nimbus-eth1/nimbus/db/aristo/aristo_delta.nim

94 lines
3.2 KiB
Nim
Raw Normal View History

# nimbus-eth1
# Copyright (c) 2023-2024 Status Research & Development GmbH
# Licensed under either of
# * Apache License, version 2.0, ([LICENSE-APACHE](LICENSE-APACHE) or
# http://www.apache.org/licenses/LICENSE-2.0)
# * MIT license ([LICENSE-MIT](LICENSE-MIT) or
# http://opensource.org/licenses/MIT)
# at your option. This file may not be copied, modified, or distributed
# except according to those terms.
## Aristo DB -- Delta filter management
## ====================================
##
import
std/tables,
eth/common,
results,
./aristo_desc/desc_backend,
"."/[aristo_desc]
# ------------------------------------------------------------------------------
# Public functions, save to backend
# ------------------------------------------------------------------------------
proc deltaPersistentOk*(db: AristoDbRef): bool =
## Check whether the read-only filter can be merged into the backend
not db.backend.isNil
proc deltaPersistent*(
db: AristoDbRef; # Database
nxtFid = 0u64; # Next filter ID (if any)
): Result[void,AristoError] =
## Resolve (i.e. move) the balancer into the physical backend database.
##
## This needs write permission on the backend DB for the descriptor argument
## `db` (see the function `aristo_desc.isCentre()`.) If the argument flag
## `reCentreOk` is passed `true`, write permission will be temporarily
## acquired when needed.
##
## When merging the current backend filter, its reverse will be is stored
## on other non-centre descriptors so there is no visible database change
## for these.
##
let be = db.backend
if be.isNil:
return err(FilBackendMissing)
# Blind or missing filter
if db.balancer.isNil:
# Add a blind storage frame. This will do no harm if `Aristo` runs
# standalone. Yet it is needed if a `Kvt` is tied to `Aristo` and has
# triggered a save cyle already which is to be completed here.
#
# There is no need to add a blind frame on any error return. If there
# is a `Kvt` tied to `Aristo`, then it must somehow run in sync and an
# error occuring here must have been detected earlier when (implicitely)
# registering `Kvt`. So that error should be considered a defect.
? be.putEndFn(? be.putBegFn())
return ok()
let lSst = SavedState(
key: EMPTY_ROOT_HASH, # placeholder for more
serial: nxtFid)
# Store structural single trie entries
let writeBatch = ? be.putBegFn()
for rvid, vtx in db.balancer.sTab:
Store keys together with node data (#2849) Currently, computed hash keys are stored in a separate column family with respect to the MPT data they're generated from - this has several disadvantages: * A lot of space is wasted because the lookup key (`RootedVertexID`) is repeated in both tables - this is 30% of the `AriKey` content! * rocksdb must maintain in-memory bloom filters and LRU caches for said keys, doubling its "minimal efficient cache size" * An extra disk traversal must be made to check for existence of cached hash key * Doubles the amount of files on disk due to each column family being its own set of files Here, the two CFs are joined such that both key and data is stored in `AriVtx`. This means: * we save ~30% disk space on repeated lookup keys * we save ~2gb of memory overhead that can be used to cache data instead of indices * we can skip storing hash keys for MPT leaf nodes - these are trivial to compute and waste a lot of space - previously they had to present in the `AriKey` CF to avoid having to look in two tables on the happy path. * There is a small increase in write amplification because when a hash value is updated for a branch node, we must write both key and branch data - previously we would write only the key * There's a small shift in CPU usage - instead of performing lookups in the database, hashes for leaf nodes are (re)-computed on the fly * We can return to slightly smaller on-disk SST files since there's fewer of them, which should reduce disk traffic a bit Internally, there are also other advantages: * when clearing keys, we no longer have to store a zero hash in memory - instead, we deduce staleness of the cached key from the presence of an updated VertexRef - this saves ~1gb of mem overhead during import * hash key cache becomes dedicated to branch keys since leaf keys are no longer stored in memory, reducing churn * key computation is a lot faster thanks to the skipped second disk traversal - a key computation for mainnet can be completed in 11 hours instead of ~2 days (!) thanks to better cache usage and less read amplification - with additional improvements to the on-disk format, we can probably get rid of the initial full traversal method of seeding the key cache on first start after import All in all, this PR reduces the size of a mainnet database from 160gb to 110gb and the peak memory footprint during import by ~1-2gb.
2024-11-20 08:56:27 +00:00
db.balancer.kMap.withValue(rvid, key) do:
be.putVtxFn(writeBatch, rvid, vtx, key[])
do:
be.putVtxFn(writeBatch, rvid, vtx, default(HashKey))
2024-06-04 15:05:13 +00:00
be.putTuvFn(writeBatch, db.balancer.vTop)
be.putLstFn(writeBatch, lSst)
? be.putEndFn writeBatch # Finalise write batch
# Copy back updated payloads
for accPath, vtx in db.balancer.accLeaves:
db.accLeaves.put(accPath, vtx)
for mixPath, vtx in db.balancer.stoLeaves:
db.stoLeaves.put(mixPath, vtx)
# Done with balancer, all saved to backend
db.balancer = LayerRef(nil)
ok()
# ------------------------------------------------------------------------------
# End
# ------------------------------------------------------------------------------