nimbus-eth1/nimbus/sync/snap/range_desc.nim

454 lines
14 KiB
Nim
Raw Normal View History

Prep for full sync after snap (#1253) * Split fetch accounts into sub-modules details: There will be separated modules for accounts snapshot, storage snapshot, and healing for either. * Allow to rebase pivot before negotiated header why: Peers seem to have not too many snapshots available. By setting back the pivot block header slightly, the chances might be higher to find more peers to serve this pivot. Experiment on mainnet showed that setting back too much (tested with 1024), the chances to find matching snapshot peers seem to decrease. * Add accounts healing * Update variable/field naming in `worker_desc` for readability * Handle leaf nodes in accounts healing why: There is no need to fetch accounts when they had been added by the healing process. On the flip side, these accounts must be checked for storage data and the batch queue updated, accordingly. * Reorganising accounts hash ranges batch queue why: The aim is to formally cover as many accounts as possible for different pivot state root environments. Formerly, this was tried by starting the accounts batch queue at a random value for each pivot (and wrapping around.) Now, each pivot environment starts with an interval set mutually disjunct from any interval set retrieved with other pivot state roots. also: Stop fishing for more pivots in `worker` if 100% download is reached * Reorganise/update accounts healing why: Error handling was wrong and the (math. complexity of) whole process could be better managed. details: Much of the algorithm is now documented at the top of the file `heal_accounts.nim`
2022-10-08 17:20:50 +00:00
# Nimbus
# Copyright (c) 2018-2021 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.
{.push raises: [].}
import
Prep for full sync after snap make 6 (#1291) * Update log ticker, using time interval rather than ticker count why: Counting and logging ticker occurrences is inherently imprecise. So time intervals are used. * Use separate storage tables for snap sync data * Left boundary proof update why: Was not properly implemented, yet. * Capture pivot in peer worker (aka buddy) tasks why: The pivot environment is linked to the `buddy` descriptor. While there is a task switch, the pivot may change. So it is passed on as function argument `env` rather than retrieved from the buddy at the start of a sub-function. * Split queues `fetchStorage` into `fetchStorageFull` and `fetchStoragePart` * Remove obsolete account range returned from `GetAccountRange` message why: Handler returned the wrong right value of the range. This range was for convenience, only. * Prioritise storage slots if the queue becomes large why: Currently, accounts processing is prioritised up until all accounts are downloaded. The new prioritisation has two thresholds for + start processing storage slots with a new worker + stop account processing and switch to storage processing also: Provide api for `SnapTodoRanges` pair of range sets in `worker_desc.nim` * Generalise left boundary proof for accounts or storage slots. why: Detailed explanation how this works is documented with `snapdb_accounts.importAccounts()`. Instead of enforcing a left boundary proof (which is still the default), the importer functions return a list of `holes` (aka node paths) found in the argument ranges of leaf nodes. This in turn is used by the book keeping software for data download. * Forgot to pass on variable in function wrapper also: + Start healing not before 99% accounts covered (previously 95%) + Logging updated/prettified
2022-11-08 18:56:04 +00:00
std/[math, sequtils, strutils, hashes],
eth/common,
stew/interval_set,
stint,
../../constants,
../../utils/prettify,
../protocol,
../types
export
types
type
ByteArray32* = array[32,byte]
## Used for 32 byte database keys
NodeKey* = distinct ByteArray32
## Hash key without the hash wrapper (as opposed to `NodeTag` which is a
## number.)
NodeTag* = distinct UInt256
## Trie leaf item, account hash etc. This data type is a representation
## for a `NodeKey` geared up for arithmetic and comparing keys.
NodeTagRange* = Interval[NodeTag,UInt256]
## Interval `[minPt,maxPt]` of` NodeTag` elements, can be managed in an
## `IntervalSet` data type.
NodeTagRangeSet* = IntervalSetRef[NodeTag,UInt256]
## Managed structure to handle non-adjacent `NodeTagRange` intervals
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
NodeSpecs* = object
## Multi purpose descriptor for a hexary trie node:
## * Missing node specs. If the `data` argument is empty, the `partialPath`
## refers to a missoing node entry. The `nodeKey` is another way of
## writing the node hash and used to verify that a potential data `Blob`
## is acceptable as node data.
## * Node data. If the `data` argument is non-empty, the `partialPath`
## fields can/will be used as function argument for various functions
## when healing.
partialPath*: Blob ## Compact encoded partial path nibbles
nodeKey*: NodeKey ## Derived from node hash
data*: Blob ## Node data (might not be present)
PackedAccountRange* = object
## Re-packed version of `SnapAccountRange`. The reason why repacking is
## needed is that the `snap/1` protocol uses another RLP encoding than is
## used for storing in the database. So the `PackedAccount` is `BaseDB`
## trie compatible.
accounts*: seq[PackedAccount] ## List of re-packed accounts data
proof*: seq[SnapProof] ## Boundary proofs
PackedAccount* = object
## In fact, the `snap/1` driver returns the `Account` structure which is
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
## unwanted overhead, here.
accKey*: NodeKey
accBlob*: Blob
AccountCodeHeader* = object
## Contract code header
accKey*: NodeKey ## Owner account
codeHash*: Hash256 ## Contarct code hash
AccountSlotsHeader* = object
## Storage root header
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
accKey*: NodeKey ## Owner account, maybe unnecessary
storageRoot*: Hash256 ## Start of storage tree
subRange*: Option[NodeTagRange] ## Sub-range of slot range covered
AccountSlotsChanged* = object
## Variant of `AccountSlotsHeader` representing some transition
account*: AccountSlotsHeader ## Account header
newRange*: Option[NodeTagRange] ## New sub-range (if-any)
AccountStorageRange* = object
## List of storage descriptors, the last `AccountSlots` storage data might
Prep for full sync after snap make 6 (#1291) * Update log ticker, using time interval rather than ticker count why: Counting and logging ticker occurrences is inherently imprecise. So time intervals are used. * Use separate storage tables for snap sync data * Left boundary proof update why: Was not properly implemented, yet. * Capture pivot in peer worker (aka buddy) tasks why: The pivot environment is linked to the `buddy` descriptor. While there is a task switch, the pivot may change. So it is passed on as function argument `env` rather than retrieved from the buddy at the start of a sub-function. * Split queues `fetchStorage` into `fetchStorageFull` and `fetchStoragePart` * Remove obsolete account range returned from `GetAccountRange` message why: Handler returned the wrong right value of the range. This range was for convenience, only. * Prioritise storage slots if the queue becomes large why: Currently, accounts processing is prioritised up until all accounts are downloaded. The new prioritisation has two thresholds for + start processing storage slots with a new worker + stop account processing and switch to storage processing also: Provide api for `SnapTodoRanges` pair of range sets in `worker_desc.nim` * Generalise left boundary proof for accounts or storage slots. why: Detailed explanation how this works is documented with `snapdb_accounts.importAccounts()`. Instead of enforcing a left boundary proof (which is still the default), the importer functions return a list of `holes` (aka node paths) found in the argument ranges of leaf nodes. This in turn is used by the book keeping software for data download. * Forgot to pass on variable in function wrapper also: + Start healing not before 99% accounts covered (previously 95%) + Logging updated/prettified
2022-11-08 18:56:04 +00:00
## be incomplete and the `proof` is needed for proving validity.
storages*: seq[AccountSlots] ## List of accounts and storage data
proof*: seq[SnapProof] ## Boundary proofs for last entry
Prep for full sync after snap make 6 (#1291) * Update log ticker, using time interval rather than ticker count why: Counting and logging ticker occurrences is inherently imprecise. So time intervals are used. * Use separate storage tables for snap sync data * Left boundary proof update why: Was not properly implemented, yet. * Capture pivot in peer worker (aka buddy) tasks why: The pivot environment is linked to the `buddy` descriptor. While there is a task switch, the pivot may change. So it is passed on as function argument `env` rather than retrieved from the buddy at the start of a sub-function. * Split queues `fetchStorage` into `fetchStorageFull` and `fetchStoragePart` * Remove obsolete account range returned from `GetAccountRange` message why: Handler returned the wrong right value of the range. This range was for convenience, only. * Prioritise storage slots if the queue becomes large why: Currently, accounts processing is prioritised up until all accounts are downloaded. The new prioritisation has two thresholds for + start processing storage slots with a new worker + stop account processing and switch to storage processing also: Provide api for `SnapTodoRanges` pair of range sets in `worker_desc.nim` * Generalise left boundary proof for accounts or storage slots. why: Detailed explanation how this works is documented with `snapdb_accounts.importAccounts()`. Instead of enforcing a left boundary proof (which is still the default), the importer functions return a list of `holes` (aka node paths) found in the argument ranges of leaf nodes. This in turn is used by the book keeping software for data download. * Forgot to pass on variable in function wrapper also: + Start healing not before 99% accounts covered (previously 95%) + Logging updated/prettified
2022-11-08 18:56:04 +00:00
base*: NodeTag ## Lower limit for last entry w/proof
AccountSlots* = object
## Account storage descriptor
account*: AccountSlotsHeader
data*: seq[SnapStorage]
# See below for definition of constant `FullNodeTagRange`
# ------------------------------------------------------------------------------
# Public helpers
# ------------------------------------------------------------------------------
proc to*(tag: NodeTag; T: type Hash256): T =
## Convert to serialised equivalent
result.data = tag.UInt256.toBytesBE
proc to*(key: NodeKey; T: type NodeTag): T =
## Convert from serialised equivalent
UInt256.fromBytesBE(key.ByteArray32).T
proc to*(key: Hash256; T: type NodeTag): T =
## Syntactic sugar
key.data.NodeKey.to(T)
proc to*(tag: NodeTag; T: type NodeKey): T =
## Syntactic sugar
tag.UInt256.toBytesBE.T
proc to*(hash: Hash256; T: type NodeKey): T =
## Syntactic sugar
hash.data.NodeKey
proc to*(key: NodeKey; T: type Hash256): T =
## Syntactic sugar
T(data: key.ByteArray32)
proc to*(key: NodeKey; T: type Blob): T =
## Syntactic sugar
key.ByteArray32.toSeq
proc to*(n: SomeUnsignedInt|UInt256; T: type NodeTag): T =
## Syntactic sugar
n.u256.T
proc digestTo*(data: Blob; T: type NodeKey): T =
keccakHash(data).data.T
proc hash*(a: NodeKey): Hash =
## Table/KeyedQueue mixin
a.ByteArray32.hash
proc `==`*(a, b: NodeKey): bool =
## Table/KeyedQueue mixin
a.ByteArray32 == b.ByteArray32
# ------------------------------------------------------------------------------
# Public constructors
# ------------------------------------------------------------------------------
proc init*(key: var NodeKey; data: openArray[byte]): bool =
## Import argument `data` into `key` which must have length either `32`, or
## `0`. The latter case is equivalent to an all zero byte array of size `32`.
if data.len == 32:
(addr key.ByteArray32[0]).copyMem(unsafeAddr data[0], data.len)
return true
elif data.len == 0:
key.reset
return true
proc init*(tag: var NodeTag; data: openArray[byte]): bool =
## Similar to `init(key: var NodeHash; .)`.
var key: NodeKey
if key.init(data):
tag = key.to(NodeTag)
return true
# ------------------------------------------------------------------------------
# Public rlp support
# ------------------------------------------------------------------------------
proc read*[T: NodeTag|NodeKey](rlp: var Rlp, W: type T): T
{.gcsafe, raises: [RlpError].} =
rlp.read(Hash256).to(T)
proc append*(writer: var RlpWriter, val: NodeTag|NodeKey) =
writer.append(val.to(Hash256))
# ------------------------------------------------------------------------------
# Public `NodeTag` and `NodeTagRange` functions
# ------------------------------------------------------------------------------
proc u256*(lp: NodeTag): UInt256 = lp.UInt256
proc low*(T: type NodeTag): T = low(UInt256).T
proc high*(T: type NodeTag): T = high(UInt256).T
proc `+`*(a: NodeTag; b: UInt256): NodeTag = (a.u256+b).NodeTag
proc `-`*(a: NodeTag; b: UInt256): NodeTag = (a.u256-b).NodeTag
proc `-`*(a, b: NodeTag): UInt256 = (a.u256 - b.u256)
proc `==`*(a, b: NodeTag): bool = a.u256 == b.u256
proc `<=`*(a, b: NodeTag): bool = a.u256 <= b.u256
proc `<`*(a, b: NodeTag): bool = a.u256 < b.u256
proc cmp*(x, y: NodeTag): int = cmp(x.UInt256, y.UInt256)
proc hash*(a: NodeTag): Hash =
## Mixin for `Table` or `keyedQueue`
a.to(Hash256).data.hash
proc digestTo*(data: Blob; T: type NodeTag): T =
## Hash the `data` argument
keccakHash(data).to(T)
const
# Cannot be defined earlier: `NodeTag` operations needed
FullNodeTagRange* = NodeTagRange.new(low(NodeTag),high(NodeTag))
# ------------------------------------------------------------------------------
# Public functions: `NodeTagRange` helpers
# ------------------------------------------------------------------------------
proc isEmpty*(lrs: NodeTagRangeSet): bool =
## Returns `true` if the argument set `lrs` of intervals is empty
lrs.chunks == 0
proc isEmpty*(lrs: openArray[NodeTagRangeSet]): bool =
## Variant of `isEmpty()` where intervals are distributed across several
## sets.
for ivSet in lrs:
if 0 < ivSet.chunks:
return false
true
Prep for full sync after snap (#1253) * Split fetch accounts into sub-modules details: There will be separated modules for accounts snapshot, storage snapshot, and healing for either. * Allow to rebase pivot before negotiated header why: Peers seem to have not too many snapshots available. By setting back the pivot block header slightly, the chances might be higher to find more peers to serve this pivot. Experiment on mainnet showed that setting back too much (tested with 1024), the chances to find matching snapshot peers seem to decrease. * Add accounts healing * Update variable/field naming in `worker_desc` for readability * Handle leaf nodes in accounts healing why: There is no need to fetch accounts when they had been added by the healing process. On the flip side, these accounts must be checked for storage data and the batch queue updated, accordingly. * Reorganising accounts hash ranges batch queue why: The aim is to formally cover as many accounts as possible for different pivot state root environments. Formerly, this was tried by starting the accounts batch queue at a random value for each pivot (and wrapping around.) Now, each pivot environment starts with an interval set mutually disjunct from any interval set retrieved with other pivot state roots. also: Stop fishing for more pivots in `worker` if 100% download is reached * Reorganise/update accounts healing why: Error handling was wrong and the (math. complexity of) whole process could be better managed. details: Much of the algorithm is now documented at the top of the file `heal_accounts.nim`
2022-10-08 17:20:50 +00:00
proc isEmpty*(iv: NodeTagRange): bool =
## Ditto for an interval range.
false # trivially by definition
proc isFull*(lrs: NodeTagRangeSet): bool =
## Returns `true` if the argument set `lrs` contains of the single
## interval [low(NodeTag),high(NodeTag)].
lrs.total == 0 and 0 < lrs.chunks
proc isFull*(lrs: openArray[NodeTagRangeSet]): bool =
## Variant of `isFull()` where intervals are distributed across several
## sets. This function makes sense only if the interval sets are mutually
## disjunct.
var accu: NodeTag
for ivSet in lrs:
if 0 < ivSet.total:
if high(NodeTag) - ivSet.total < accu:
return true
accu = accu + ivSet.total
elif 0 < ivSet.chunks:
# number of points in `ivSet` is `2^256 + 1`
return true
proc isFull*(iv: NodeTagRange): bool =
## Ditto for an interval range.
iv == FullNodeTagRange
proc emptyFactor*(lrs: NodeTagRangeSet): float =
Prep for full sync after snap (#1253) * Split fetch accounts into sub-modules details: There will be separated modules for accounts snapshot, storage snapshot, and healing for either. * Allow to rebase pivot before negotiated header why: Peers seem to have not too many snapshots available. By setting back the pivot block header slightly, the chances might be higher to find more peers to serve this pivot. Experiment on mainnet showed that setting back too much (tested with 1024), the chances to find matching snapshot peers seem to decrease. * Add accounts healing * Update variable/field naming in `worker_desc` for readability * Handle leaf nodes in accounts healing why: There is no need to fetch accounts when they had been added by the healing process. On the flip side, these accounts must be checked for storage data and the batch queue updated, accordingly. * Reorganising accounts hash ranges batch queue why: The aim is to formally cover as many accounts as possible for different pivot state root environments. Formerly, this was tried by starting the accounts batch queue at a random value for each pivot (and wrapping around.) Now, each pivot environment starts with an interval set mutually disjunct from any interval set retrieved with other pivot state roots. also: Stop fishing for more pivots in `worker` if 100% download is reached * Reorganise/update accounts healing why: Error handling was wrong and the (math. complexity of) whole process could be better managed. details: Much of the algorithm is now documented at the top of the file `heal_accounts.nim`
2022-10-08 17:20:50 +00:00
## Relative uncovered total, i.e. `#points-not-covered / 2^256` to be used
## in statistics or triggers.
if 0 < lrs.total:
((high(NodeTag) - lrs.total).u256 + 1).to(float) / (2.0^256)
elif lrs.chunks == 0:
1.0 # `total` represents the residue class `mod 2^256` from `0`..`(2^256-1)`
else:
Prep for full sync after snap (#1253) * Split fetch accounts into sub-modules details: There will be separated modules for accounts snapshot, storage snapshot, and healing for either. * Allow to rebase pivot before negotiated header why: Peers seem to have not too many snapshots available. By setting back the pivot block header slightly, the chances might be higher to find more peers to serve this pivot. Experiment on mainnet showed that setting back too much (tested with 1024), the chances to find matching snapshot peers seem to decrease. * Add accounts healing * Update variable/field naming in `worker_desc` for readability * Handle leaf nodes in accounts healing why: There is no need to fetch accounts when they had been added by the healing process. On the flip side, these accounts must be checked for storage data and the batch queue updated, accordingly. * Reorganising accounts hash ranges batch queue why: The aim is to formally cover as many accounts as possible for different pivot state root environments. Formerly, this was tried by starting the accounts batch queue at a random value for each pivot (and wrapping around.) Now, each pivot environment starts with an interval set mutually disjunct from any interval set retrieved with other pivot state roots. also: Stop fishing for more pivots in `worker` if 100% download is reached * Reorganise/update accounts healing why: Error handling was wrong and the (math. complexity of) whole process could be better managed. details: Much of the algorithm is now documented at the top of the file `heal_accounts.nim`
2022-10-08 17:20:50 +00:00
0.0 # number of points in `lrs` is `2^256 + 1`
proc emptyFactor*(lrs: openArray[NodeTagRangeSet]): float =
Prep for full sync after snap (#1253) * Split fetch accounts into sub-modules details: There will be separated modules for accounts snapshot, storage snapshot, and healing for either. * Allow to rebase pivot before negotiated header why: Peers seem to have not too many snapshots available. By setting back the pivot block header slightly, the chances might be higher to find more peers to serve this pivot. Experiment on mainnet showed that setting back too much (tested with 1024), the chances to find matching snapshot peers seem to decrease. * Add accounts healing * Update variable/field naming in `worker_desc` for readability * Handle leaf nodes in accounts healing why: There is no need to fetch accounts when they had been added by the healing process. On the flip side, these accounts must be checked for storage data and the batch queue updated, accordingly. * Reorganising accounts hash ranges batch queue why: The aim is to formally cover as many accounts as possible for different pivot state root environments. Formerly, this was tried by starting the accounts batch queue at a random value for each pivot (and wrapping around.) Now, each pivot environment starts with an interval set mutually disjunct from any interval set retrieved with other pivot state roots. also: Stop fishing for more pivots in `worker` if 100% download is reached * Reorganise/update accounts healing why: Error handling was wrong and the (math. complexity of) whole process could be better managed. details: Much of the algorithm is now documented at the top of the file `heal_accounts.nim`
2022-10-08 17:20:50 +00:00
## Variant of `emptyFactor()` where intervals are distributed across several
## sets. This function makes sense only if the interval sets are mutually
## disjunct.
var accu: NodeTag
Prep for full sync after snap (#1253) * Split fetch accounts into sub-modules details: There will be separated modules for accounts snapshot, storage snapshot, and healing for either. * Allow to rebase pivot before negotiated header why: Peers seem to have not too many snapshots available. By setting back the pivot block header slightly, the chances might be higher to find more peers to serve this pivot. Experiment on mainnet showed that setting back too much (tested with 1024), the chances to find matching snapshot peers seem to decrease. * Add accounts healing * Update variable/field naming in `worker_desc` for readability * Handle leaf nodes in accounts healing why: There is no need to fetch accounts when they had been added by the healing process. On the flip side, these accounts must be checked for storage data and the batch queue updated, accordingly. * Reorganising accounts hash ranges batch queue why: The aim is to formally cover as many accounts as possible for different pivot state root environments. Formerly, this was tried by starting the accounts batch queue at a random value for each pivot (and wrapping around.) Now, each pivot environment starts with an interval set mutually disjunct from any interval set retrieved with other pivot state roots. also: Stop fishing for more pivots in `worker` if 100% download is reached * Reorganise/update accounts healing why: Error handling was wrong and the (math. complexity of) whole process could be better managed. details: Much of the algorithm is now documented at the top of the file `heal_accounts.nim`
2022-10-08 17:20:50 +00:00
for ivSet in lrs:
if 0 < ivSet.total:
if high(NodeTag) - ivSet.total < accu:
return 0.0
accu = accu + ivSet.total
elif ivSet.chunks == 0:
discard
else: # number of points in `ivSet` is `2^256 + 1`
return 0.0
# Calculate: (2^256 - accu) / 2^256
if accu == 0.to(NodeTag):
1.0
else:
((high(NodeTag) - accu).u256 + 1).to(float) / (2.0^256)
proc fullFactor*(lrs: NodeTagRangeSet): float =
Prep for full sync after snap (#1253) * Split fetch accounts into sub-modules details: There will be separated modules for accounts snapshot, storage snapshot, and healing for either. * Allow to rebase pivot before negotiated header why: Peers seem to have not too many snapshots available. By setting back the pivot block header slightly, the chances might be higher to find more peers to serve this pivot. Experiment on mainnet showed that setting back too much (tested with 1024), the chances to find matching snapshot peers seem to decrease. * Add accounts healing * Update variable/field naming in `worker_desc` for readability * Handle leaf nodes in accounts healing why: There is no need to fetch accounts when they had been added by the healing process. On the flip side, these accounts must be checked for storage data and the batch queue updated, accordingly. * Reorganising accounts hash ranges batch queue why: The aim is to formally cover as many accounts as possible for different pivot state root environments. Formerly, this was tried by starting the accounts batch queue at a random value for each pivot (and wrapping around.) Now, each pivot environment starts with an interval set mutually disjunct from any interval set retrieved with other pivot state roots. also: Stop fishing for more pivots in `worker` if 100% download is reached * Reorganise/update accounts healing why: Error handling was wrong and the (math. complexity of) whole process could be better managed. details: Much of the algorithm is now documented at the top of the file `heal_accounts.nim`
2022-10-08 17:20:50 +00:00
## Relative covered total, i.e. `#points-covered / 2^256` to be used
## in statistics or triggers
if 0 < lrs.total:
lrs.total.u256.to(float) / (2.0^256)
elif lrs.chunks == 0:
Prep for full sync after snap (#1253) * Split fetch accounts into sub-modules details: There will be separated modules for accounts snapshot, storage snapshot, and healing for either. * Allow to rebase pivot before negotiated header why: Peers seem to have not too many snapshots available. By setting back the pivot block header slightly, the chances might be higher to find more peers to serve this pivot. Experiment on mainnet showed that setting back too much (tested with 1024), the chances to find matching snapshot peers seem to decrease. * Add accounts healing * Update variable/field naming in `worker_desc` for readability * Handle leaf nodes in accounts healing why: There is no need to fetch accounts when they had been added by the healing process. On the flip side, these accounts must be checked for storage data and the batch queue updated, accordingly. * Reorganising accounts hash ranges batch queue why: The aim is to formally cover as many accounts as possible for different pivot state root environments. Formerly, this was tried by starting the accounts batch queue at a random value for each pivot (and wrapping around.) Now, each pivot environment starts with an interval set mutually disjunct from any interval set retrieved with other pivot state roots. also: Stop fishing for more pivots in `worker` if 100% download is reached * Reorganise/update accounts healing why: Error handling was wrong and the (math. complexity of) whole process could be better managed. details: Much of the algorithm is now documented at the top of the file `heal_accounts.nim`
2022-10-08 17:20:50 +00:00
0.0 # `total` represents the residue class `mod 2^256` from `0`..`(2^256-1)`
else:
Prep for full sync after snap (#1253) * Split fetch accounts into sub-modules details: There will be separated modules for accounts snapshot, storage snapshot, and healing for either. * Allow to rebase pivot before negotiated header why: Peers seem to have not too many snapshots available. By setting back the pivot block header slightly, the chances might be higher to find more peers to serve this pivot. Experiment on mainnet showed that setting back too much (tested with 1024), the chances to find matching snapshot peers seem to decrease. * Add accounts healing * Update variable/field naming in `worker_desc` for readability * Handle leaf nodes in accounts healing why: There is no need to fetch accounts when they had been added by the healing process. On the flip side, these accounts must be checked for storage data and the batch queue updated, accordingly. * Reorganising accounts hash ranges batch queue why: The aim is to formally cover as many accounts as possible for different pivot state root environments. Formerly, this was tried by starting the accounts batch queue at a random value for each pivot (and wrapping around.) Now, each pivot environment starts with an interval set mutually disjunct from any interval set retrieved with other pivot state roots. also: Stop fishing for more pivots in `worker` if 100% download is reached * Reorganise/update accounts healing why: Error handling was wrong and the (math. complexity of) whole process could be better managed. details: Much of the algorithm is now documented at the top of the file `heal_accounts.nim`
2022-10-08 17:20:50 +00:00
1.0 # number of points in `lrs` is `2^256 + 1`
proc fullFactor*(lrs: openArray[NodeTagRangeSet]): float =
## Variant of `fullFactor()` where intervals are distributed across several
## sets. This function makes sense only if the interval sets are mutually
## disjunct.
var accu: NodeTag
for ivSet in lrs:
if 0 < ivSet.total:
if high(NodeTag) - ivSet.total < accu:
return 1.0
accu = accu + ivSet.total
elif ivSet.chunks == 0:
discard
else: # number of points in `ivSet` is `2^256 + 1`
return 1.0
accu.u256.to(float) / (2.0^256)
proc fullFactor*(iv: NodeTagRange): float =
## Relative covered length of an inetrval, i.e. `#points-covered / 2^256`
if 0 < iv.len:
iv.len.u256.to(float) / (2.0^256)
else:
1.0 # number of points in `iv` is `2^256 + 1`
# ------------------------------------------------------------------------------
# Public functions: printing & pretty printing
# ------------------------------------------------------------------------------
Prep for full sync after snap (#1253) * Split fetch accounts into sub-modules details: There will be separated modules for accounts snapshot, storage snapshot, and healing for either. * Allow to rebase pivot before negotiated header why: Peers seem to have not too many snapshots available. By setting back the pivot block header slightly, the chances might be higher to find more peers to serve this pivot. Experiment on mainnet showed that setting back too much (tested with 1024), the chances to find matching snapshot peers seem to decrease. * Add accounts healing * Update variable/field naming in `worker_desc` for readability * Handle leaf nodes in accounts healing why: There is no need to fetch accounts when they had been added by the healing process. On the flip side, these accounts must be checked for storage data and the batch queue updated, accordingly. * Reorganising accounts hash ranges batch queue why: The aim is to formally cover as many accounts as possible for different pivot state root environments. Formerly, this was tried by starting the accounts batch queue at a random value for each pivot (and wrapping around.) Now, each pivot environment starts with an interval set mutually disjunct from any interval set retrieved with other pivot state roots. also: Stop fishing for more pivots in `worker` if 100% download is reached * Reorganise/update accounts healing why: Error handling was wrong and the (math. complexity of) whole process could be better managed. details: Much of the algorithm is now documented at the top of the file `heal_accounts.nim`
2022-10-08 17:20:50 +00:00
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
proc `$`*(nodeTag: NodeTag): string =
if nodeTag == high(NodeTag):
Prep for full sync after snap make 6 (#1291) * Update log ticker, using time interval rather than ticker count why: Counting and logging ticker occurrences is inherently imprecise. So time intervals are used. * Use separate storage tables for snap sync data * Left boundary proof update why: Was not properly implemented, yet. * Capture pivot in peer worker (aka buddy) tasks why: The pivot environment is linked to the `buddy` descriptor. While there is a task switch, the pivot may change. So it is passed on as function argument `env` rather than retrieved from the buddy at the start of a sub-function. * Split queues `fetchStorage` into `fetchStorageFull` and `fetchStoragePart` * Remove obsolete account range returned from `GetAccountRange` message why: Handler returned the wrong right value of the range. This range was for convenience, only. * Prioritise storage slots if the queue becomes large why: Currently, accounts processing is prioritised up until all accounts are downloaded. The new prioritisation has two thresholds for + start processing storage slots with a new worker + stop account processing and switch to storage processing also: Provide api for `SnapTodoRanges` pair of range sets in `worker_desc.nim` * Generalise left boundary proof for accounts or storage slots. why: Detailed explanation how this works is documented with `snapdb_accounts.importAccounts()`. Instead of enforcing a left boundary proof (which is still the default), the importer functions return a list of `holes` (aka node paths) found in the argument ranges of leaf nodes. This in turn is used by the book keeping software for data download. * Forgot to pass on variable in function wrapper also: + Start healing not before 99% accounts covered (previously 95%) + Logging updated/prettified
2022-11-08 18:56:04 +00:00
"2^256-1"
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
elif nodeTag == 0.u256.NodeTag:
"0"
elif nodeTag == 2.u256.pow(255).NodeTag:
"2^255" # 800...
elif nodeTag == 2.u256.pow(254).NodeTag:
"2^254" # 400..
elif nodeTag == 2.u256.pow(253).NodeTag:
"2^253" # 200...
elif nodeTag == 2.u256.pow(251).NodeTag:
"2^252" # 100...
else:
nodeTag.UInt256.toHex
Prep for full sync after snap make 4 (#1282) * Re-arrange fetching storage slots in batch module why; Previously, fetching partial slot ranges first has a chance of terminating the worker peer 9due to network error) while there were many inheritable storage slots on the queue. Now, inheritance is checked first, then full slot ranges and finally partial ranges. * Update logging * Bundled node information for healing into single object `NodeSpecs` why: Previously, partial paths and node keys were kept in separate variables. This approach was error prone due to copying/reassembling function argument objects. As all partial paths, keys, and node data types are more or less handled as `Blob`s over the network (using Eth/6x, or Snap/1) it makes sense to hold these `Blob`s as named field in a single object (even if not all fields are active for the current purpose.) * For good housekeeping, using `NodeKey` type only for account keys why: previously, a mixture of `NodeKey` and `Hash256` was used. Now, only state or storage root keys use the `Hash256` type. * Always accept latest pivot (and not a slightly older one) why; For testing it was tried to use a slightly older pivot state root than available. Some anecdotal tests seemed to suggest an advantage so that more peers are willing to serve on that older pivot. But this could not be confirmed in subsequent tests (still anecdotal, though.) As a side note, the distance of the latest pivot to its predecessor is at least 128 (or whatever the constant `minPivotBlockDistance` is assigned to.) * Reshuffle name components for some file and function names why: Clarifies purpose: "storages" becomes: "storage slots" "store" becomes: "range fetch" * Stash away currently unused modules in sub-folder named "notused"
2022-10-27 13:49:28 +00:00
proc `$`*(nodeKey: NodeKey): string =
$nodeKey.to(NodeTag)
proc leafRangePp*(a, b: NodeTag): string =
## Needed for macro generated DSL files like `snap.nim` because the
## `distinct` flavour of `NodeTag` is discarded there.
result = "[" & $a
if a != b:
result &= ',' & $b
result &= "]"
Snap sync refactor accounts healing (#1392) * Relocated mothballing (i.e. swap-in preparation) logic details: Mothballing was previously tested & started after downloading account ranges in `range_fetch_accounts`. Whenever current download or healing stops because of a pivot change, swap-in preparation is needed (otherwise some storage slots may get lost when swap-in takes place.) Also, `execSnapSyncAction()` has been moved back to `pivot_helper`. * Reorganised source file directories details: Grouped pivot focused modules into `pivot` directory * Renamed `checkNodes`, `sickSubTries` as `nodes.check`, `nodes.missing` why: Both lists are typically used together as pair. Renaming `sickSubTries` reflects moving away from a healing centric view towards a swap-in attitude. * Multi times coverage recording details: Per pivot account ranges are accumulated into coverage range set. This set fill eventually contain a singe range of account hashes [0..2^256] which amounts to 100% capacity. A counter has been added that is incremented whenever max capacity is reached. The accumulated range is then reset to empty. The effect of this setting is that the coverage can be evenly duplicated. So 200% would not accumulate on a particular region. * Update range length comparisons (mod 2^256) why: A range interval can have sizes 1..2^256 as it cannot be empty by definition. The number of points in a range intervals set can have 0..2^256 points. As the scalar range is a residue class modulo 2^256, the residue class 0 means length 2^256 for a range interval, but can be 0 or 2^256 for the number of points in a range intervals set. * Generalised `hexaryEnvelopeDecompose()` details: Compile the complement of the union of some (processed) intervals and express this complement as a list of envelopes of sub-tries. This facility is directly applicable to swap-in book-keeping. * Re-factor `swapIn()` why: Good idea but baloney implementation. The main algorithm is based on the generalised version of `hexaryEnvelopeDecompose()` which has been derived from this implementation. * Refactor `healAccounts()` using `hexaryEnvelopeDecompose()` as main driver why: Previously, the hexary trie was searched recursively for dangling nodes which has a poor worst case performance already when the trie is reasonably populated. The function `hexaryEnvelopeDecompose()` is a magnitude faster because it does not peruse existing sub-tries in order to find missing nodes although result is not fully compatible with the previous function. So recursive search is used in a limited mode only when the decomposer will not deliver a useful result. * Logging & maintenance fixes details: Preparation for abandoning buddy-global healing variables `node`, `resumeCtx`, and `lockTriePerusal`. These variable are trie-perusal centric which will be run on the back burner in favour of `hexaryEnvelopeDecompose()` which is used for accounts healing already.
2022-12-19 21:22:09 +00:00
proc leafRangePp*(iv: NodeTagRange): string =
## Variant of `leafRangePp()`
leafRangePp(iv.minPt, iv.maxPt)
proc `$`*(a, b: NodeTag): string =
## Prettyfied prototype
leafRangePp(a,b)
proc `$`*(iv: NodeTagRange): string =
Snap sync refactor accounts healing (#1392) * Relocated mothballing (i.e. swap-in preparation) logic details: Mothballing was previously tested & started after downloading account ranges in `range_fetch_accounts`. Whenever current download or healing stops because of a pivot change, swap-in preparation is needed (otherwise some storage slots may get lost when swap-in takes place.) Also, `execSnapSyncAction()` has been moved back to `pivot_helper`. * Reorganised source file directories details: Grouped pivot focused modules into `pivot` directory * Renamed `checkNodes`, `sickSubTries` as `nodes.check`, `nodes.missing` why: Both lists are typically used together as pair. Renaming `sickSubTries` reflects moving away from a healing centric view towards a swap-in attitude. * Multi times coverage recording details: Per pivot account ranges are accumulated into coverage range set. This set fill eventually contain a singe range of account hashes [0..2^256] which amounts to 100% capacity. A counter has been added that is incremented whenever max capacity is reached. The accumulated range is then reset to empty. The effect of this setting is that the coverage can be evenly duplicated. So 200% would not accumulate on a particular region. * Update range length comparisons (mod 2^256) why: A range interval can have sizes 1..2^256 as it cannot be empty by definition. The number of points in a range intervals set can have 0..2^256 points. As the scalar range is a residue class modulo 2^256, the residue class 0 means length 2^256 for a range interval, but can be 0 or 2^256 for the number of points in a range intervals set. * Generalised `hexaryEnvelopeDecompose()` details: Compile the complement of the union of some (processed) intervals and express this complement as a list of envelopes of sub-tries. This facility is directly applicable to swap-in book-keeping. * Re-factor `swapIn()` why: Good idea but baloney implementation. The main algorithm is based on the generalised version of `hexaryEnvelopeDecompose()` which has been derived from this implementation. * Refactor `healAccounts()` using `hexaryEnvelopeDecompose()` as main driver why: Previously, the hexary trie was searched recursively for dangling nodes which has a poor worst case performance already when the trie is reasonably populated. The function `hexaryEnvelopeDecompose()` is a magnitude faster because it does not peruse existing sub-tries in order to find missing nodes although result is not fully compatible with the previous function. So recursive search is used in a limited mode only when the decomposer will not deliver a useful result. * Logging & maintenance fixes details: Preparation for abandoning buddy-global healing variables `node`, `resumeCtx`, and `lockTriePerusal`. These variable are trie-perusal centric which will be run on the back burner in favour of `hexaryEnvelopeDecompose()` which is used for accounts healing already.
2022-12-19 21:22:09 +00:00
leafRangePp iv
proc fullPC3*(w: NodeTagRangeSet|NodeTagRange): string =
## Pretty print fill state of range sets.
if w.isEmpty:
"0%"
elif w.isFull:
"100%"
else:
let ff = w.fullFactor
if ff <= 0.99999:
ff.toPC(3)
else:
"99.999"
proc fullPC3*(w: openArray[NodeTagRangeSet]): string =
## Variant of `fullPC3()` where intervals are distributed across several
## sets. This function makes sense only if the interval sets are mutually
## disjunct.
if w.isEmpty:
"0%"
else:
let partition = "~" & $w.mapIt(it.chunks).foldl(a+b)
if w.isFull:
"100%" & partition
else:
let ff = w.fullFactor
if ff <= 0.99999:
ff.toPC(3) & partition
else:
"99.999" & partition
Prep for full sync after snap make 6 (#1291) * Update log ticker, using time interval rather than ticker count why: Counting and logging ticker occurrences is inherently imprecise. So time intervals are used. * Use separate storage tables for snap sync data * Left boundary proof update why: Was not properly implemented, yet. * Capture pivot in peer worker (aka buddy) tasks why: The pivot environment is linked to the `buddy` descriptor. While there is a task switch, the pivot may change. So it is passed on as function argument `env` rather than retrieved from the buddy at the start of a sub-function. * Split queues `fetchStorage` into `fetchStorageFull` and `fetchStoragePart` * Remove obsolete account range returned from `GetAccountRange` message why: Handler returned the wrong right value of the range. This range was for convenience, only. * Prioritise storage slots if the queue becomes large why: Currently, accounts processing is prioritised up until all accounts are downloaded. The new prioritisation has two thresholds for + start processing storage slots with a new worker + stop account processing and switch to storage processing also: Provide api for `SnapTodoRanges` pair of range sets in `worker_desc.nim` * Generalise left boundary proof for accounts or storage slots. why: Detailed explanation how this works is documented with `snapdb_accounts.importAccounts()`. Instead of enforcing a left boundary proof (which is still the default), the importer functions return a list of `holes` (aka node paths) found in the argument ranges of leaf nodes. This in turn is used by the book keeping software for data download. * Forgot to pass on variable in function wrapper also: + Start healing not before 99% accounts covered (previously 95%) + Logging updated/prettified
2022-11-08 18:56:04 +00:00
proc dump*(
ranges: openArray[NodeTagRangeSet];
moan: proc(overlap: UInt256; iv: NodeTagRange) {.gcsafe.};
printRangesMax = high(int);
): string =
## Dump/anlalyse range sets
var
cache: NodeTagRangeSet
ivTotal = 0.u256
ivCarry = false
if ranges.len == 1:
cache = ranges[0]
ivTotal = cache.total
if ivTotal == 0.u256 and 0 < cache.chunks:
ivCarry = true
else:
cache = NodeTagRangeSet.init()
for ivSet in ranges:
if ivSet.total == 0.u256 and 0 < ivSet.chunks:
ivCarry = true
elif ivTotal <= high(UInt256) - ivSet.total:
ivTotal += ivSet.total
else:
ivCarry = true
for iv in ivSet.increasing():
let n = cache.merge(iv)
if n != iv.len and not moan.isNil:
moan(iv.len - n, iv)
if 0 == cache.total and 0 < cache.chunks:
result = "2^256"
if not ivCarry:
result &= ":" & $ivTotal
else:
result = $cache.total
if ivCarry:
result &= ":2^256"
elif ivTotal != cache.total:
result &= ":" & $ivTotal
result &= ":"
if cache.chunks <= printRangesMax:
result &= toSeq(cache.increasing).mapIt($it).join(",")
else:
result &= toSeq(cache.increasing).mapIt($it)[0 ..< printRangesMax].join(",")
result &= " " & $(cache.chunks - printRangesMax) & " more .."
proc dump*(
range: NodeTagRangeSet;
printRangesMax = high(int);
): string =
## Ditto
[range].dump(nil, printRangesMax)
# ------------------------------------------------------------------------------
# End
# ------------------------------------------------------------------------------