nimbus-eth2/tests/test_beacon_chain_db.nim

358 lines
11 KiB
Nim
Raw Normal View History

# Nimbus
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
# Copyright (c) 2018-2021 Status Research & Development GmbH
# Licensed under either of
# * Apache License, version 2.0, ([LICENSE-APACHE](LICENSE-APACHE) or https://www.apache.org/licenses/LICENSE-2.0)
# * MIT license ([LICENSE-MIT](LICENSE-MIT) or https://opensource.org/licenses/MIT)
# at your option. This file may not be copied, modified, or distributed except according to those terms.
{.used.}
import
std/[algorithm, options, sequtils],
unittest2,
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
../beacon_chain/[beacon_chain_db, extras, interop, ssz],
../beacon_chain/spec/[
beaconstate, crypto, digest, forkedbeaconstate_helpers, presets,
state_transition],
../beacon_chain/spec/datatypes/[phase0, altair],
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
../beacon_chain/consensus_object_pools/blockchain_dag,
2020-04-27 16:36:28 +00:00
eth/db/kvstore,
# test utilies
./testutil, ./testdbutil, ./testblockutil, ./teststateutil
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
when isMainModule:
import chronicles # or some random compile error happens...
proc getPhase0StateRef(db: BeaconChainDB, root: Eth2Digest):
phase0.NilableBeaconStateRef =
# load beaconstate the way the block pool does it - into an existing instance
let res = (phase0.BeaconStateRef)()
if db.getState(root, res[], noRollback):
return res
proc getAltairStateRef(db: BeaconChainDB, root: Eth2Digest):
altair.NilableBeaconStateRef =
# load beaconstate the way the block pool does it - into an existing instance
let res = (altair.BeaconStateRef)()
if db.getAltairState(root, res[], noRollback):
return res
func withDigest(blck: phase0.TrustedBeaconBlock):
phase0.TrustedSignedBeaconBlock =
phase0.TrustedSignedBeaconBlock(
message: blck,
root: hash_tree_root(blck)
)
func withDigest(blck: altair.TrustedBeaconBlock):
altair.TrustedSignedBeaconBlock =
altair.TrustedSignedBeaconBlock(
message: blck,
root: hash_tree_root(blck)
)
suite "Beacon chain DB" & preset():
test "empty database" & preset():
var
db = BeaconChainDB.new(defaultRuntimePreset, "", inMemory = true)
check:
db.getPhase0StateRef(Eth2Digest()).isNil
db.getBlock(Eth2Digest()).isNone
test "sanity check phase 0 blocks" & preset():
var db = BeaconChainDB.new(defaultRuntimePreset, "", inMemory = true)
let
signedBlock = withDigest((phase0.TrustedBeaconBlock)())
root = hash_tree_root(signedBlock.message)
db.putBlock(signedBlock)
check:
db.containsBlock(root)
db.getBlock(root).get() == signedBlock
db.delBlock(root)
check:
not db.containsBlock(root)
db.getBlock(root).isErr()
db.putStateRoot(root, signedBlock.message.slot, root)
var root2 = root
root2.data[0] = root.data[0] + 1
db.putStateRoot(root, signedBlock.message.slot + 1, root2)
check:
db.getStateRoot(root, signedBlock.message.slot).get() == root
db.getStateRoot(root, signedBlock.message.slot + 1).get() == root2
db.close()
test "sanity check Altair blocks" & preset():
var db = BeaconChainDB.new(defaultRuntimePreset, "", inMemory = true)
let
signedBlock = withDigest((altair.TrustedBeaconBlock)())
root = hash_tree_root(signedBlock.message)
db.putBlock(signedBlock)
check:
db.containsBlock(root)
db.getAltairBlock(root).get() == signedBlock
db.delBlock(root)
check:
not db.containsBlock(root)
db.getAltairBlock(root).isErr()
db.putStateRoot(root, signedBlock.message.slot, root)
2020-08-07 20:17:24 +00:00
var root2 = root
root2.data[0] = root.data[0] + 1
db.putStateRoot(root, signedBlock.message.slot + 1, root2)
check:
db.getStateRoot(root, signedBlock.message.slot).get() == root
2020-08-07 20:17:24 +00:00
db.getStateRoot(root, signedBlock.message.slot + 1).get() == root2
db.close()
test "sanity check phase 0 states" & preset():
var
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
db = makeTestDB(SLOTS_PER_EPOCH)
dag = init(ChainDAGRef, defaultRuntimePreset, db)
testStates = getTestStates(dag.headState.data)
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
# Ensure transitions beyond just adding validators and increasing slots
sort(testStates) do (x, y: ref ForkedHashedBeaconState) -> int:
cmp($getStateRoot(x[]), $getStateRoot(y[]))
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
for state in testStates:
db.putState(state[].hbsPhase0.data)
let root = hash_tree_root(state[])
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
check:
db.containsState(root)
hash_tree_root(db.getPhase0StateRef(root)[]) == root
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
db.delState(root)
check:
not db.containsState(root)
db.getPhase0StateRef(root).isNil
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
db.close()
test "sanity check Altair states" & preset():
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
var
db = makeTestDB(SLOTS_PER_EPOCH)
dag = init(ChainDAGRef, defaultRuntimePreset, db)
testStates = getTestStates(dag.headState.data, true)
# Ensure transitions beyond just adding validators and increasing slots
sort(testStates) do (x, y: ref ForkedHashedBeaconState) -> int:
cmp($getStateRoot(x[]), $getStateRoot(y[]))
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
for state in testStates:
db.putState(state[].hbsAltair.data)
let root = hash_tree_root(state[])
check:
db.containsState(root)
hash_tree_root(db.getAltairStateRef(root)[]) == root
db.delState(root)
check:
not db.containsState(root)
db.getAltairStateRef(root).isNil
db.close()
test "sanity check phase 0 states, reusing buffers" & preset():
var
db = makeTestDB(SLOTS_PER_EPOCH)
dag = init(ChainDAGRef, defaultRuntimePreset, db)
let stateBuffer = (phase0.BeaconStateRef)()
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
var testStates = getTestStates(dag.headState.data)
# Ensure transitions beyond just adding validators and increasing slots
sort(testStates) do (x, y: ref ForkedHashedBeaconState) -> int:
cmp($getStateRoot(x[]), $getStateRoot(y[]))
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
for state in testStates:
db.putState(state[].hbsPhase0.data)
let root = hash_tree_root(state[])
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
check:
db.getState(root, stateBuffer[], noRollback)
db.containsState(root)
hash_tree_root(stateBuffer[]) == root
db.delState(root)
check:
not db.containsState(root)
not db.getState(root, stateBuffer[], noRollback)
db.close()
test "sanity check Altair states, reusing buffers" & preset():
var
db = makeTestDB(SLOTS_PER_EPOCH)
dag = init(ChainDAGRef, defaultRuntimePreset, db)
let stateBuffer = (altair.BeaconStateRef)()
var testStates = getTestStates(dag.headState.data, true)
# Ensure transitions beyond just adding validators and increasing slots
sort(testStates) do (x, y: ref ForkedHashedBeaconState) -> int:
cmp($getStateRoot(x[]), $getStateRoot(y[]))
for state in testStates:
db.putState(state[].hbsAltair.data)
let root = hash_tree_root(state[])
check:
db.getAltairState(root, stateBuffer[], noRollback)
db.containsState(root)
hash_tree_root(stateBuffer[]) == root
db.delState(root)
check:
not db.containsState(root)
not db.getAltairState(root, stateBuffer[], noRollback)
immutable validator database factoring (#2297) * initial immutable validator database factoring * remove changes from chain_dag: this abstraction properly belongs in beacon_chain_db * add merging mutable/immutable validator portions; individually test database roundtripping of immutable validators and states-sans-immutable-validators * update test summaries * use stew/assign2 instead of Nim assignment * add reading/writing of immutable validators in chaindag * remove unused import * replace chunked k/v store of immutable validators with per-row SQL table storage * use List instead of HashList * un-stub some ncli_db code so that it uses * switch HashArray to array; move BeaconStateNoImmutableValidators from datatypes to beacon_chain_db * begin only-mutable-part state storage * uncomment some assigns * work around https://github.com/nim-lang/Nim/issues/17253 * fix most of the issues/oversights; local sim runs again * fix test suite by adding missing beaconstate field to copy function * have ncli bench also store immutable validators * extract some immutable-validator-specific code from the beacon chain db module * add more rigorous database state roundtripping, with changing validator sets * adjust ncli_db to use new schema * simplify putState/getState by moving all immutable validator accounting into beacon state DB * remove redundant test case and move code to immutable-beacon-chain module * more efficient, but still brute-force, mutable+immutable validator merging * reuse BeaconState in getState * ensure HashList/HashArray caches are cleared when reusing getState buffers; add ncli_db and a unit test to verify this * HashList.clear() -> HashList.clearCache() * only copy incrementally necessary immutable validators * increase strictness of test cases and fix/work around resulting HashList cache invalidation issues * remove explanatory scaffolding * allow for storage of full (with all validators) states for backwards/forwards-compatibility * adjust DbSeq type usage * store full, with-validators, state every 64 epochs to enable reverting versions * reduce memory allocation and intermediate objects in state storage codepath * eliminate allocation/copying through intermediate BeaconStateNoImmutableValidators objects * skip benchmarking initial genesis-validator-heavy state store * always store new-style state and sometimes old-style state * document intent behind BeaconState/Validator type-punnery * more accurate failure message on SQLite in-memory database initialization failure
2021-03-15 14:11:51 +00:00
db.close()
test "sanity check phase 0 getState rollback" & preset():
var
db = makeTestDB(SLOTS_PER_EPOCH)
dag = init(ChainDAGRef, defaultRuntimePreset, db)
state = (ref ForkedHashedBeaconState)(
beaconStateFork: forkPhase0,
hbsPhase0: phase0.HashedBeaconState(data: phase0.BeaconState(
slot: 10.Slot)))
root = Eth2Digest()
db.putCorruptPhase0State(root)
let restoreAddr = addr dag.headState
func restore() =
assign(state[], restoreAddr[].data)
check:
state[].hbsPhase0.data.slot == 10.Slot
not db.getState(root, state[].hbsPhase0.data, restore)
state[].hbsPhase0.data.slot != 10.Slot
test "sanity check Altair and cross-fork getState rollback" & preset():
var
db = makeTestDB(SLOTS_PER_EPOCH)
dag = init(ChainDAGRef, defaultRuntimePreset, db)
state = (ref ForkedHashedBeaconState)(
beaconStateFork: forkAltair,
hbsAltair: altair.HashedBeaconState(data: altair.BeaconState(
slot: 10.Slot)))
root = Eth2Digest()
db.putCorruptAltairState(root)
let restoreAddr = addr dag.headState
func restore() =
assign(state[], restoreAddr[].data)
check:
state[].hbsAltair.data.slot == 10.Slot
not db.getAltairState(root, state[].hbsAltair.data, restore)
# assign() has switched the case object fork
state[].beaconStateFork == forkPhase0
state[].hbsPhase0.data.slot != 10.Slot
test "find ancestors" & preset():
var
db = BeaconChainDB.new(defaultRuntimePreset, "", inMemory = true)
let
a0 = withDigest(
(phase0.TrustedBeaconBlock)(slot: GENESIS_SLOT + 0))
a1 = withDigest(
(phase0.TrustedBeaconBlock)(slot: GENESIS_SLOT + 1, parent_root: a0.root))
a2 = withDigest(
(phase0.TrustedBeaconBlock)(slot: GENESIS_SLOT + 2, parent_root: a1.root))
doAssert toSeq(db.getAncestors(a0.root)) == []
doAssert toSeq(db.getAncestors(a2.root)) == []
doAssert toSeq(db.getAncestorSummaries(a0.root)).len == 0
doAssert toSeq(db.getAncestorSummaries(a2.root)).len == 0
db.putBlock(a2)
doAssert toSeq(db.getAncestors(a0.root)) == []
doAssert toSeq(db.getAncestors(a2.root)) == [a2]
doAssert toSeq(db.getAncestorSummaries(a0.root)).len == 0
doAssert toSeq(db.getAncestorSummaries(a2.root)).len == 1
db.putBlock(a1)
doAssert toSeq(db.getAncestors(a0.root)) == []
doAssert toSeq(db.getAncestors(a2.root)) == [a2, a1]
doAssert toSeq(db.getAncestorSummaries(a0.root)).len == 0
doAssert toSeq(db.getAncestorSummaries(a2.root)).len == 2
db.putBlock(a0)
doAssert toSeq(db.getAncestors(a0.root)) == [a0]
doAssert toSeq(db.getAncestors(a2.root)) == [a2, a1, a0]
doAssert toSeq(db.getAncestorSummaries(a0.root)).len == 1
doAssert toSeq(db.getAncestorSummaries(a2.root)).len == 3
test "sanity check genesis roundtrip" & preset():
# This is a really dumb way of checking that we can roundtrip a genesis
# state. We've been bit by this because we've had a bug in the BLS
# serialization where an all-zero default-initialized bls signature could
# not be deserialized because the deserialization was too strict.
var
db = BeaconChainDB.new(defaultRuntimePreset, "", inMemory = true)
let
state = initialize_beacon_state_from_eth1(
defaultRuntimePreset, eth1BlockHash, 0,
makeInitialDeposits(SLOTS_PER_EPOCH), {skipBlsValidation})
root = hash_tree_root(state[])
db.putState(state[])
check db.containsState(root)
let state2 = db.getPhase0StateRef(root)
db.delState(root)
check not db.containsState(root)
db.close()
check:
hash_tree_root(state2[]) == root
test "sanity check state diff roundtrip" & preset():
var
db = BeaconChainDB.new(defaultRuntimePreset, "", inMemory = true)
# TODO htr(diff) probably not interesting/useful, but stand-in
let
stateDiff = BeaconStateDiff()
root = hash_tree_root(stateDiff)
db.putStateDiff(root, stateDiff)
let state2 = db.getStateDiff(root)
db.delStateDiff(root)
check db.getStateDiff(root).isNone()
db.close()
check:
hash_tree_root(state2[]) == root