fix: fix Kusama sync; add storageState lock in core.HandleTransactionMessage #1783

noot · 2021-09-09T20:35:56Z

Changes

add storageState lock in core.HandleTransactionMessage to use while validating transaction
there was also a race condition / edge case between calling GetRuntime and TrieState as the best block might change between these two calls, and if the runtime changes between the two calls this might be an issue. so it's updated to get the best block hash at the start of the func and use that
note: this seems to somewhat improve the issue but not fully solve it, as there are still bugs in the current syncQueue (syncing refactor appears address this, currently running a node to confirm)

Tests

make gossamer
./bin/gossamer --chain kusama init --force
./bin/gossamer --chain kusama

Issues

closes Kusama Node not syncing shortly after launch #1770

Primary Reviewer

@timwu20

codecov · 2021-09-09T21:05:06Z

Codecov Report

Merging #1783 (db119ba) into development (a91de8b) will increase coverage by 0.06%.
The diff coverage is 76.00%.

@@               Coverage Diff               @@
##           development    #1783      +/-   ##
===============================================
+ Coverage        59.15%   59.22%   +0.06%     
===============================================
  Files              186      186              
  Lines            19998    20008      +10     
===============================================
+ Hits             11830    11849      +19     
+ Misses            6112     6099      -13     
- Partials          2056     2060       +4

Flag	Coverage Δ
unit-tests	`59.22% <76.00%> (+0.06%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
dot/core/messages.go	`75.75% <76.00%> (-6.86%)`	⬇️
lib/babe/babe.go	`56.70% <0.00%> (-1.15%)`	⬇️
dot/core/service.go	`50.39% <0.00%> (-0.79%)`	⬇️
lib/blocktree/blocktree.go	`56.74% <0.00%> (+1.12%)`	⬆️
lib/babe/epoch.go	`66.66% <0.00%> (+2.15%)`	⬆️
dot/network/sync.go	`63.70% <0.00%> (+2.76%)`	⬆️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a91de8b...db119ba. Read the comment docs.

…/fix-kusama

arijitAD · 2021-09-13T10:33:28Z

dot/core/messages.go

-			return false, err
-		}
+		err = func() error {
+			s.storageState.Lock()


Should we move this lock outside of the loop?

I figured to put it inside the loop since if it locks outside the loop instead of for each transaction, it might cause more lock contention

timwu20 · 2021-09-13T17:55:44Z

dot/core/interface.go

+	Lock()
+	Unlock()


Suggested change

Lock()

Unlock()

sync.Locker

# [0.6.0](v0.5.0...v0.6.0) (2021-12-03) ### Bug Fixes * **babe:** Fix extrinsic format in block. ([#1530](#1530)) ([1a03b2a](1a03b2a)) * **ci:** add missing go-version matrix to fix development branch CI ([#2037](#2037)) ([6babe76](6babe76)) * **cmd/cfg:** Use Babe Lead value from toml config ([#2032](#2032)) ([06aa3e3](06aa3e3)) * cmd/gossamer: Generate random name if --name flag not set ([#1506](#1506)) ([3c05a88](3c05a88)) * confirm block import notifier is closed properly ([#1736](#1736)) ([ad2d85e](ad2d85e)) * **docs:** improve build-spec usage docs ([#1706](#1706)) ([2e164b4](2e164b4)) * **dot/core:** Add only extrinsic during chain reorg. ([#1609](#1609)) ([29413d4](29413d4)) * **dot/core:** Batch process transaction message. ([#1780](#1780)) ([0064836](0064836)) * **dot/core:** check transaction Validity.Propagate field to determine whether to propagate tx ([#1643](#1643)) ([81f23cc](81f23cc)) * **dot/core:** Fix handle transaction message test. ([#1607](#1607)) ([58b8725](58b8725)) * **dot/network, lib/grandpa:** fix handshake decoding and grandpa message handler sigabort ([#1631](#1631)) ([887f72c](887f72c)) * **dot/network, lib/grandpa:** fix node sync, improve devnet finality ([bcc7935](bcc7935)) * **dot/network:** add nil checks in connManager ([#2069](#2069)) ([7f9c042](7f9c042)) * **dot/network:** Check for size when decoding leb128. ([#1634](#1634)) ([d082b9e](d082b9e)) * **dot/network:** check if peer supports protocol ([#1617](#1617)) ([6bf66a4](6bf66a4)) * **dot/network:** decrease DHT find peers interval for gssmr nodes ([#1703](#1703)) ([08516a0](08516a0)) * **dot/network:** fix bugs in notifications protocol handlers; add metrics for inbound/outbound streams ([#2010](#2010)) ([8c2993d](8c2993d)) * **dot/network:** fix dht connection on discovery on devnet ([#2059](#2059)) ([da065b8](da065b8)) * **dot/network:** fix discovery between gossamer nodes ([#1594](#1594)) ([f4c79d3](f4c79d3)) * **dot/network:** fix justification request at head logic ([#1510](#1510)) ([98d1413](98d1413)) * **dot/network:** fix memory allocations with `sizedBufferPool` ([#1963](#1963)) ([e0b126b](e0b126b)) * **dot/network:** Fix missing digest in header ([#2092](#2092)) ([21ea85e](21ea85e)) * **dot/network:** Fix notification handshake and reuse stream. ([#1545](#1545)) ([a632dc4](a632dc4)) * **dot/network:** fix receiving notifications messages from substrate peers ([#1517](#1517)) ([fdf3c53](fdf3c53)) * **dot/network:** fix stream manager tests ([#1683](#1683)) ([e02eca4](e02eca4)) * **dot/network:** implement a handshake timeout ([#1615](#1615)) ([87c2f63](87c2f63)) * **dot/network:** Implement time based handle transaction ([#1942](#1942)) ([dd08424](dd08424)) * **dot/network:** move low reputation peer removal from network ConnManager to peer scoring logic (dot/peerstate) ([#2068](#2068)) ([ac16285](ac16285)), closes [#2039](#2039) * **dot/network:** Return on EOF error while reading stream. ([#1733](#1733)) ([f447eac](f447eac)) * **dot/network:** split stored streams and handshakeData into inbound and outbound ([#1553](#1553)) ([637050b](637050b)) * **dot/network:** update notificationsProtocol handshakeData to sync.Map ([#1492](#1492)) ([22f7269](22f7269)) * **dot/node:** Start websocket server only with `--ws` flag ([#1671](#1671)) ([6ecef3b](6ecef3b)) * **dot/state, lib/babe, lib/trie:** improve syncing between gossamer authority nodes ([#1613](#1613)) ([ca99fbf](ca99fbf)) * **dot/state, lib/grandpa:** update justification and SignedVote handling in database ([#1682](#1682)) ([bbdcd6f](bbdcd6f)) * **dot/state,dot/network:** improve memory usage when syncing ([#1491](#1491)) ([3b2ad8d](3b2ad8d)) * **dot/state:** add StorageState Lock/Unlock API for usage by babe and sync ([#1700](#1700)) ([3c22ace](3c22ace)) * **dot/state:** fix deadlock, fixes bootstrap syncing ([#1959](#1959)) ([dd80c09](dd80c09)) * **dot/state:** fix usage of trie.Snapshot ([#1489](#1489)) ([3880a40](3880a40)) * **dot/state:** track runtime per-block, fix runtime upgrades differing between forks ([#1638](#1638)) ([e133884](e133884)) * **dot/state:** update `*state.BlockState.AddBlockToBlockTree` to store block in `unfinalisedBlocksMap` ([#2006](#2006)) ([55d997f](55d997f)) * **dot/sync:** add nil header checks ([#2099](#2099)) ([a7d4be0](a7d4be0)) * **dot/sync:** fix block request and response logic ([#1907](#1907)) ([9c6283e](9c6283e)) * **dot/sync:** fix creating block response, fixes node sync between gossamer nodes ([#1572](#1572)) ([1328c80](1328c80)) * **dot/telemetry:** refactor telemetry to reduce CPU usage ([#1597](#1597)) ([bc31ac7](bc31ac7)) * **dot/types:** *types.Body to be of type []types.Extrinsic ([#1807](#1807)) ([4c09715](4c09715)) * **dot/types:** fix max value for digest ([#1687](#1687)) ([48405e7](48405e7)) * **dot:** fix `TestNewNode` ([#2070](#2070)) ([42908d0](42908d0)) * fix edit link ([#1507](#1507)) ([5089327](5089327)) * fix Kusama sync; add storageState lock in core.HandleTransactionMessage ([#1783](#1783)) ([1d688e4](1d688e4)) * **lib/babe, lib/runtime/wasmer:** fixes for v0.9.8+ runtime ([#2075](#2075)) ([2f9f80c](2f9f80c)) * **lib/babe:** add `--babe-lead` flag, update epoch handling logic ([#1895](#1895)) ([7abcce6](7abcce6)) * **lib/babe:** add pre-runtime digest before calling initialize_block ([#1581](#1581)) ([c1b26d3](c1b26d3)) * **lib/babe:** always use 2/3 of slot to produce block, re-add potentially valid txs to queue ([#1679](#1679)) ([cf93ad3](cf93ad3)) * **lib/babe:** call AddBlock in BABE synchronously ([#1585](#1585)) ([86acc43](86acc43)) * **lib/babe:** fix BABE state storing after building block ([#1536](#1536)) ([1a3dea2](1a3dea2)) * **lib/babe:** fix err log ([#1801](#1801)) ([a96f06a](a96f06a)) * **lib/babe:** fix setting first slot of network, fix loading BABE epoch params ([#1640](#1640)) ([5c3dbfe](5c3dbfe)) * **lib/babe:** fix timing for transition between epochs ([#1636](#1636)) ([57027db](57027db)) * **lib/blocktree:** fix blocktree bug ([#2060](#2060)) ([c17b53a](c17b53a)) * **lib/blocktree:** fix potential nil pointer dereference in `HighestCommonAncestor`, core `handleBlocksAsync` ([#1993](#1993)) ([f7f4463](f7f4463)) * **lib/blocktree:** fix setting leaves after blocktree pruning ([#1605](#1605)) ([58c0854](58c0854)) * **lib/blocktree:** removes the inconsistency to choose a deepest leaf ([#2094](#2094)) ([43d68e3](43d68e3)) * **lib/crypto/ed25519:** update ed25519 to use go-schnorrkel bip39 derivation ([#1488](#1488)) ([dfb95d2](dfb95d2)) * **lib/genesis:** Update missing and incorrect fields in genesis file. ([#1681](#1681)) ([8207704](8207704)) * **lib/grandpa:** fix grandpa stall and various bugs ([#1708](#1708)) ([67c93f4](67c93f4)) * **lib/grandpa:** fix grandpa vote message switch ([#2095](#2095)) ([461890c](461890c)) * **lib/grandpa:** fix threshold checking to be strictly greater than 2/3 ([#1891](#1891)) ([66ffe51](66ffe51)) * **lib/grandpa:** use `defaultGrandpaInterval` if not set, fixes error on startup ([#1982](#1982)) ([75627b5](75627b5)) * **lib/runtime/life:** remove import C from life ([#1923](#1923)) ([ed507d2](ed507d2)) * **lib/runtime:** update HOST_API_TEST_RUNTIME_URL to reference specific commit ([#1885](#1885)) ([666ed06](666ed06)) * **log-levels:** do not ignore configuration file log levels ([#2016](#2016)) ([80879b2](80879b2)) * pending bubble hidden after block included ([#1592](#1592)) ([5826322](5826322)) * persist node name ([#1543](#1543)) ([88b88f2](88b88f2)) * **pprof:** only run pprof service if enabled ([#2073](#2073)) ([55669c5](55669c5)) * **release:** Trigger release when pushed to main branch. ([#1566](#1566)) ([d445c97](d445c97)) * **rpc/subscription:** subscribe runtime version notify when version changes ([#1686](#1686)) ([9a76d39](9a76d39)) * Staging CI workflow ([#2034](#2034)) ([84ec792](84ec792)) * **trie:** memory leak fix in `lib/trie` ([#2009](#2009)) ([0ad5eb7](0ad5eb7)) * update deprecated package ([#1603](#1603)) ([f195204](f195204)) * update go-schnorrkel version ([#1557](#1557)) ([b86c7ff](b86c7ff)) * update gssmr genesis to use pallet_babe::SameAuthoritiesForever ([#1696](#1696)) ([fb0a751](fb0a751)) * update HOST_API_TEST_RUNTIME_URL ([#1898](#1898)) ([2ef59a8](2ef59a8)) * **utils:** create a specific folder for database ([#1598](#1598)) ([8c67795](8c67795)) ### Features * add --chain dev option ([#1561](#1561)) ([04a2969](04a2969)) * Add properties and chainId on build-spec command ([#1520](#1520)) ([b18290c](b18290c)) * cmd: implement import-runtime subcommand ([#1483](#1483)) ([d82b2da](d82b2da)) * **cmd/gossamer:** implement --telemetry-url parameter ([#1890](#1890)) ([b202e89](b202e89)), closes [#1502](#1502) * **cmd:** implement offline pruning of state trie ([#1564](#1564)) ([af9c925](af9c925)) * **devnet:** Local Gossamer Devnet ([#2008](#2008)) ([a520001](a520001)) * **dot/network, lib/grandpa:** request justification on receiving NeighbourMessage, verify justification on receipt ([#1529](#1529)) ([e1f9f42](e1f9f42)) * **dot/network:** Add cache for network message. ([#1511](#1511)) ([accaf69](accaf69)) * **dot/network:** add propagate return bool to messageHandler func type to determine whether to propagate message or not ([#1555](#1555)) ([0d6f488](0d6f488)) * **dot/network:** implement persistent peers functionality ([#1512](#1512)) ([7850532](7850532)) * **dot/network:** implement streamManager to cleanup not recently used streams ([#1611](#1611)) ([ba861bf](ba861bf)) * **dot/network:** request block justifications when near head ([#1499](#1499)) ([ae7012b](ae7012b)) * **dot/peerset:** Implement peer scoring ([#1791](#1791)) ([1c989ad](1c989ad)) * **dot/rpc/modules:** add `system_addReservedPeer` and `system_removeReservedPeer` RPC call ([#1712](#1712)) ([dba5922](dba5922)) * **dot/rpc:** Add `system_localListenAddresses` RPC call ([#1689](#1689)) ([c981d2e](c981d2e)) * **dot/rpc:** Implement `childstate_getKeys` rpc call ([#1800](#1800)) ([9b2f41e](9b2f41e)) * **dot/rpc:** implement sync_state_genSyncSpec RPC call ([#1827](#1827)) ([2186caf](2186caf)) * **dot/state:** implement online pruning of historical state tries ([#1596](#1596)) ([3eb9399](3eb9399)) * **dot/sync:** implement codeSubstitutes ([#1635](#1635)) ([d87aaeb](d87aaeb)) * dot/telemetry: Implement basic telemetry connection ([#1497](#1497)) ([fcb4159](fcb4159)) * **dot/telemetry:** Added connection retry ([#1904](#1904)) ([579a791](579a791)) * **dot/telemetry:** Added more telemetry messages in grandpa client ([#2043](#2043)) ([2e57d15](2e57d15)), closes [#1841](#1841) [#1842](#1842) * **dot/telemetry:** implement notify.finalized telemetry interface ([#1877](#1877)) ([de1a60d](de1a60d)) * **dot/telemetry:** implement substrate_number_leaves metrics ([#1926](#1926)) ([69823c0](69823c0)) * **dot/telemetry:** implement telemetry message network_state ([#1618](#1618)) ([a81844e](a81844e)) * **flags:** read log levels from flags ([#1953](#1953)) ([9694e46](9694e46)) * implement ext_default_child_storage_storage_kill_version_2 ([#1799](#1799)) ([c2908ae](c2908ae)) * implement ext_offchain_index_set_version_1 for wasmer runtime ([#1739](#1739)) ([96c30a6](96c30a6)) * **lib/babe:** add check of types.ConfigData.SecondarySlots for disabling secondary verification ([#1910](#1910)) ([cd27ae4](cd27ae4)) * **lib/grandpa:** fully verify justifications using GrandpaState ([#1544](#1544)) ([028d25e](028d25e)) * **lib/grandpa:** Include equivocatory nodes while creating justification ([#1911](#1911)) ([aca86b6](aca86b6)) * **lib/grandpa:** send NeighbourMessage to peers ([#1558](#1558)) ([322ccf9](322ccf9)) * **lib/runtime/wasmer:** implement ext_default_child_storage_storage_kill_version_3 ([#1878](#1878)) ([a719a60](a719a60)) * **lib/runtime/wasmer:** implement ext_offchain_local_storage_version_1 ([#1821](#1821)) ([0f63b17](0f63b17)) * **lib/runtime:** Implement `ext_offchain_http_request_add_header_version_1` host function ([#1994](#1994)) ([0a30b3d](0a30b3d)) * **lib/runtime:** Implement `ext_offchain_http_request_start_version_1` host function ([#1947](#1947)) ([974b1fc](974b1fc)) * **lib/runtime:** Implement `trie_blake2_256_verify_proof` host function ([#1920](#1920)) ([506565d](506565d)) * **lib/trie:** Implement `verify_proof` function ([#1883](#1883)) ([67bb5ef](67bb5ef)) * **lib/trie:** Implement limit for trie.ClearPrefix ([#1905](#1905)) ([becec9e](becec9e)) * **lib/trie:** Parallel hash trie. ([#1657](#1657)) ([22827e7](22827e7)) * **pprof:** Pprof HTTP server service ([#1991](#1991)) ([ce24ea9](ce24ea9)) * **rpc/subscription:** implement state_unsubscribeStorage ([#1574](#1574)) ([7574f10](7574f10)) * **rpc:** Implement `childstate_getChildStorage` RPC call ([#1832](#1832)) ([3d949f2](3d949f2)) * **rpc:** Implement `childstate_getStorageHash` RPC call ([#1805](#1805)) ([e539bd3](e539bd3)) * **rpc:** Implement `childstate_getStorageSize` RPC call ([#1810](#1810)) ([a04deb6](a04deb6)) * **rpc:** Implement `payment_queryInfo` RPC call ([#1826](#1826)) ([7a5deec](7a5deec)) * **rpc:** Implement `state_getReadProof` rpc call ([#1768](#1768)) ([865f80f](865f80f)) * **runtime:** implement custom logging handler that print function name ([#1825](#1825)) ([2b1276d](2b1276d)) * **telemetry:** send telemetry messages when GRANDPA receieves commit or vote messages ([#2015](#2015)) ([7bf40e1](7bf40e1)), closes [#1840](#1840) [#1839](#1839) [#1838](#1838) * **telemetry:** send txpool.import telemetry msg ([#1966](#1966)) ([ffc81bf](ffc81bf)) ### Reverts * Revert "feat(dot/rpc) implement `author_hasSessionKeys` RPC call (#1704)" (#1714) ([65380fd](65380fd)), closes [#1704](#1704) [#1714](#1714)

github-actions · 2021-12-03T21:09:56Z

🎉 This PR is included in version 0.6.0 🎉

The release is available on GitHub release

Your semantic-release bot 📦🚀

…Message (ChainSafe#1783)

add storageState lock in core.HandleTransactionMessage

b27e3d5

noot requested review from arijitAD, EclesioMeloJunior, edwardmack, jimjbrettj and timwu20 as code owners September 9, 2021 20:35

noot self-assigned this Sep 9, 2021

noot added the Status: Review Needed label Sep 9, 2021

log for each msg

a9b7a40

noot marked this pull request as draft September 9, 2021 21:27

noot added 3 commits September 10, 2021 11:55

get best block hash at start of HandleTransactionMessage

6a9a5dc

Merge branch 'development' of github.com:ChainSafe/gossamer into noot…

4aa177b

…/fix-kusama

lint

cae4919

noot marked this pull request as ready for review September 10, 2021 15:58

noot changed the title ~~fix: fix Kusama sync - add storageState lock in core.HandleTransactionMessage~~ fix: fix Kusama sync; add storageState lock in core.HandleTransactionMessage Sep 10, 2021

noot added 2 commits September 10, 2021 11:58

fix

4c897de

fix again

1eef284

arijitAD approved these changes Sep 13, 2021

View reviewed changes

arijitAD reviewed Sep 13, 2021

View reviewed changes

timwu20 requested changes Sep 13, 2021

View reviewed changes

use sync.Locker interface

802020e

noot requested a review from timwu20 September 13, 2021 18:14

remove files

db119ba

timwu20 approved these changes Sep 13, 2021

View reviewed changes

noot merged commit 1d688e4 into development Sep 13, 2021

noot deleted the noot/fix-kusama branch September 13, 2021 21:03

timwu20 pushed a commit to timwu20/gossamer that referenced this pull request Dec 6, 2021

fix: fix Kusama sync; add storageState lock in core.HandleTransaction…

d03f6d1

…Message (ChainSafe#1783)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix: fix Kusama sync; add storageState lock in core.HandleTransactionMessage #1783

fix: fix Kusama sync; add storageState lock in core.HandleTransactionMessage #1783

noot commented Sep 9, 2021 •

edited

Loading

codecov bot commented Sep 9, 2021 •

edited

Loading

arijitAD Sep 13, 2021

noot Sep 13, 2021

timwu20 Sep 13, 2021

noot Sep 13, 2021

github-actions bot commented Dec 3, 2021

fix: fix Kusama sync; add storageState lock in core.HandleTransactionMessage #1783

fix: fix Kusama sync; add storageState lock in core.HandleTransactionMessage #1783

Conversation

noot commented Sep 9, 2021 • edited Loading

Changes

Tests

Issues

Primary Reviewer

codecov bot commented Sep 9, 2021 • edited Loading

Codecov Report

arijitAD Sep 13, 2021

Choose a reason for hiding this comment

noot Sep 13, 2021

Choose a reason for hiding this comment

timwu20 Sep 13, 2021

Choose a reason for hiding this comment

noot Sep 13, 2021

Choose a reason for hiding this comment

github-actions bot commented Dec 3, 2021

noot commented Sep 9, 2021 •

edited

Loading

codecov bot commented Sep 9, 2021 •

edited

Loading