nimbus-eth1/tests/test_aristo/test_portal_proof.nim

261 lines
8.1 KiB
Nim
Raw Permalink Normal View History

# Nimbus
# 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.
{.used.}
import
std/[json, os, sets, strutils, tables],
eth/common,
stew/byteutils,
results,
unittest2,
../test_helpers,
../../nimbus/db/aristo,
../../nimbus/db/aristo/[aristo_desc, aristo_get, aristo_hike, aristo_layers,
aristo_part],
../../nimbus/db/aristo/aristo_part/part_debug
type
ProofData = ref object
chain: seq[seq[byte]]
missing: bool
error: AristoError
hike: Hike
# ------------------------------------------------------------------------------
# Private helper
# ------------------------------------------------------------------------------
proc createPartDb(ps: PartStateRef; data: seq[seq[byte]]; info: static[string]) =
# Set up production MPT
block:
let rc = ps.partPut(data, AutomaticPayload)
if rc.isErr: raiseAssert info & ": partPut => " & $rc.error
# Save keys to database
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
# TODO support for partial databases
# for (rvid,key) in ps.vkPairs:
# ps.db.layersPutKey(rvid, key)
# Make sure all is OK
block:
let rc = ps.check()
if rc.isErr: raiseAssert info & ": check => " & $rc.error
proc preLoadAristoDb(jKvp: JsonNode): PartStateRef =
const info = "preLoadAristoDb"
let ps = PartStateRef.init AristoDbRef.init()
# Collect rlp-encodede node blobs
var proof: seq[seq[byte]]
for (k,v) in jKvp.pairs:
let
key = hexToSeqByte(k)
val = hexToSeqByte(v.getStr())
if key.len == 32:
doAssert key == val.keccak256.data
if val != @[0x80u8]: # Exclude empty item
proof.add val
ps.createPartDb(proof, info)
ps
proc collectAddresses(node: JsonNode, collect: var HashSet[Address]) =
case node.kind:
of JObject:
for k,v in node.pairs:
if k == "address" and v.kind == JString:
collect.incl Address.fromHex v.getStr
else:
v.collectAddresses collect
of JArray:
for v in node.items:
v.collectAddresses collect
else:
discard
proc payloadAsBlob(pyl: LeafPayload; ps: PartStateRef): seq[byte] =
## Modified function `aristo_serialise.serialise()`.
##
const info = "payloadAsBlob"
case pyl.pType:
of AccountData:
let key = block:
if pyl.stoID.isValid:
Pre-allocate vids for branches (#2882) Each branch node may have up to 16 sub-items - currently, these are given VertexID based when they are first needed leading to a mostly-random order of vertexid for each subitem. Here, we pre-allocate all 16 vertex ids such that when a branch subitem is filled, it already has a vertexid waiting for it. This brings several important benefits: * subitems are sorted and "close" in their id sequencing - this means that when rocksdb stores them, they are likely to end up in the same data block thus improving read efficiency * because the ids are consequtive, we can store just the starting id and a bitmap representing which subitems are in use - this reduces disk space usage for branches allowing more of them fit into a single disk read, further improving disk read and caching performance - disk usage at block 18M is down from 84 to 78gb! * the in-memory footprint of VertexRef reduced allowing more instances to fit into caches and less memory to be used overall. Because of the increased locality of reference, it turns out that we no longer need to iterate over the entire database to efficiently generate the hash key database because the normal computation is now faster - this significantly benefits "live" chain processing as well where each dirtied key must be accompanied by a read of all branch subitems next to it - most of the performance benefit in this branch comes from this locality-of-reference improvement. On a sample resync, there's already ~20% improvement with later blocks seeing increasing benefit (because the trie is deeper in later blocks leading to more benefit from branch read perf improvements) ``` blocks: 18729664, baseline: 190h43m49s, contender: 153h59m0s Time (total): -36h44m48s, -19.27% ``` Note: clients need to be resynced as the PR changes the on-disk format R.I.P. little bloom filter - your life in the repo was short but valuable
2024-12-04 10:42:04 +00:00
let rc = ps.db.getKeyRc((VertexID(1),pyl.stoID.vid), {})
if rc.isErr:
raiseAssert info & ": getKey => " & $rc.error
Pre-allocate vids for branches (#2882) Each branch node may have up to 16 sub-items - currently, these are given VertexID based when they are first needed leading to a mostly-random order of vertexid for each subitem. Here, we pre-allocate all 16 vertex ids such that when a branch subitem is filled, it already has a vertexid waiting for it. This brings several important benefits: * subitems are sorted and "close" in their id sequencing - this means that when rocksdb stores them, they are likely to end up in the same data block thus improving read efficiency * because the ids are consequtive, we can store just the starting id and a bitmap representing which subitems are in use - this reduces disk space usage for branches allowing more of them fit into a single disk read, further improving disk read and caching performance - disk usage at block 18M is down from 84 to 78gb! * the in-memory footprint of VertexRef reduced allowing more instances to fit into caches and less memory to be used overall. Because of the increased locality of reference, it turns out that we no longer need to iterate over the entire database to efficiently generate the hash key database because the normal computation is now faster - this significantly benefits "live" chain processing as well where each dirtied key must be accompanied by a read of all branch subitems next to it - most of the performance benefit in this branch comes from this locality-of-reference improvement. On a sample resync, there's already ~20% improvement with later blocks seeing increasing benefit (because the trie is deeper in later blocks leading to more benefit from branch read perf improvements) ``` blocks: 18729664, baseline: 190h43m49s, contender: 153h59m0s Time (total): -36h44m48s, -19.27% ``` Note: clients need to be resynced as the PR changes the on-disk format R.I.P. little bloom filter - your life in the repo was short but valuable
2024-12-04 10:42:04 +00:00
rc.value[0][0]
else:
VOID_HASH_KEY
rlp.encode Account(
nonce: pyl.account.nonce,
balance: pyl.account.balance,
storageRoot: key.to(Hash32),
codeHash: pyl.account.codeHash)
of StoData:
rlp.encode pyl.stoData
func asExtension(b: seq[byte]; path: Hash32): seq[byte] =
var node = rlpFromBytes b
if node.listLen == 17:
let nibble = NibblesBuf.fromBytes(path.data)[0]
var wr = initRlpWriter()
wr.startList(2)
wr.append NibblesBuf.fromBytes(@[nibble]).slice(1).toHexPrefix(isleaf=false).data()
wr.append node.listElem(nibble.int).toBytes
wr.finish()
else:
b
when false:
# just keep for potential debugging
proc sq(s: string): string =
## For long strings print `begin..end` only
let n = (s.len + 1) div 2
result = if s.len < 20: s else: s[0 .. 5] & ".." & s[s.len-8 .. ^1]
result &= "[" & (if 0 < n: "#" & $n else: "") & "]"
# ------------------------------------------------------------------------------
# Private test functions
# ------------------------------------------------------------------------------
proc testCreatePortalProof(node: JsonNode, testStatusIMPL: var TestStatus) {.deprecated: "need to be rewritten to use non-generic data".} =
block: # TODO remove after rewrite
skip
return
const info = "testCreateProofTwig"
# Create partial database
let ps = node["state"].preLoadAristoDb()
# Collect addresses from json structure
var addresses: HashSet[Address]
node.collectAddresses addresses
# Convert addresses to valid paths (not all addresses might work)
var sample: Table[Hash32,ProofData]
for a in addresses:
let
path = a.data.keccak256
var hike: Hike
let rc = path.hikeUp(VertexID(1), ps.db, Opt.none(VertexRef), hike)
sample[path] = ProofData(
error: (if rc.isErr: rc.error[1] else: AristoError(0)),
hike: hike) # keep `hike` for potential debugging
# Verify that there is somehing to do, at all
check 0 < sample.values.toSeq.filterIt(it.error == AristoError 0).len
# Create proof chains
for (path,proof) in sample.pairs:
let rc = ps.db.partAccountTwig path
if proof.error == AristoError(0):
check rc.isOk and rc.value[1] == true
proof.chain = rc.value[0]
elif proof.error != HikeBranchMissingEdge:
# Note that this is a partial data base and in this case the proof for a
# non-existing entry might not work properly when the vertex is missing.
check rc.isOk and rc.value[1] == false
proof.chain = rc.value[0]
proof.missing = true
# Verify proof chains
for (path,proof) in sample.pairs:
if proof.missing:
# Proof for missing entries
let
rVid = proof.hike.root
root = ps.db.getKey((rVid,rVid)).to(Hash32)
block:
let rc = proof.chain.partUntwigPath(root, path)
check rc.isOk and rc.value.isNone
# Just for completeness (same a above combined into a single function)
check proof.chain.partUntwigPathOk(root, path, Opt.none seq[byte]).isOk
elif proof.error == AristoError 0:
let
rVid = proof.hike.root
pyl = proof.hike.legs[^1].wp.vtx.lData.payloadAsBlob(ps)
block:
# Use these root and chain
let chain = proof.chain
# Create another partial database from tree
let pq = PartStateRef.init AristoDbRef.init()
pq.createPartDb(chain, info)
# Create the same proof again which must result into the same as before
block:
let rc = pq.db.partAccountTwig path
if rc.isOk and rc.value[1] == true:
check rc.value[0] == proof.chain
# Verify proof
let root = pq.db.getKey((rVid,rVid)).to(Hash32)
block:
let rc = proof.chain.partUntwigPath(root, path)
check rc.isOk
if rc.isOk:
check rc.value == Opt.some(pyl)
# Just for completeness (same a above combined into a single function)
check proof.chain.partUntwigPathOk(root, path, Opt.some pyl).isOk
# Extension nodes are rare, so there is one created, inserted and the
# previous test repeated.
block:
let
ext = proof.chain[0].asExtension(path)
tail = @(proof.chain.toOpenArray(1,proof.chain.len-1))
chain = @[ext] & tail
# Create a third partial database from modified proof
let pq = PartStateRef.init AristoDbRef.init()
pq.createPartDb(chain, info)
# Re-create proof again
block:
let rc = pq.db.partAccountTwig path
check rc.isOk and rc.value[1] == true
if rc.isOk and rc.value[1] == true:
check rc.value[0] == chain
let root = pq.db.getKey((rVid,rVid)).to(Hash32)
block:
let rc = chain.partUntwigPath(root, path)
check rc.isOk
if rc.isOk:
check rc.value == Opt.some(pyl)
check chain.partUntwigPathOk(root, path, Opt.some pyl).isOk
# ------------------------------------------------------------------------------
# Test
# ------------------------------------------------------------------------------
suite "Encoding & verification of portal proof twigs for Aristo DB":
# Piggyback on tracer test suite environment
jsonTest("TracerTests", testCreatePortalProof)
# ------------------------------------------------------------------------------
# End
# ------------------------------------------------------------------------------