Commit Graph

19 Commits

Author SHA1 Message Date
Alex Sharov
c5ffc971c5
[WIP] Badger v2 (#378)
* badger v2 investigation

* buf pool - use native New method and avoid double checks

* db.Open prototype

* db.Tx/Bucket/Cursor prototypes

* Chained config

* Item concept added

* save changes to test on master

* make hack resumable

* Design document v0

* Cursor concept

* less brackets syntax of cursor builder

* benchmarks

* cleanup fs

* test for context cancelations

* test for context cancelations

* test for cursor.Prefix option

* add ForEachKey method

* add ForEachKey method

* add naming explanation

* experiment of non-pointers cursor/bucket

* .Bucket() and .Cursor() doesn't returns error

* .Bucket() and .Cursor() doesn't returns error

* .Bucket() and .Cursor() doesn't returns error

* remove CursorOpts concept

* more test-cases

* simplify open api

* Tx, Bucket, Cursor - now are interfaces

* Tx, Bucket, Cursor - now are interfaces

* switch to interfaces

* rebase master

Co-authored-by: alex.sharov <alex.sharov@lazada.com>
2020-03-11 11:02:37 +00:00
Alex Sharov
ce96cf75b2
Intermediate hash phase 3 (#377)
* #remove debug prints

* remove storage-mode="i"

* minnet re-execute hack with checkpoints

* minnet re-execute hack with checkpoints

* rollback to master setup

* mainnet re-exec hack

* rollback some changes

* v0 of "push down" functionality

* move all logic to own functions

* handle case when re-created account already has some storage

* clear path for storage

* try to rely on tree structure (but maybe need to rely on DB because can be intra-block re-creations of account)

* fix some bugs with indexes, moving to tests

* tests added

* make linter happy

* make linter happy

* simplify logic

* adjust comparison of keys with and without incarnation

* test for keyIsBefore

* test for keyIsBefore

* better nibbles alignment

* better nibbles alignment

* cleanup

* continue work on tests

* simplify test

* check tombstone existence before pushing it down.

* put tombstone only when account deleted, not created

* put tombstone only when account has storage

* make linter happy

* test for storage resolver

* make fixedbytes work without incarnation

* fix panic on short keys

* use special comparison only when working with keys from cache

* add blockNr for better tracing

* fix: incorrect tombstone check

* fix: incorrect tombstone check

* trigger ci

* hack for problem block

* more test-cases

* add test case for too long keys

* speedup cached resolver by removing bucket creation transaction

* remove parent type check in pruning, remove unused copy from mutation.put

* dump resolving info on fail

* dump resolving info on fail

* set tombstone everytime for now to check if it will help

* on unload: check parent type, not type of node

* fix wrong order of checking node type

* fix wrong order of checking node type

* rebase to new master

* make linter happy

* rebase to new master

* place tombstone only if acc has storage

* rebase master

* rebase master

* rebase master

* rebase master

Co-authored-by: alex.sharov <alex.sharov@lazada.com>
2020-03-11 10:31:49 +00:00
ledgerwatch
80f06b279a
Fix --download-only mode (#374)
* Fix download only

* Fix lint

* Reset references

* Only reset on error

* Potential fixes

* no NPE

* no NPE

* Not use multi-put

* Reduce ideal batch size for download only

* Handle tds == nil

* remove nested mutation

* Return multiput

* Better reporting

* Reduce batch size for download only

* Avoid extra copying

* Avoid extra copying

* IdealBatchSize

* Not write tx lookup entries

* Larger batches

* Go back to normal batch size

* Fix lint

* Gen tx lookup

* print progress

* Add filling up the lookup array

* Show tx count

* Introduce second round

* Add generating tx lookup

* Fix lint

* properly stop at specified block

* measure the duration of the last phase

* not to fail if the bucket is not found

* Fix lint

* Alternative tx generation

* Fix out of memory

* Fix out of memory

* Split in parts to conserve memory

* Copy keys

* Fix lint

* Fix lint
2020-03-01 09:00:14 +00:00
b00ris
83e65fdd81 migrations
fix build

add storage mode check

add intermitiate migration state&snapshot migrations
2020-02-26 00:25:30 +03:00
b00ris
958ace4088 split storage and account encoding for changeset 2020-02-24 22:36:16 +03:00
b00ris
e69f53db8d Merge branch 'master' of github.com:ledgerwatch/turbo-geth into split_cs_buckets_squash 2020-02-24 21:47:26 +03:00
b00ris
aedecfc820 split changeset bucket
fmt

badger changes

badget puts

fix shadowing

fix mutation test. restore DecodeChangeSet

fix pruner test
2020-02-24 20:56:32 +03:00
Alex Sharov
fdbba5202b
Trie: store self-destructed accounts (#355)
* squash commits

* enable storage cache

* make linter happy

* fix subtree prefix len check

* save cahnges to test master

* remove restriction on prefix len

* fix comparison of last bits

* fix wrong alignment

* remove debug prints

* commit current state

* commit current state

* avoid changing state of resolver from multiwalk

* remove debug code

* remove debug code

* remove debug code

* remove unnecessary copy

* make code more readable

* reduce rebuildHashes initial resolution

* fix test after rebase to master

* make code more readable

* improve pruner

* pruner add IntermediateCache bucket

* fix panic in Walk on short keys

* reduce allocations for storage keys decompression by increasing default buffer size

* re-run CI

* fix iterator behaviour

* rename cache to hash for unification

* re-run ci

* avoid using underlying DB

* hash all subtree nodes before unload

* fix getNode method

* need to check node type, not parent - before put to hashBucket

* return back parent type check, doesn't work without it.

* don't recalculate hash again

* move unloadFunc from trie to pruner

* rename bucket to shorter name

* rename bucket to shorter name

* clean

* rebase to master
2020-02-12 13:52:59 +00:00
ledgerwatch
767a374c2e
Tx pool use triedbstate, fixes to GetNodeData and move it to experimental (#364)
* Use TrieDbState for tx pool

* Not initialise tx pool until state is loaded

* Add preimage

* Fix account

* Print codehash

* Print correct code hash

* Print incarnatin

* Print incarnatin

* Use proper incarnation

* Print dbValue

* Actually fix

* Actually fix

* Fix verifySnapshot

* readAccount to get code hash

* Next incarnation

* Print addrHashes with 0 incarnations

* Print storage history

* Print storage history

* Print storage history

* Print storage history

* Print all storage history

* print change set keys

* print change set keys

* print change set keys

* print change set keys

* Not print codebucket info

* Fixes

* Fix for incarnation

* Fix for storage history bucket

* Try to fix the leak

* Try to fix the leak

* Try to fix the leak

* Try to fix the leak

* Try to fix the leak

* Try to fix the leak

* Try to fix the leak

* Fix embedded nodes

* Hasher

* Fix

* Test fixes

* Add experimental debug flag

* Fix tx_pool_test

* Disable GetNodeData test unless in experiment

* Fix more tests

* Fix lint and revert some changes

* Fix lint

* Fix lint
2020-02-10 17:05:32 +00:00
Andrew Ashikhmin
8585516071
Serve GetNodeData for nodes in memory (Issue #300) 2020-02-06 11:53:09 +01:00
Evgeny Danilenko
17a4a56634
Refactor mining. Remove a few mining goroutines (#338)
* add context

* extract chain events

* run commit in goroutines

* mine only on canonical

* typo

* linters

* fmt

* mark unused methods

* restore stress test

* test single miner

* remove unsafe Trie storage

* remove locks from miner

* restore interrupt

* remove result goroutine

* remove unconfirmedBlocks

* cherry-pick 04a1d475ff1a36ad8f92fec80385df18c52bdc1f

* extract uncles

* one miner succeeded

* restore context cancel

* cleanup

* skip an unstable test

* remove pending state

* use context instead of interrupt func

* calculate sealHash only once

* comment out unstable test

* after merge

* fix after merge

Co-authored-by: ledgerwatch <akhounov@gmail.com>
2020-02-03 15:02:26 +03:00
ledgerwatch
4f5c858f00
Collect data about Keccak256 invocations for evaluation of STARK proof sizes and performance (#315)
* Gathering start data on witnesses

* Fix number parsing

* Proper filename, actually outputting

* Correct block

* Visualise states

* Code and value fix

* Visualise code hash and storage hash

* Current block number

* Print bucket

* Fix lint

* Fix lint

* Fix lint

* Fix lint

* Fix typo

* Fixes after merging master

* Fix lint

Co-authored-by: b00ris <b00ris@mail.ru>
2020-01-15 17:33:36 +00:00
ledgerwatch
dbc37b5660
bolt vs badger log parser and chart (#284) 2019-12-24 13:26:29 +00:00
ledgerwatch
a521aef296
Run archive node by default (#204)
* Make archive mode default again

* Fix test

* Remove bad file

* Change default of NoHistory to false
2019-11-27 15:50:40 +00:00
ledgerwatch
b2ca635d86
Debug tool for recursively comparing state (in DB or in a file) with geth archive node (#191)
* Fetching results of eth_getProof

* Dump 5 levels of the trie in a file for repeated runs

* Drill down to 6 levels of the trie

* Fix lint

* Fix lint

* Fix lint

* verifySnapshot to check accounts with emptyRoot

* Descend into short nodes

* Latest tool fixes

* Fix lint

* Fix state properly working
2019-11-25 13:36:21 +00:00
ledgerwatch
244d70fb9c
Further fixes for the no-mod-root (#186)
* Further fixes

* Repace 1000 with a symbol
2019-11-21 15:56:39 +00:00
b00ris
6e880cc945
suffix to changeset rename (#152) 2019-11-07 18:51:25 +03:00
ledgerwatch
d1c8416270 Merge pull request #114 from ledgerwatch/fix_sync_crash
Fix the sync crash due to the mis-accounting of prunable entries
2019-11-04 14:15:26 +01:00
Alexey Akhunov
fe01bccbb8 Apply Turbo-Geth modifications to go-ethereum codebase 2019-11-01 21:52:03 +01:00