[Merged by Bors] - Prepare proposer #3043

paulhauner · 2022-02-27T01:41:44Z

Issue Addressed

Resolves #2936

Proposed Changes

Adds functionality for calling validator/prepare_beacon_proposer in advance.

There is a BeaconChain::prepare_beacon_proposer method which, which called, computes the proposer for the next slot. If that proposer has been registered via the validator/prepare_beacon_proposer API method, then the beacon_chain.execution_layer will be provided the PayloadAttributes for us in all future forkchoiceUpdated calls. An artificial forkchoiceUpdated call will be created 4s before each slot, when the head updates and when a validator updates their information.

Additionally, I added strict ordering for calls from the BeaconChain to the ExecutionLayer. I'm not certain the ExecutionLayer will always maintain this ordering, but it's a good start to have consistency from the BeaconChain. There are some deadlock opportunities introduced, they are documented in the code.

Additional Info

~~Blocked on [Merged by Bors] - Retrospective invalidation of exec. payloads for opt. sync #2837~~

realbigsean

I had a couple comments around simplifying this, but let me know if there are edge cases I am missing that are informing the design here.

Generally I was thinking the easiest/safest way to maintain BN <> EL consistency is to minimize when we call notify_forkchoice_updated and to only call it from fork_choice. We'd be relying on fork choice's ordering which seems like the correct thing to do to me but please correct me here otherwise.

The one place outside of fork choice where calling notify_forkchoice_updated seems unavoidable is when we fallback from get_payloadV1 due to a payload id missing from our cache (here). But calling a synthetic fork choice update here doesn't really seem like a good thing anyways... could we just run a call to chain.fork_choice() in this scenario instead?

If we're able to only call notify_forkchoice_updated from fork choice, I think we could then remove all the notify_forkchoice_updated specific locking that's added here as well.

realbigsean · 2022-03-02T15:33:14Z

beacon_node/beacon_chain/src/beacon_chain.rs

+            //
+            // This seems OK. It's not a significant waste of EL<>CL bandwidth or resources, as
+            // far as I know.
+            if let Err(e) = self.prepare_beacon_proposer_blocking() {


What's the benefit of running this after update_execution_engine_forkchoice_blocking? This seems intentional, but I am only really seeing the potential for duplicate notify_forkchoice_updated requests, which like you point out doesn't seem bad but I'm not seeing why it might be useful

So prepare_beacon_proposer_blocking will only send a fcU (forkchciceUpdated) message to the EL if something has changed or if the timing is right.

As it's currently written, we need to call prepare_beacon_proposer_blocking and notify_forkchoice_updated to ensure that at least one fcU call goes through to the EE.

We could probably avoid this by breaking out the fcU call from prepare_beacon_proposer_blocking it get it to return a "should call fcU bool". I'm not certain it's worth it, but I'll give it some thought.

Calling fcU before calculating the proposer in order to notify the EE as soon as possible seems reason enough to me to keep it as is 👌

realbigsean · 2022-03-02T15:37:54Z

beacon_node/beacon_chain/src/proposer_prep_service.rs

+    }
+}
+
+/// Loop indefinitely, calling `BeaconChain::prepare_beacon_proposer_async` at an interval.


It seems like this service covers the scenario where we have a skip slot right before we propose, and want to make sure we still notify the execution engine of our PayloadAttributes. But shouldn't that already be covered by the state_advance_timer which triggers at the same time and calls fork_choice which calls prepare_beacon_proposer_async?

Good thinking, but unfortunately the state_advance_timer only moves forward a single slot:

lighthouse/beacon_node/beacon_chain/src/state_advance_timer.rs

Lines 230 to 237 in c3a793f

// Protect against advancing a state more than a single slot.

//

// Advancing more than one slot without storing the intermediate state would corrupt the

// database. Future works might store temporary, intermediate states inside this function.

return Err(Error::BadStateSlot {

_block_slot: head_slot,

_state_slot: state.slot(),

});

So after one skip slot and following slots won't be advanced.

Oh gotcha, didn't realize this

realbigsean · 2022-03-02T16:08:34Z

beacon_node/http_api/src/lib.rs

@@ -2229,6 +2229,13 @@ pub fn serve<T: BeaconChainTypes>(
                            )
                        })?;

+                    chain.prepare_beacon_proposer_blocking().map_err(|e| {


I sort of think this endpoint should only handle updating the proposer preparation cache, and we should make sure fork choice is the only place we call prepare_beacon_proposer_async from.

I could see the benefit of this potentially allowing last second suggested fee recipients changes. But I'm not sure that's something we should try to code for. To me it seems like the later this information is provided the less profitable the block could be and the more likely it is to just be missed completely, so there's already strong incentives to provide the fee recipient early and not to change it frequently.

If we are worried about not being able to get the fee recipient on startup for validators proposing immediately, we could also persist this cache or allow the beacon chain to read from the suggested fee recipients file.

paulhauner · 2022-03-02T19:45:00Z

Generally I was thinking the easiest/safest way to maintain BN <> EL consistency is to minimize when we call notify_forkchoice_updated and to only call it from fork_choice.

Even if we do call it inside fork choice, we still need to deal with ordering. Even if the call to fcU is inside the fork_choice_internal function, it still doesn't mean the calls are ordered. Something can call fork_choice_internal multiple times, unless we've prevented concurrency internally via some lock. We could use the main chain.fork_choice lock, however I'm hesitant to hold that lock whilst we call out to remote servers. That is likely to cause long-lived contention.

I'm open to just returning a bool from prepare_beacon_proposer_blocking and callers can use that value to call fork choice, if they want. However, I think there's still downsides to that. For example, during fork_choice_internal, it seems valuable to me to call fcU as quickly as possible, then compute the proposers, then maybe call it again. Some ELs might avoid syncing non-head-chains until we provide them a fcU message saying it's the head. In those cases, we want the EL to be aware of the head ASAP so they can start syncing it. In the case where the shuffling isn't in the cache, it seems ideal to send 2x fcU messages; one that alerts about the head change and a later one that informs about the proposer.

realbigsean · 2022-03-02T21:08:16Z

Even if we do call it inside fork choice, we still need to deal with ordering.

Yes I was misunderstanding this. I was looking for a way around the mutex lock on fcU because it'll be called from a lot of different places, but I guess it is the only way to guarantee order.

paulhauner · 2022-03-03T23:17:06Z

It turns out there's a bit of an edge-case with the transition block that isn't covered here. I've documented it in #3058.

I think we should merge this PR as-is since it's a significant improvement and has no issues before/after the transition block. We can come back for #3058.

Squashed commit of the following: commit c03670e Author: Paul Hauner <[email protected]> Date: Fri Mar 4 09:41:11 2022 +1100 Filter out zero-hashes commit 6d269dd Author: Paul Hauner <[email protected]> Date: Fri Mar 4 09:09:28 2022 +1100 Avoid sending fcU with zero hash

paulhauner · 2022-03-07T23:59:47Z

Alrighty, I've been testing this on merge-devnet-5 with our 1k validators. It's looking good there so I'm happy to merge.

There are two changes since last review:

Add in the changes from Avoid sending a zero-hash head_block_hash to EEs #3057
Add some metrics.

You can see these changes here: paulhauner/lighthouse@76a1d1d...d029c7b

If I can get an approval for these changes then I'll merge :)

paulhauner · 2022-03-08T00:48:47Z

This is failing CI due to Geth changing some CLI flags. See #3046 (comment)

paulhauner · 2022-03-08T03:58:21Z

I've also added 4811406 after observing a panic:

thread 'tokio-runtime-worker' panicked at 'Cannot start a runtime from within a runtime. This happens because a function (like `block_on`) attempted to block the current thread while the thread is being used to drive asynchronous tasks.'

That commit has approval from @AgeManning.

Squashed commit of the following: commit 4811406 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 14:39:58 2022 +1100 Use spawn_blocking for sync fork choice call commit b0e0829 Merge: d029c7b 4186d11 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 10:53:54 2022 +1100 Merge branch 'unstable' into prepare-proposer commit d029c7b Author: Paul Hauner <[email protected]> Date: Tue Mar 8 10:40:40 2022 +1100 Add metrics commit 70e3b82 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 09:45:47 2022 +1100 Squash merge of #3057 Squashed commit of the following: commit c03670e Author: Paul Hauner <[email protected]> Date: Fri Mar 4 09:41:11 2022 +1100 Filter out zero-hashes commit 6d269dd Author: Paul Hauner <[email protected]> Date: Fri Mar 4 09:09:28 2022 +1100 Avoid sending fcU with zero hash commit 76a1d1d Author: realbigsean <[email protected]> Date: Thu Mar 3 14:17:54 2022 -0500 Update beacon_node/beacon_chain/src/beacon_chain.rs commit 4a7e3a7 Author: realbigsean <[email protected]> Date: Thu Mar 3 14:17:35 2022 -0500 Update beacon_node/beacon_chain/tests/payload_invalidation.rs commit c651994 Merge: bb1f7d0 aea43b6 Author: realbigsean <[email protected]> Date: Thu Mar 3 14:05:26 2022 -0500 Merge branch 'unstable' into prepare-proposer commit bb1f7d0 Author: Paul Hauner <[email protected]> Date: Thu Mar 3 11:07:24 2022 +1100 Update beacon_node/execution_layer/src/lib.rs Co-authored-by: realbigsean <[email protected]> commit 68e4648 Author: Paul Hauner <[email protected]> Date: Thu Mar 3 11:04:48 2022 +1100 Prevent running during sync commit 5ddeae8 Author: Paul Hauner <[email protected]> Date: Thu Mar 3 10:57:28 2022 +1100 Spawn proposer prepare routine instead of waiting commit 2306b09 Author: Paul Hauner <[email protected]> Date: Wed Mar 2 19:01:06 2022 +1100 Fix failing test commit 8eca77a Author: Paul Hauner <[email protected]> Date: Wed Mar 2 18:54:26 2022 +1100 Fix bugs commit 15d6b71 Author: Paul Hauner <[email protected]> Date: Wed Mar 2 13:03:40 2022 +1100 Read proposer during preparation commit b1ac426 Author: Paul Hauner <[email protected]> Date: Tue Mar 1 20:17:34 2022 +1100 Add more tests commit 357ff8b Author: Paul Hauner <[email protected]> Date: Tue Mar 1 16:41:36 2022 +1100 Save prev request in exec layer commit e19fa44 Author: Paul Hauner <[email protected]> Date: Tue Mar 1 16:41:17 2022 +1100 Upgrade error to dbg commit d2d0099 Author: Paul Hauner <[email protected]> Date: Mon Feb 28 16:29:02 2022 +1100 Add forkchoice lock commit 28b94cb Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:30:55 2022 +1100 Tidy, add comments commit b24a3ea Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:09:20 2022 +1100 Don't start update service if there's no EL commit 3f56ed0 Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:08:52 2022 +1100 Reduce nesting commit 61c50b7 Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:02:42 2022 +1100 Call prepare_proposer after head change commit 8fb610c Author: Paul Hauner <[email protected]> Date: Mon Feb 28 14:44:51 2022 +1100 Tidy commit 7e008cb Author: Paul Hauner <[email protected]> Date: Sun Feb 27 12:39:29 2022 +1100 Add prepare call to HTTP api commit 44d090e Author: Paul Hauner <[email protected]> Date: Sun Feb 27 12:36:32 2022 +1100 Add proposer prep service commit 8193fae Author: Paul Hauner <[email protected]> Date: Sun Feb 27 12:12:45 2022 +1100 Add fc call to preparation endpoint commit 2ad4e18 Author: Paul Hauner <[email protected]> Date: Sun Feb 27 11:56:45 2022 +1100 Repair after rebase commit f1e6056 Author: Paul Hauner <[email protected]> Date: Sun Feb 27 11:41:34 2022 +1100 Start adding payload prep delay commit 0f72e34 Author: Paul Hauner <[email protected]> Date: Sun Feb 27 11:31:33 2022 +1100 Add proposer prep to beacon chain commit 10d545f Author: Paul Hauner <[email protected]> Date: Sun Feb 27 10:42:08 2022 +1100 Add `proposers` cache to EL

paulhauner · 2022-03-08T06:35:49Z

Sorry for the noise here. As you can see, I found some bugs whilst testing with our validators on merge-devnet-5.

I think this will require additional review, sorry @realbigsean. Unfortunately the diff between the previously-approved changes and the current commit includes a merge of unstable (I had to resolve a conflict, sorry).

We should be good to go after this round of review, this branch is performing well presently 🚀

…epare-proposer � Conflicts: � Cargo.lock � beacon_node/execution_layer/Cargo.toml

realbigsean

New changes look good to me, left a couple nits. Resolved a merge conflict in Cargo.lock and a Cargo.toml as well to try to get CI running

beacon_node/beacon_chain/tests/payload_invalidation.rs

beacon_node/execution_layer/src/metrics.rs

Co-authored-by: realbigsean <[email protected]>

paulhauner · 2022-03-08T22:41:45Z

Thanks for the review and CI fix @realbigsean 🙏

I'll see how this goes on CI and the testnets and merge if all goes well.

Squashed commit of the following: commit f0153f1 Author: Paul Hauner <[email protected]> Date: Wed Mar 9 09:39:48 2022 +1100 Apply suggestions from code review Co-authored-by: realbigsean <[email protected]> commit b6b6a34 Author: realbigsean <[email protected]> Date: Tue Mar 8 08:56:23 2022 -0500 Cargo.lock commit e270c61 Merge: c86ada0 381d0ec Author: realbigsean <[email protected]> Date: Tue Mar 8 08:55:57 2022 -0500 Merge branch 'unstable' of https://github.com/sigp/lighthouse into prepare-proposer � Conflicts: � Cargo.lock � beacon_node/execution_layer/Cargo.toml commit c86ada0 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 16:58:39 2022 +1100 Add payload id hit/miss metrics commit 4811406 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 14:39:58 2022 +1100 Use spawn_blocking for sync fork choice call commit b0e0829 Merge: d029c7b 4186d11 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 10:53:54 2022 +1100 Merge branch 'unstable' into prepare-proposer commit d029c7b Author: Paul Hauner <[email protected]> Date: Tue Mar 8 10:40:40 2022 +1100 Add metrics commit 70e3b82 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 09:45:47 2022 +1100 Squash merge of #3057 Squashed commit of the following: commit c03670e Author: Paul Hauner <[email protected]> Date: Fri Mar 4 09:41:11 2022 +1100 Filter out zero-hashes commit 6d269dd Author: Paul Hauner <[email protected]> Date: Fri Mar 4 09:09:28 2022 +1100 Avoid sending fcU with zero hash commit 76a1d1d Author: realbigsean <[email protected]> Date: Thu Mar 3 14:17:54 2022 -0500 Update beacon_node/beacon_chain/src/beacon_chain.rs commit 4a7e3a7 Author: realbigsean <[email protected]> Date: Thu Mar 3 14:17:35 2022 -0500 Update beacon_node/beacon_chain/tests/payload_invalidation.rs commit c651994 Merge: bb1f7d0 aea43b6 Author: realbigsean <[email protected]> Date: Thu Mar 3 14:05:26 2022 -0500 Merge branch 'unstable' into prepare-proposer commit bb1f7d0 Author: Paul Hauner <[email protected]> Date: Thu Mar 3 11:07:24 2022 +1100 Update beacon_node/execution_layer/src/lib.rs Co-authored-by: realbigsean <[email protected]> commit 68e4648 Author: Paul Hauner <[email protected]> Date: Thu Mar 3 11:04:48 2022 +1100 Prevent running during sync commit 5ddeae8 Author: Paul Hauner <[email protected]> Date: Thu Mar 3 10:57:28 2022 +1100 Spawn proposer prepare routine instead of waiting commit 2306b09 Author: Paul Hauner <[email protected]> Date: Wed Mar 2 19:01:06 2022 +1100 Fix failing test commit 8eca77a Author: Paul Hauner <[email protected]> Date: Wed Mar 2 18:54:26 2022 +1100 Fix bugs commit 15d6b71 Author: Paul Hauner <[email protected]> Date: Wed Mar 2 13:03:40 2022 +1100 Read proposer during preparation commit b1ac426 Author: Paul Hauner <[email protected]> Date: Tue Mar 1 20:17:34 2022 +1100 Add more tests commit 357ff8b Author: Paul Hauner <[email protected]> Date: Tue Mar 1 16:41:36 2022 +1100 Save prev request in exec layer commit e19fa44 Author: Paul Hauner <[email protected]> Date: Tue Mar 1 16:41:17 2022 +1100 Upgrade error to dbg commit d2d0099 Author: Paul Hauner <[email protected]> Date: Mon Feb 28 16:29:02 2022 +1100 Add forkchoice lock commit 28b94cb Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:30:55 2022 +1100 Tidy, add comments commit b24a3ea Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:09:20 2022 +1100 Don't start update service if there's no EL commit 3f56ed0 Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:08:52 2022 +1100 Reduce nesting commit 61c50b7 Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:02:42 2022 +1100 Call prepare_proposer after head change commit 8fb610c Author: Paul Hauner <[email protected]> Date: Mon Feb 28 14:44:51 2022 +1100 Tidy commit 7e008cb Author: Paul Hauner <[email protected]> Date: Sun Feb 27 12:39:29 2022 +1100 Add prepare call to HTTP api commit 44d090e Author: Paul Hauner <[email protected]> Date: Sun Feb 27 12:36:32 2022 +1100 Add proposer prep service commit 8193fae Author: Paul Hauner <[email protected]> Date: Sun Feb 27 12:12:45 2022 +1100 Add fc call to preparation endpoint commit 2ad4e18 Author: Paul Hauner <[email protected]> Date: Sun Feb 27 11:56:45 2022 +1100 Repair after rebase commit f1e6056 Author: Paul Hauner <[email protected]> Date: Sun Feb 27 11:41:34 2022 +1100 Start adding payload prep delay commit 0f72e34 Author: Paul Hauner <[email protected]> Date: Sun Feb 27 11:31:33 2022 +1100 Add proposer prep to beacon chain commit 10d545f Author: Paul Hauner <[email protected]> Date: Sun Feb 27 10:42:08 2022 +1100 Add `proposers` cache to EL

Squashed commit of the following: commit f0153f1 Author: Paul Hauner <[email protected]> Date: Wed Mar 9 09:39:48 2022 +1100 Apply suggestions from code review Co-authored-by: realbigsean <[email protected]> commit b6b6a34 Author: realbigsean <[email protected]> Date: Tue Mar 8 08:56:23 2022 -0500 Cargo.lock commit e270c61 Merge: c86ada0 381d0ec Author: realbigsean <[email protected]> Date: Tue Mar 8 08:55:57 2022 -0500 Merge branch 'unstable' of https://github.com/sigp/lighthouse into prepare-proposer � Conflicts: � Cargo.lock � beacon_node/execution_layer/Cargo.toml commit c86ada0 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 16:58:39 2022 +1100 Add payload id hit/miss metrics commit 4811406 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 14:39:58 2022 +1100 Use spawn_blocking for sync fork choice call commit b0e0829 Merge: d029c7b 4186d11 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 10:53:54 2022 +1100 Merge branch 'unstable' into prepare-proposer commit d029c7b Author: Paul Hauner <[email protected]> Date: Tue Mar 8 10:40:40 2022 +1100 Add metrics commit 70e3b82 Author: Paul Hauner <[email protected]> Date: Tue Mar 8 09:45:47 2022 +1100 Squash merge of sigp#3057 Squashed commit of the following: commit c03670e Author: Paul Hauner <[email protected]> Date: Fri Mar 4 09:41:11 2022 +1100 Filter out zero-hashes commit 6d269dd Author: Paul Hauner <[email protected]> Date: Fri Mar 4 09:09:28 2022 +1100 Avoid sending fcU with zero hash commit 76a1d1d Author: realbigsean <[email protected]> Date: Thu Mar 3 14:17:54 2022 -0500 Update beacon_node/beacon_chain/src/beacon_chain.rs commit 4a7e3a7 Author: realbigsean <[email protected]> Date: Thu Mar 3 14:17:35 2022 -0500 Update beacon_node/beacon_chain/tests/payload_invalidation.rs commit c651994 Merge: bb1f7d0 aea43b6 Author: realbigsean <[email protected]> Date: Thu Mar 3 14:05:26 2022 -0500 Merge branch 'unstable' into prepare-proposer commit bb1f7d0 Author: Paul Hauner <[email protected]> Date: Thu Mar 3 11:07:24 2022 +1100 Update beacon_node/execution_layer/src/lib.rs Co-authored-by: realbigsean <[email protected]> commit 68e4648 Author: Paul Hauner <[email protected]> Date: Thu Mar 3 11:04:48 2022 +1100 Prevent running during sync commit 5ddeae8 Author: Paul Hauner <[email protected]> Date: Thu Mar 3 10:57:28 2022 +1100 Spawn proposer prepare routine instead of waiting commit 2306b09 Author: Paul Hauner <[email protected]> Date: Wed Mar 2 19:01:06 2022 +1100 Fix failing test commit 8eca77a Author: Paul Hauner <[email protected]> Date: Wed Mar 2 18:54:26 2022 +1100 Fix bugs commit 15d6b71 Author: Paul Hauner <[email protected]> Date: Wed Mar 2 13:03:40 2022 +1100 Read proposer during preparation commit b1ac426 Author: Paul Hauner <[email protected]> Date: Tue Mar 1 20:17:34 2022 +1100 Add more tests commit 357ff8b Author: Paul Hauner <[email protected]> Date: Tue Mar 1 16:41:36 2022 +1100 Save prev request in exec layer commit e19fa44 Author: Paul Hauner <[email protected]> Date: Tue Mar 1 16:41:17 2022 +1100 Upgrade error to dbg commit d2d0099 Author: Paul Hauner <[email protected]> Date: Mon Feb 28 16:29:02 2022 +1100 Add forkchoice lock commit 28b94cb Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:30:55 2022 +1100 Tidy, add comments commit b24a3ea Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:09:20 2022 +1100 Don't start update service if there's no EL commit 3f56ed0 Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:08:52 2022 +1100 Reduce nesting commit 61c50b7 Author: Paul Hauner <[email protected]> Date: Mon Feb 28 15:02:42 2022 +1100 Call prepare_proposer after head change commit 8fb610c Author: Paul Hauner <[email protected]> Date: Mon Feb 28 14:44:51 2022 +1100 Tidy commit 7e008cb Author: Paul Hauner <[email protected]> Date: Sun Feb 27 12:39:29 2022 +1100 Add prepare call to HTTP api commit 44d090e Author: Paul Hauner <[email protected]> Date: Sun Feb 27 12:36:32 2022 +1100 Add proposer prep service commit 8193fae Author: Paul Hauner <[email protected]> Date: Sun Feb 27 12:12:45 2022 +1100 Add fc call to preparation endpoint commit 2ad4e18 Author: Paul Hauner <[email protected]> Date: Sun Feb 27 11:56:45 2022 +1100 Repair after rebase commit f1e6056 Author: Paul Hauner <[email protected]> Date: Sun Feb 27 11:41:34 2022 +1100 Start adding payload prep delay commit 0f72e34 Author: Paul Hauner <[email protected]> Date: Sun Feb 27 11:31:33 2022 +1100 Add proposer prep to beacon chain commit 10d545f Author: Paul Hauner <[email protected]> Date: Sun Feb 27 10:42:08 2022 +1100 Add `proposers` cache to EL

paulhauner · 2022-03-09T00:41:49Z

Looks like windows tests are just being dodgy again :(

Everything is looking good on the testnet nodes 🎉

bors r+

## Issue Addressed Resolves #2936 ## Proposed Changes Adds functionality for calling [`validator/prepare_beacon_proposer`](https://ethereum.github.io/beacon-APIs/?urls.primaryName=dev#/Validator/prepareBeaconProposer) in advance. There is a `BeaconChain::prepare_beacon_proposer` method which, which called, computes the proposer for the next slot. If that proposer has been registered via the `validator/prepare_beacon_proposer` API method, then the `beacon_chain.execution_layer` will be provided the `PayloadAttributes` for us in all future forkchoiceUpdated calls. An artificial forkchoiceUpdated call will be created 4s before each slot, when the head updates and when a validator updates their information. Additionally, I added strict ordering for calls from the `BeaconChain` to the `ExecutionLayer`. I'm not certain the `ExecutionLayer` will always maintain this ordering, but it's a good start to have consistency from the `BeaconChain`. There are some deadlock opportunities introduced, they are documented in the code. ## Additional Info - ~~Blocked on #2837~~ Co-authored-by: realbigsean <[email protected]>

bors · 2022-03-09T02:32:44Z

Pull request successfully merged into unstable.

Build succeeded:

## Issue Addressed This address an issue which was preventing checkpoint-sync. When the node starts from checkpoint sync, the head block and the finalized block are the same value. We did not respect this when sending a `forkchoiceUpdated` (fcU) call to the EL and were expecting fork choice to hold the *finalized ancestor of the head* and returning an error when it didn't. This PR uses *only fork choice* for sending fcU updates. This is actually quite nice and avoids some atomicity issues between `chain.canonical_head` and `chain.fork_choice`. Now, whenever `chain.fork_choice.get_head` returns a value we also cache the values required for the next fcU call. ## TODO - [x] ~~Blocked on #3043~~ - [x] Ensure there isn't a warn message at startup.

paulhauner added blocked work-in-progress PR is a work-in-progress bellatrix Required to support the Bellatrix Upgrade labels Feb 27, 2022

paulhauner mentioned this pull request Feb 27, 2022

Update to kiln spec v2 #3037

Closed

paulhauner force-pushed the prepare-proposer branch from 1ef1b14 to cd0bac7 Compare February 28, 2022 03:24

paulhauner self-assigned this Feb 28, 2022

paulhauner added 13 commits March 1, 2022 13:24

Add proposers cache to EL

10d545f

Add proposer prep to beacon chain

0f72e34

Start adding payload prep delay

f1e6056

Repair after rebase

2ad4e18

Add fc call to preparation endpoint

8193fae

Add proposer prep service

44d090e

Add prepare call to HTTP api

7e008cb

Tidy

8fb610c

Call prepare_proposer after head change

61c50b7

Reduce nesting

3f56ed0

Don't start update service if there's no EL

b24a3ea

Tidy, add comments

28b94cb

Add forkchoice lock

d2d0099

paulhauner force-pushed the prepare-proposer branch from f851f4b to d2d0099 Compare March 1, 2022 02:34

paulhauner removed the blocked label Mar 1, 2022

paulhauner added 6 commits March 1, 2022 16:41

Upgrade error to dbg

e19fa44

Save prev request in exec layer

357ff8b

Add more tests

b1ac426

Read proposer during preparation

15d6b71

Fix bugs

8eca77a

Fix failing test

2306b09

realbigsean reviewed Mar 2, 2022

View reviewed changes

paulhauner mentioned this pull request Mar 7, 2022

Avoid sending a zero-hash head_block_hash to EEs #3057

Closed

paulhauner added 2 commits March 8, 2022 10:46

Add metrics

d029c7b

Merge branch 'unstable' into prepare-proposer

b0e0829

Use spawn_blocking for sync fork choice call

4811406

paulhauner mentioned this pull request Mar 8, 2022

Testnet: merge-devnet-5 #3054

Closed

Add payload id hit/miss metrics

c86ada0

realbigsean added 2 commits March 8, 2022 08:55

Merge branch 'unstable' of https://github.com/sigp/lighthouse into pr…

e270c61

…epare-proposer � Conflicts: � Cargo.lock � beacon_node/execution_layer/Cargo.toml

Cargo.lock

b6b6a34

realbigsean approved these changes Mar 8, 2022

View reviewed changes

beacon_node/beacon_chain/tests/payload_invalidation.rs Outdated Show resolved Hide resolved

beacon_node/execution_layer/src/metrics.rs Outdated Show resolved Hide resolved

Apply suggestions from code review

f0153f1

Co-authored-by: realbigsean <[email protected]>

paulhauner added ready-for-merge This PR is ready to merge. and removed ready-for-review The code is ready for review labels Mar 9, 2022

paulhauner mentioned this pull request Mar 9, 2022

[Merged by Bors] - Fix post-merge checkpoint sync #3065

Closed

2 tasks

bors bot changed the title ~~Prepare proposer~~ [Merged by Bors] - Prepare proposer Mar 9, 2022

bors bot closed this Mar 9, 2022

paulhauner mentioned this pull request Mar 9, 2022

Provide payloadAttributes for block proposers in advance #2936

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[Merged by Bors] - Prepare proposer #3043

[Merged by Bors] - Prepare proposer #3043

paulhauner commented Feb 27, 2022 •

edited

Loading

realbigsean left a comment

realbigsean Mar 2, 2022 •

edited

Loading

paulhauner Mar 2, 2022

realbigsean Mar 2, 2022

realbigsean Mar 2, 2022 •

edited

Loading

paulhauner Mar 2, 2022

realbigsean Mar 2, 2022

realbigsean Mar 2, 2022

paulhauner commented Mar 2, 2022 •

edited

Loading

realbigsean commented Mar 2, 2022

paulhauner commented Mar 3, 2022

paulhauner commented Mar 7, 2022

paulhauner commented Mar 8, 2022

paulhauner commented Mar 8, 2022

paulhauner commented Mar 8, 2022 •

edited

Loading

realbigsean left a comment •

edited

Loading

paulhauner commented Mar 8, 2022

paulhauner commented Mar 9, 2022

bors bot commented Mar 9, 2022

	// Protect against advancing a state more than a single slot.
	//
	// Advancing more than one slot without storing the intermediate state would corrupt the
	// database. Future works might store temporary, intermediate states inside this function.
	return Err(Error::BadStateSlot {
	_block_slot: head_slot,
	_state_slot: state.slot(),
	});

[Merged by Bors] - Prepare proposer #3043

[Merged by Bors] - Prepare proposer #3043

Conversation

paulhauner commented Feb 27, 2022 • edited Loading

Issue Addressed

Proposed Changes

Additional Info

realbigsean left a comment

Choose a reason for hiding this comment

realbigsean Mar 2, 2022 • edited Loading

Choose a reason for hiding this comment

paulhauner Mar 2, 2022

Choose a reason for hiding this comment

realbigsean Mar 2, 2022

Choose a reason for hiding this comment

realbigsean Mar 2, 2022 • edited Loading

Choose a reason for hiding this comment

paulhauner Mar 2, 2022

Choose a reason for hiding this comment

realbigsean Mar 2, 2022

Choose a reason for hiding this comment

realbigsean Mar 2, 2022

Choose a reason for hiding this comment

paulhauner commented Mar 2, 2022 • edited Loading

realbigsean commented Mar 2, 2022

paulhauner commented Mar 3, 2022

paulhauner commented Mar 7, 2022

paulhauner commented Mar 8, 2022

paulhauner commented Mar 8, 2022

paulhauner commented Mar 8, 2022 • edited Loading

realbigsean left a comment • edited Loading

Choose a reason for hiding this comment

paulhauner commented Mar 8, 2022

paulhauner commented Mar 9, 2022

bors bot commented Mar 9, 2022

paulhauner commented Feb 27, 2022 •

edited

Loading

realbigsean Mar 2, 2022 •

edited

Loading

realbigsean Mar 2, 2022 •

edited

Loading

paulhauner commented Mar 2, 2022 •

edited

Loading

paulhauner commented Mar 8, 2022 •

edited

Loading

realbigsean left a comment •

edited

Loading