Hierarchical state diffs in hot DB (#6750)

This PR implements https://github.com/sigp/lighthouse/pull/5978 (tree-states) but on the hot DB. It allows Lighthouse to massively reduce its disk footprint during non-finality and overall I/O in all cases.

Closes https://github.com/sigp/lighthouse/issues/6580

Conga into https://github.com/sigp/lighthouse/pull/6744

### TODOs

- [x] Fix OOM in CI https://github.com/sigp/lighthouse/pull/7176
- [x] optimise store_hot_state to avoid storing a duplicate state if the summary already exists (should be safe from races now that pruning is cleaner)
- [x] mispelled: get_ancenstor_state_root
- [x] get_ancestor_state_root should use state summaries
- [x] Prevent split from changing during ancestor calc
- [x] Use same hierarchy for hot and cold

### TODO Good optimization for future PRs

- [ ] On the migration, if the latest hot snapshot is aligned with the cold snapshot migrate the diffs instead of the full states.
```
align slot  time
10485760    Nov-26-2024
12582912    Sep-14-2025
14680064    Jul-02-2026
```

### TODO Maybe things good to have

- [ ] Rename anchor_slot https://github.com/sigp/lighthouse/compare/tree-states-hot-rebase-oom...dapplion:lighthouse:tree-states-hot-anchor-slot-rename?expand=1
- [ ] Make anchor fields not public such that they must be mutated through a method. To prevent un-wanted changes of the anchor_slot

### NOTTODO

- [ ] Use fork-choice and a new method [`descendants_of_checkpoint`](ca2388e196 (diff-046fbdb517ca16b80e4464c2c824cf001a74a0a94ac0065e635768ac391062a8)) to filter only the state summaries that descend of finalized checkpoint]
This commit is contained in:
Lion - dapplion
2025-06-19 04:43:25 +02:00
committed by GitHub
parent 6786b9d12a
commit dd98534158
33 changed files with 2695 additions and 812 deletions

View File

@@ -1,102 +0,0 @@
use crate::*;
use ssz::{DecodeError, Encode};
use ssz_derive::Encode;
pub fn store_full_state<E: EthSpec>(
state_root: &Hash256,
state: &BeaconState<E>,
ops: &mut Vec<KeyValueStoreOp>,
) -> Result<(), Error> {
let bytes = {
let _overhead_timer = metrics::start_timer(&metrics::BEACON_STATE_WRITE_OVERHEAD_TIMES);
StorageContainer::new(state).as_ssz_bytes()
};
metrics::inc_counter_by(&metrics::BEACON_STATE_WRITE_BYTES, bytes.len() as u64);
metrics::inc_counter(&metrics::BEACON_STATE_WRITE_COUNT);
ops.push(KeyValueStoreOp::PutKeyValue(
DBColumn::BeaconState,
state_root.as_slice().to_vec(),
bytes,
));
Ok(())
}
pub fn get_full_state<KV: KeyValueStore<E>, E: EthSpec>(
db: &KV,
state_root: &Hash256,
spec: &ChainSpec,
) -> Result<Option<BeaconState<E>>, Error> {
let total_timer = metrics::start_timer(&metrics::BEACON_STATE_READ_TIMES);
match db.get_bytes(DBColumn::BeaconState, state_root.as_slice())? {
Some(bytes) => {
let overhead_timer = metrics::start_timer(&metrics::BEACON_STATE_READ_OVERHEAD_TIMES);
let container = StorageContainer::from_ssz_bytes(&bytes, spec)?;
metrics::stop_timer(overhead_timer);
metrics::stop_timer(total_timer);
metrics::inc_counter(&metrics::BEACON_STATE_READ_COUNT);
metrics::inc_counter_by(&metrics::BEACON_STATE_READ_BYTES, bytes.len() as u64);
Ok(Some(container.try_into()?))
}
None => Ok(None),
}
}
/// A container for storing `BeaconState` components.
// TODO: would be more space efficient with the caches stored separately and referenced by hash
#[derive(Encode)]
pub struct StorageContainer<E: EthSpec> {
state: BeaconState<E>,
committee_caches: Vec<Arc<CommitteeCache>>,
}
impl<E: EthSpec> StorageContainer<E> {
/// Create a new instance for storing a `BeaconState`.
pub fn new(state: &BeaconState<E>) -> Self {
Self {
state: state.clone(),
committee_caches: state.committee_caches().to_vec(),
}
}
pub fn from_ssz_bytes(bytes: &[u8], spec: &ChainSpec) -> Result<Self, ssz::DecodeError> {
// We need to use the slot-switching `from_ssz_bytes` of `BeaconState`, which doesn't
// compose with the other SSZ utils, so we duplicate some parts of `ssz_derive` here.
let mut builder = ssz::SszDecoderBuilder::new(bytes);
builder.register_anonymous_variable_length_item()?;
builder.register_type::<Vec<CommitteeCache>>()?;
let mut decoder = builder.build()?;
let state = decoder.decode_next_with(|bytes| BeaconState::from_ssz_bytes(bytes, spec))?;
let committee_caches = decoder.decode_next()?;
Ok(Self {
state,
committee_caches,
})
}
}
impl<E: EthSpec> TryInto<BeaconState<E>> for StorageContainer<E> {
type Error = Error;
fn try_into(mut self) -> Result<BeaconState<E>, Error> {
let mut state = self.state;
for i in (0..CACHED_EPOCHS).rev() {
if i >= self.committee_caches.len() {
return Err(Error::SszDecodeError(DecodeError::BytesInvalid(
"Insufficient committees for BeaconState".to_string(),
)));
};
state.committee_caches_mut()[i] = self.committee_caches.remove(i);
}
Ok(state)
}
}