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

169 lines
5.5 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.
## Read vertex record on the layered Aristo DB delta architecture
## ==============================================================
{.push raises: [].}
import
std/tables,
results,
"."/[aristo_desc, aristo_layers]
# ------------------------------------------------------------------------------
# Public functions
# ------------------------------------------------------------------------------
2024-06-04 15:05:13 +00:00
proc getTuvUbe*(
db: AristoDbRef;
2024-06-04 15:05:13 +00:00
): Result[VertexID,AristoError] =
## Get the ID generator state from the unfiltered backened if available.
let be = db.backend
if not be.isNil:
2024-06-04 15:05:13 +00:00
return be.getTuvFn()
err(GetTuvNotFound)
proc getLstUbe*(
db: AristoDbRef;
): Result[SavedState,AristoError] =
## Get the last saved state
let be = db.backend
if not be.isNil:
return be.getLstFn()
err(GetLstNotFound)
proc getVtxUbe*(
db: AristoDbRef;
rvid: RootedVertexID;
flags: set[GetVtxFlag] = {};
): Result[VertexRef,AristoError] =
## Get the vertex from the unfiltered backened if available.
let be = db.backend
if not be.isNil:
return be.getVtxFn(rvid, flags)
err GetVtxNotFound
proc getKeyUbe*(
db: AristoDbRef;
rvid: RootedVertexID;
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
flags: set[GetVtxFlag];
): Result[(HashKey, VertexRef),AristoError] =
## Get the Merkle hash/key from the unfiltered backend if available.
let be = db.backend
if not be.isNil:
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
return be.getKeyFn(rvid, flags)
err GetKeyNotFound
# ------------------
2024-06-04 15:05:13 +00:00
proc getTuvBE*(
db: AristoDbRef;
2024-06-04 15:05:13 +00:00
): Result[VertexID,AristoError] =
## Get the ID generator state the `backened` layer if available.
if not db.balancer.isNil:
2024-06-04 15:05:13 +00:00
return ok(db.balancer.vTop)
db.getTuvUbe()
proc getVtxBE*(
db: AristoDbRef;
rvid: RootedVertexID;
flags: set[GetVtxFlag] = {};
): Result[(VertexRef, int),AristoError] =
## Get the vertex from the (filtered) backened if available.
if not db.balancer.isNil:
db.balancer.sTab.withValue(rvid, w):
if w[].isValid:
return ok (w[], -1)
return err(GetVtxNotFound)
ok (? db.getVtxUbe(rvid, flags), -2)
proc getKeyBE*(
db: AristoDbRef;
rvid: RootedVertexID;
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
flags: set[GetVtxFlag];
): Result[((HashKey, VertexRef), int),AristoError] =
## Get the merkle hash/key from the (filtered) backend if available.
if not db.balancer.isNil:
db.balancer.kMap.withValue(rvid, w):
if w[].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
return ok(((w[], nil), -1))
db.balancer.sTab.withValue(rvid, s):
if s[].isValid:
return ok(((VOID_HASH_KEY, s[]), -1))
return err(GetKeyNotFound)
ok ((?db.getKeyUbe(rvid, flags)), -2)
# ------------------
proc getVtxRc*(
db: AristoDbRef;
rvid: RootedVertexID;
flags: set[GetVtxFlag] = {};
): Result[(VertexRef, int),AristoError] =
## Cascaded attempt to fetch a vertex from the cache layers or the backend.
##
block body:
# If the vertex marked is to be deleted on the backend, a `VertexRef(nil)`
# entry is kept in the local table in which case it is returned as the
# error symbol `GetVtxNotFound`.
let vtx = db.layersGetVtx(rvid).valueOr:
break body
if vtx[0].isValid:
return ok vtx
else:
return err(GetVtxNotFound)
db.getVtxBE(rvid, flags)
Core db update providing additional results code interface (#1776) * Split `core_db/base.nim` into several sources * Rename `core_db/legacy.nim` => `core_db/legacy_db.nim` * Update `CoreDb` API, dual methods returning `Result[]` or plain value detail: Plain value methods implemet the legacy API, they defect on error results * Redesign `CoreDB` direct backend access why: Made the `backend` directive integral part of the API * Discontinue providing unused or otherwise available functions details: + setTransactionID() removed, not used and not easily replicable in Aristo + maybeGet() removed, available via direct backend access + newPhk() removed, never used & was experimental anyway * Update/reorg backend API why: + Added error print function `$$()` + General descriptor completion (and optional validation) via `bless()` * Update `Aristo`/`Kvt` exception handling why: Avoid `CatchableError` exceptions, rather pass them as error code where appropriate. * More `CoreDB` compliant `Aristo` and `Kvt` methods details: + Providing functions like `contains()`, `getVtxRc()` (returns `Result[]`). + Additional error code: `NotImplemented` * Rewrite/reorg of Aristo DB constructor why: Previously used global object `DefaultQidLayoutRef` as default initialiser. This object was created at compile time which lead to non-gc safe functions. * Update nimbus/db/core_db/legacy_db.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> * Update nimbus/db/aristo/aristo_transcode.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> * Update nimbus/db/core_db/legacy_db.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> --------- Co-authored-by: Kim De Mey <kim.demey@gmail.com>
2023-09-26 09:21:13 +00:00
proc getVtx*(db: AristoDbRef; rvid: RootedVertexID, flags: set[GetVtxFlag] = {}): VertexRef =
## Cascaded attempt to fetch a vertex from the cache layers or the backend.
Core db update providing additional results code interface (#1776) * Split `core_db/base.nim` into several sources * Rename `core_db/legacy.nim` => `core_db/legacy_db.nim` * Update `CoreDb` API, dual methods returning `Result[]` or plain value detail: Plain value methods implemet the legacy API, they defect on error results * Redesign `CoreDB` direct backend access why: Made the `backend` directive integral part of the API * Discontinue providing unused or otherwise available functions details: + setTransactionID() removed, not used and not easily replicable in Aristo + maybeGet() removed, available via direct backend access + newPhk() removed, never used & was experimental anyway * Update/reorg backend API why: + Added error print function `$$()` + General descriptor completion (and optional validation) via `bless()` * Update `Aristo`/`Kvt` exception handling why: Avoid `CatchableError` exceptions, rather pass them as error code where appropriate. * More `CoreDB` compliant `Aristo` and `Kvt` methods details: + Providing functions like `contains()`, `getVtxRc()` (returns `Result[]`). + Additional error code: `NotImplemented` * Rewrite/reorg of Aristo DB constructor why: Previously used global object `DefaultQidLayoutRef` as default initialiser. This object was created at compile time which lead to non-gc safe functions. * Update nimbus/db/core_db/legacy_db.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> * Update nimbus/db/aristo/aristo_transcode.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> * Update nimbus/db/core_db/legacy_db.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> --------- Co-authored-by: Kim De Mey <kim.demey@gmail.com>
2023-09-26 09:21:13 +00:00
## The function returns `nil` on error or failure.
##
db.getVtxRc(rvid).valueOr((VertexRef(nil), 0))[0]
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
proc getKeyRc*(
db: AristoDbRef; rvid: RootedVertexID, flags: set[GetVtxFlag]): Result[((HashKey, VertexRef), int),AristoError] =
## Cascaded attempt to fetch a Merkle hash from the cache layers or the
## backend. This function will never return a `VOID_HASH_KEY` but rather
## some `GetKeyNotFound` or `GetKeyUpdateNeeded` error.
##
block body:
let key = db.layersGetKey(rvid).valueOr:
break body
# If there is a zero key value, the entry is either marked for being
# updated or for deletion on the database. So check below.
if key[0].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
return ok ((key[0], nil), key[1])
# The zero key value does not refer to an update mark if there is no
# valid vertex (either on the cache or the backend whatever comes first.)
let vtx = db.layersGetVtx(rvid).valueOr:
# There was no vertex on the cache. So there must be one the backend (the
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
# reason for the key label to exists, at all.)
return err(GetKeyNotFound)
if vtx[0].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
return ok ((VOID_HASH_KEY, vtx[0]), vtx[1])
else:
# The vertex is to be deleted. So is the value key.
return err(GetKeyNotFound)
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
db.getKeyBE(rvid, flags)
Core db update providing additional results code interface (#1776) * Split `core_db/base.nim` into several sources * Rename `core_db/legacy.nim` => `core_db/legacy_db.nim` * Update `CoreDb` API, dual methods returning `Result[]` or plain value detail: Plain value methods implemet the legacy API, they defect on error results * Redesign `CoreDB` direct backend access why: Made the `backend` directive integral part of the API * Discontinue providing unused or otherwise available functions details: + setTransactionID() removed, not used and not easily replicable in Aristo + maybeGet() removed, available via direct backend access + newPhk() removed, never used & was experimental anyway * Update/reorg backend API why: + Added error print function `$$()` + General descriptor completion (and optional validation) via `bless()` * Update `Aristo`/`Kvt` exception handling why: Avoid `CatchableError` exceptions, rather pass them as error code where appropriate. * More `CoreDB` compliant `Aristo` and `Kvt` methods details: + Providing functions like `contains()`, `getVtxRc()` (returns `Result[]`). + Additional error code: `NotImplemented` * Rewrite/reorg of Aristo DB constructor why: Previously used global object `DefaultQidLayoutRef` as default initialiser. This object was created at compile time which lead to non-gc safe functions. * Update nimbus/db/core_db/legacy_db.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> * Update nimbus/db/aristo/aristo_transcode.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> * Update nimbus/db/core_db/legacy_db.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> --------- Co-authored-by: Kim De Mey <kim.demey@gmail.com>
2023-09-26 09:21:13 +00:00
proc getKey*(db: AristoDbRef; rvid: RootedVertexID): HashKey =
## Cascaded attempt to fetch a vertex from the cache layers or the backend.
Core db update providing additional results code interface (#1776) * Split `core_db/base.nim` into several sources * Rename `core_db/legacy.nim` => `core_db/legacy_db.nim` * Update `CoreDb` API, dual methods returning `Result[]` or plain value detail: Plain value methods implemet the legacy API, they defect on error results * Redesign `CoreDB` direct backend access why: Made the `backend` directive integral part of the API * Discontinue providing unused or otherwise available functions details: + setTransactionID() removed, not used and not easily replicable in Aristo + maybeGet() removed, available via direct backend access + newPhk() removed, never used & was experimental anyway * Update/reorg backend API why: + Added error print function `$$()` + General descriptor completion (and optional validation) via `bless()` * Update `Aristo`/`Kvt` exception handling why: Avoid `CatchableError` exceptions, rather pass them as error code where appropriate. * More `CoreDB` compliant `Aristo` and `Kvt` methods details: + Providing functions like `contains()`, `getVtxRc()` (returns `Result[]`). + Additional error code: `NotImplemented` * Rewrite/reorg of Aristo DB constructor why: Previously used global object `DefaultQidLayoutRef` as default initialiser. This object was created at compile time which lead to non-gc safe functions. * Update nimbus/db/core_db/legacy_db.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> * Update nimbus/db/aristo/aristo_transcode.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> * Update nimbus/db/core_db/legacy_db.nim Co-authored-by: Kim De Mey <kim.demey@gmail.com> --------- Co-authored-by: Kim De Mey <kim.demey@gmail.com>
2023-09-26 09:21:13 +00:00
## The function returns `nil` on error or failure.
##
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
(db.getKeyRc(rvid, {}).valueOr(((VOID_HASH_KEY, nil), 0)))[0][0]
# ------------------------------------------------------------------------------
# End
# ------------------------------------------------------------------------------