker, shutdown may be underway
Apr 15 09:37:39.350 INFO Saved beacon chain to disk service: beacon
Panic (fatal error)
```
## Proposed Changes
* Bump Go from 1.17 to 1.20. The latest Geth release v1.11.0 requires 1.18 minimum.
* Prevent a cache miss during payload building by using the right fee recipient. This prevents Geth v1.11.0 from building a block with 0 transactions. The payload building mechanism is overhauled in the new Geth to improve the payload every 2s, and the tests were failing because we were falling back on a `getPayload` call with no lookahead due to `get_payload_id` cache miss caused by the mismatched fee recipient. Alternatively we could hack the tests to send `proposer_preparation_data`, but I think the static fee recipient is simpler for now.
* Add support for optionally enabling Lighthouse logs in the integration tests. Enable using `cargo run --release --features logging/test_logger`. This was very useful for debugging.
## Issue Addressed
The documentation referring to build from source mismatches with the what gitworkflow uses.
aa5b7ef783/book/src/installation-source.md (L118-L120)
## Proposed Changes
Because the github workflow uses `cross` to build from source and for that build there is different env variable `CROSS_FEATURES` so need pass at the compile time.
## Additional Info
Verified that existing `-dev` builds does not contains the `minimal` spec enabled.
```bash
> docker run --rm --name node-5-cl-lighthouse sigp/lighthouse:latest-amd64-unstable-dev lighthouse --version
Lighthouse v3.4.0-aa5b7ef
BLS library: blst-portable
SHA256 hardware acceleration: true
Allocator: jemalloc
Specs: mainnet (true), minimal (false), gnosis (true)
```
## Proposed Changes
There are some features that are enabled/disabled with the `FEATURES` env variable. This PR would introduce a pattern to introduce docker images based on those features. This can be useful later on to have specific images for some experimental features in the future.
## Additional Info
We at Lodesart need to have `minimal` spec support for some cross-client network testing. To make it efficient on the CI, we tend to use minimal preset.
## Proposed Changes
Another `tree-states` motivated PR, this adds `jemalloc` as the default allocator, with an option to use the system allocator by compiling with `FEATURES="" make`.
- [x] Metrics
- [x] Test on Windows
- [x] Test on macOS
- [x] Test with `musl`
- [x] Metrics dashboard on `lighthouse-metrics` (https://github.com/sigp/lighthouse-metrics/pull/37)
Co-authored-by: Michael Sproul <micsproul@gmail.com>
I've needed to do this work in order to do some episub testing.
This version of libp2p has not yet been released, so this is left as a draft for when we wish to update.
Co-authored-by: Diva M <divma@protonmail.com>
## Proposed Changes
With proposer boosting implemented (#2822) we have an opportunity to re-org out late blocks.
This PR adds three flags to the BN to control this behaviour:
* `--disable-proposer-reorgs`: turn aggressive re-orging off (it's on by default).
* `--proposer-reorg-threshold N`: attempt to orphan blocks with less than N% of the committee vote. If this parameter isn't set then N defaults to 20% when the feature is enabled.
* `--proposer-reorg-epochs-since-finalization N`: only attempt to re-org late blocks when the number of epochs since finalization is less than or equal to N. The default is 2 epochs, meaning re-orgs will only be attempted when the chain is finalizing optimally.
For safety Lighthouse will only attempt a re-org under very specific conditions:
1. The block being proposed is 1 slot after the canonical head, and the canonical head is 1 slot after its parent. i.e. at slot `n + 1` rather than building on the block from slot `n` we build on the block from slot `n - 1`.
2. The current canonical head received less than N% of the committee vote. N should be set depending on the proposer boost fraction itself, the fraction of the network that is believed to be applying it, and the size of the largest entity that could be hoarding votes.
3. The current canonical head arrived after the attestation deadline from our perspective. This condition was only added to support suppression of forkchoiceUpdated messages, but makes intuitive sense.
4. The block is being proposed in the first 2 seconds of the slot. This gives it time to propagate and receive the proposer boost.
## Additional Info
For the initial idea and background, see: https://github.com/ethereum/consensus-specs/pull/2353#issuecomment-950238004
There is also a specification for this feature here: https://github.com/ethereum/consensus-specs/pull/3034
Co-authored-by: Michael Sproul <micsproul@gmail.com>
Co-authored-by: pawan <pawandhananjay@gmail.com>
## Issue Addressed
Closes#3709
## Proposed Changes
Add the job `compile-with-beta-compiler` to `test-suite`. This job has the following steps:
1. Use `actions/checkout@v3`. (Needed to run make in a later step.)
2. Install the dependencies listed in [build from source guide](https://lighthouse-book.sigmaprime.io/installation-source.html).
3. Change the compiler to the current beta version with `rustup override`.
4. Run `make`.
## Issue Addressed
Closes https://github.com/sigp/lighthouse/issues/3656
## Proposed Changes
* Replace `set-output` by `$GITHUB_OUTPUT` usage
* Avoid rate-limits when installing `protoc` by making authenticated requests (continuation of https://github.com/sigp/lighthouse/pull/3621)
* Upgrade all Ubuntu 18.04 usage to 22.04 (18.04 is end of life)
* Upgrade macOS-latest to explicit macOS-12 to silence warning
* Use `actions/checkout@v3` and `actions/cache@v3` to avoid deprecated NodeJS v12
## Additional Info
Can't silence the NodeJS warnings entirely due to https://github.com/sigp/lighthouse/issues/3705. Can fix that in future.
## Issue Addressed
The release CI is currently broken due to the addition of the `protoc` dependency. Here's a failure of the release flow running on my fork: https://github.com/michaelsproul/lighthouse/actions/runs/3155541478/jobs/5134317334
## Proposed Changes
- Install `protoc` on Windows and Mac so that it's available for `cargo install`.
- Install an x86_64 binary in the Cross image for the aarch64 platform: we need a binary that runs on the host, _not_ on the target.
- Fix `macos` local testnet CI by using the Github API key to dodge rate limiting (this issue: https://github.com/actions/runner-images/issues/602).
## Proposed Changes
Add a new Cargo compilation profile called `maxperf` which enables more aggressive compiler optimisations at the expense of compilation time.
Some rough initial benchmarks show that this can provide up to a 25% reduction to run time for CPU bound tasks like block processing: https://docs.google.com/spreadsheets/d/15jHuZe7lLHhZq9Nw8kc6EL0Qh_N_YAYqkW2NQ_Afmtk/edit
The numbers in that spreadsheet compare the `consensus-context` branch from #3604 to the same branch compiled with the `maxperf` profile using:
```
PROFILE=maxperf make install-lcli
```
## Additional Info
The downsides of the maxperf profile are:
- It increases compile times substantially, which will particularly impact low-spec hardware. Compiling `lcli` is about 3x slower. Compiling Lighthouse is about 5x slower on my 5950X: 17m 38s rather than 3m 28s.
As a result I think we should not enable this everywhere by default.
- **Option 1**: enable by default for our released binaries. This gives the majority of users the fastest version of `lighthouse` possible, at the expense of slowing down our release CI. Source builds will continue to use the default `release` profile unless users opt-in to `maxperf`.
- **Option 2**: enable by default for source builds. This gives users building from source an edge, but makes them pay for it with compilation time.
I think I would prefer Option 1. I'll try doing some benchmarking to see how long a maxperf build of Lighthouse would take on GitHub actions.
Credit to Nicholas Nethercote for documenting these options in the Rust Performance Book: https://nnethercote.github.io/perf-book/build-configuration.html.
## Issue Addressed
I think the antithesis is failing due to an OOM which may be resolved by updating the ubuntu image it runs on. The lcli build looks like it's failing because the image lacks the `libclang` dependency
Co-authored-by: realbigsean <sean@sigmaprime.io>
## Issue Addressed
NA
## Proposed Changes
Bump versions to v3.0.0
## Additional Info
- ~~Blocked on #3439~~
- ~~Blocked on #3459~~
- ~~Blocked on #3463~~
- ~~Blocked on #3462~~
- ~~Requires further testing~~
Co-authored-by: Michael Sproul <michael@sigmaprime.io>
## Proposed Changes
Enable multiple database backends for the slasher, either MDBX (default) or LMDB. The backend can be selected using `--slasher-backend={lmdb,mdbx}`.
## Additional Info
In order to abstract over the two library's different handling of database lifetimes I've used `Box::leak` to give the `Environment` type a `'static` lifetime. This was the only way I could think of using 100% safe code to construct a self-referential struct `SlasherDB`, where the `OpenDatabases` refers to the `Environment`. I think this is OK, as the `Environment` is expected to live for the life of the program, and both database engines leave the database in a consistent state after each write. The memory claimed for memory-mapping will be freed by the OS and appropriately flushed regardless of whether the `Environment` is actually dropped.
We are depending on two `sigp` forks of `libmdbx-rs` and `lmdb-rs`, to give us greater control over MDBX OS support and LMDB's version.
## Issue Addressed
I think we're running into this in our linkcheck, so I'm going to frist verify linkcheck fails on the current version, and then try downgrading it to see if it passes https://github.com/chronotope/chrono/issues/755
Co-authored-by: realbigsean <sean@sigmaprime.io>
