Repository navigation
Releases
v2.2.2
Release notes
A patch release on top of 2.2.1 with the October 2026 out-of-band .NET security updates. See the Microsoft announcement for details of the fixed vulnerabilities.
There are no Nethermind code, database-format, or configuration changes, so you can upgrade in place from any 2.2.x version. Consensus behavior and fork schedules are the same as in 2.2.1.
Overview
2 changes since 2.2.1:
- .NET 10.0.14 runtime in Docker images and binaries: Docker images are now based on
mcr.microsoft.com/dotnet/aspnet:10.0.14-resolute(-resolute-chiseledfor the chiseled variant). The self-contained binaries are built with SDK 10.0.403 and bundle the 10.0.14 runtime. Docker users only need to pull the new tag (#15251) - .NET packages updated to 10.0.14: ASP.NET Core DataProtection, Microsoft.Extensions.*, System.Security.Cryptography.*, System.IO.Hashing, System.CommandLine and others (#15211)
Build signatures
The packages are signed with the following OpenPGP key: AD12 7976 5093 C675 9CD8 A400 24A7 7461 6F1E 617E
What's Changed
- chore(deps): update .NET packages by @rubo in #15211
- chore: update Dockerfiles to .NET 10.0.14 by @stdevMac in #15251
Full Changelog: 2.2.1...2.2.2
Full Changelog: 2.2.1...2.2.2
zkVM guests v2.3.0-preview
Release notes
- Heavily optimized with significant performance improvements
- Compliant with [email protected]
Guest binaries
Built from 76ff2473d895d04e55e6a056b8520d1403bad0af, with verification keys generated by ere-server 0.19.5. The SHA-256 checksums are listed below; the OpenVM ELF is the one inside its archive:
- ZisK 1.3.2-alpha
- ELF:
fd31c5afdb66c64da48afcab76a52915b2b2e5745bc93b17a462fb73b2eac41a - VK:
9223053b558bb369ab3c5cbf8dcb97dd342d0b92116c06a71e64d1fcbab7e92a
- ELF:
- SP1 6.9.0
- ELF:
9251e6af8d436ce70425c84614ea697a2af6aa1e8a1409b0dab888bc0cdef905 - VK:
bf7086f529767878f61b768c475f193d55b588da3c8663aaf1cc03ec876f59d2
- ELF:
- OpenVM 2.1.0-rv64
- ELF:
80c3422a9adfa0906762151d5887aa4fc6ee3687240d8475ba5193444ee54c09
- ELF:
Build signature
SHA256SUMS lists every file and is signed with the following OpenPGP key: AD12 7976 5093 C675 9CD8 A400 24A7 7461 6F1E 617E
Each file listed in SHA256SUMS also has a build provenance attestation: gh attestation verify <file> --repo NethermindEth/nethermind
Nethermind Bootnode 1.0.0
Nethermind Bootnode is a lightweight, standalone service for Ethereum peer discovery. It supports discv4 and discv5, keeps a persistent identity, and exposes REST, JSON-RPC, and Prometheus metrics. It helps nodes find peers; it does not replace a full Ethereum node.
Minimum requirements (Bootnode only): 1 vCPU, 1 GiB RAM, and 1 GiB free disk.
Native
On Linux or macOS, this selects and downloads the 1.0.0 archive for your OS and architecture:
case "$(uname -s)" in
Linux) os=linux ;;
Darwin) os=macos ;;
*) echo "Use the Windows x64 ZIP from the release page"; exit 1 ;;
esac
case "$(uname -m)" in
x86_64|amd64) arch=x64 ;;
aarch64|arm64) arch=arm64 ;;
*) echo "Unsupported architecture"; exit 1 ;;
esac
asset="nethermind-bootnode-1.0.0-${os}-${arch}.tar.gz"
curl -fL -o "$asset" \
"https://github.com/NethermindEth/nethermind/releases/download/bootnode-1.0.0/$asset"
mkdir -p bootnode-1.0.0
tar -xzf "$asset" -C bootnode-1.0.0
cd bootnode-1.0.0
./bootnode --data-dir ./bootnode-dataAdd Grafana
For a native Bootnode on Linux, start it with --metrics-host 0.0.0.0 so Prometheus can scrape it. Restrict access to port 6060.
./bootnode --data-dir ./bootnode-data --metrics-host 0.0.0.0In another terminal, from the extracted directory:
export GRAFANA_ADMIN_PASSWORD='choose-a-password'
docker compose -f observability/docker-compose.yml up -dOpen http://localhost:3000. Monitoring setup.
Docker
The Docker image is avaiable on Docker Hub.
docker pull nethermind/bootnode:1.0.0
docker run -d --name bootnode \
-p 30303:30303/udp \
-p 127.0.0.1:8546:8546 \
-p 127.0.0.1:6060:6060 \
-v bootnode-data:/nethermind-bootnode/data \
nethermind/bootnode:1.0.0
v2.2.1
Important
Hoodi node operators must upgrade before Mon, 26 Oct 2026, 17:42:48 UTC. This release removes the Glamsterdam fork from Hoodi. A Hoodi node still on 2.2.0 will activate Glamsterdam at that time and fork off the network. Your consensus client must also be on a release that removes the Glamsterdam fork from Hoodi.
Release notes
A patch release on top of 2.2.0 that removes the Glamsterdam (Amsterdam) fork from Hoodi, which was planned for Mon, 26 Oct 2026, 17:42:48 UTC (timestamp 1793036568).
The fork is on hold on Hoodi following the security concerns discussed on ACDE #248. A new date will come in a later release.
This is the only change. There are no database-format or configuration changes, so you can upgrade in place from any 2.2.x version. No other network is affected.
Overview
1 change since 2.2.0:
- Glamsterdam removed from Hoodi: Hoodi stays on BPO2 (fork hash
0x23aa1351) with no next fork announced (#15110)
What's Changed
- fix(specs): deschedule Hoodi Amsterdam fork by @stdevMac in #15110
- chore: update Directory.Build.props for 2.2.1 by @stdevMac in #15126
Full Changelog: 2.2.0...2.2.1
Build signatures
The packages are signed with the following OpenPGP key: AD12 7976 5093 C675 9CD8 A400 24A7 7461 6F1E 617E
zkVM guests v2.3.0-preview
Release notes
- Heavily optimized with significant performance improvements
- Compliant with [email protected]
Guest binaries
Built from 7aa160726f37a5f1fc5d1d66a559ef087166f3a0, with verification keys generated by ere-server 0.19.3. The SHA-256 checksums are listed below; the OpenVM ELF is the one inside its archive:
- ZisK 1.3.2-alpha
- ELF:
bf0eed7fda86865b0757d40e1633f2c38ff361314eafd69aed15837c9d1d5796 - VK:
cef9a5d84730957dca61315e1c1c5eb9fce02ee5fc481a1c94bab73014cb1012
- ELF:
- SP1 6.6.0
- ELF:
c569d8800b94987d72b884b090263eb7454dc02a1d5815cb8f4a03c807542969 - VK:
d9f252ef2137b035284e550e00fe76cee8444298b0287b7cd8c04f9dab8ff917
- ELF:
- OpenVM 2.1.0-rv64
- ELF:
246dc90e68c294ccae54c3479171d0f33272fbacd4b1dc75acaa4f51a25da390
- ELF:
Build signature
SHA256SUMS lists every file and is signed with the following OpenPGP key: AD12 7976 5093 C675 9CD8 A400 24A7 7461 6F1E 617E
Each file listed in SHA256SUMS also has a build provenance attestation: gh attestation verify <file> --repo NethermindEth/nethermind
zkVM guests v2.3.0-preview
Release notes
Temporary release for testing purposes.
Guest binaries
Built from b8407de1accde001365d649a57ce5f2b03c19ff8, with verification keys generated by ere-server 0.19.2. The SHA-256 checksums are listed below; the OpenVM ELF is the one inside its archive:
- ZisK 1.3.2-alpha
- ELF:
211187394623cc32fdff34d6f6af64ddd54eea9e72db42365914beb2c08b5d58 - VK:
e05ddc0d95931124426c7e7ab0d34fbd928810fceb6012a7fb3139b8447078e4
- ELF:
- SP1 6.6.0
- ELF:
e102b8ad12149b550eb880b0240ef4be0c52cf6c867e9b2a388c6783c2a19c36 - VK:
f230681d9607bdf2756d8219f016393f04c537fbc32262fe0b1d507dfe262d64
- ELF:
- OpenVM 2.1.0-rv64
- ELF:
9c100c3779a18e2b8a97b1f5c402b2afeb7a4afeaef82ec306e25be44dbc24b4
- ELF:
Build signature
SHA256SUMS lists every file and is signed with the following OpenPGP key: AD12 7976 5093 C675 9CD8 A400 24A7 7461 6F1E 617E
Each file listed in SHA256SUMS also has a build provenance attestation: gh attestation verify <file> --repo NethermindEth/nethermind
v2.2.0
Release notes
586 merged pull requests since 2.1.0, from 31 contributors.
Important
Hoodi: Glamsterdam activates on Mon, 26 Oct 2026, 17:42:48 UTC (timestamp 1793036568, epoch 132352). Hoodi nodes must run 2.2.0 or later before then. Your consensus client must support Glamsterdam on Hoodi as well.
Note
Upgrading from 2.0.x or 1.39.x? The breaking changes below cover 2.1.0 → 2.2.0 only. Apply the 2.1.0 breaking changes first (and, from 1.39.x, the 2.0.0 breaking changes and upgrade checklist).
For most nodes this is an in-place upgrade with no resync and no config change. Read Breaking changes first if you set JsonRpc.Host / JsonRpc.EngineHost to a host name, run with Init.DiagnosticMode=VerifyTrie or FlatDb.ImportFromPruningTrieState=true, use KeyStore.BlockAuthorAccount, pipeline requests over WebSocket or IPC, consume Parity-style trace_* output, or build plugins against Nethermind.
Highlights
- Amsterdam (Glamsterdam) on Hoodi. The fork is scheduled at
1793036568(Mon, 26 Oct 2026, 17:42:48 UTC), and Hoodi's chainspec now carries its correct beacon-chain genesis timestamp. (#14489, #13732) - Catch up finalized blocks from block access lists. On chains with EIP-7928, a node that is behind rebuilds the state of consensus-finalized ancestors from their verified block access lists and receipts instead of executing them, and executes only the unfinalized tail. It is on by default (
Sync.ReconstructFinalizedStateFromBlockAccessLists=true). Missing data or a commitment mismatch falls back to execution. (#14157) - Faster tracing on Amsterdam blocks. A single-transaction trace (
debug_traceTransaction,trace_transaction,trace_replayTransaction) on a block with a block access list no longer replays the transactions before the target, and whole-block traces run in parallel. All parallel block tracing now shares one node-wide budget,JsonRpc.TraceBlockParallelism(default 4). (#13906, #14363) - Blocks that load many large contracts. Glamsterdam allows 64 KiB contracts and 200M-gas blocks, so one block can touch ~58k distinct maximum-size contracts. The code DB moves from Snappy to LZ4 (mean block time on such a devnet range 2.6 s → 1.7 s, no block ≥ 6 s), missed code is copied once instead of twice (allocations per block 7.17 → 3.59 GB), and a block keeps the code it loaded for the rest of the block. (#13907, #14086, #14322)
- Lower
engine_newPayloadlatency. Storage trie nodes on the flat backend are written after the block is reported valid, committed storage writes are applied on an idle thread during execution, block-processing stages share one worker budget, sender recovery no longer competes with the transactions root, Kestrel reads large payload bodies in 64 KiB blocks, and the payload pipeline is warmed before RPC opens (first-payload latency 66.8 ms → 5.4 ms in the PR's measurement). (#14058, #14059, #14182, #14164, #14213, #14443, #13931, #14481) - Prewarmer hand-off. Block processing takes over a transaction the prewarmer already executed when its recorded footprint matches the state the block has reached, instead of executing it again (hidden
Blocks.PreWarmHandoff, on by default; not used for blocks with a block access list or under EIP-8037). (#14312, #14476) - Parity-style trace standardisation. A long series of
trace_*fixes aligns Nethermind with the draft trace spec (execution-apis#890) and with Erigon/Reth: failed precheck frames, reverted-frame results, Parity error labels,trace_filtermode/blockHash,trace_getbytraceAddresspath,nullfor unknown transactions, and stricter parameter validation. Several of these change output; see Breaking changes. (#13957, #13981, #13857, #14111, #13858, #13937) - Geth-compatible debug tracing.
keccak256PreimageTracer, Geth opcode-logger byte limits,debug_traceCalltxIndex, block-widecallTracerlog indices, corrected trace-file records, and per-request V8 runtimes so JavaScript traces run in parallel. (#13662, #13670, #13700, #13693, #13709, #13727, #13737, #13541) - Faster JSON-RPC serialization. Blocks, RPC transactions and logs are written by source-generated writers (full
eth_getBlockByNumberwith 200 txs −13% time / −93% allocations;eth_getLogsof 1,000 logs −14%), and public WebSocket/IPC connections process up to 16 requests concurrently. (#14299, #14345) - New
mainnet_aztecconfig with a rolling 4,096-block trace window on the flat-history backend and the per-transaction trace index. (#14038) - SendPolicy plugin (opt-in). Checks every transaction submitted through the node's JSON-RPC against an owner-written rule file (allowed senders/destinations, token-call allow-lists, fee and spend caps) and refuses the ones that break it. Off by default (
SendPolicy.Enabled=false). (#14396) - Next-fork prototypes (not scheduled on any public network). EIP-8141 frame transactions with EIP-8250 keyed nonces and EIP-8272 recent roots, EIP-7979, EIP-8131, EIP-3298, EIP-7668, EIP-8253, EIP-8279. They activate only through explicit chainspec/genesis keys or the Bogota fork alias. (#12526, #13932, #13929, #13928, #13934, #13943, #13942, #14386)
Performance
- Code DB on LZ4: on a Glamsterdam devnet range of blocks that each read ~3.7 GB of 64 KiB code, mean block time fell from 2.58–2.63 s to 1.70–1.72 s and the slowest block from 7.8 s to 5.3 s (measured on LZ4-recompacted files); point reads of large values 88–93 µs → 47–49 µs; the code DB shrank 475 MB → 303 MB. New and compacted SST files use LZ4. (#13907)
- Large-contract blocks: missed code is copied once (−31% block time, −50% allocations on a 58k × 64 KiB block); code reads no longer fill the RocksDB row cache, which is turned off for the code DB. (#14086)
- Payload start-up: median first-payload latency 66.8 ms → 5.4 ms with the pipeline warm-up (process start to first VALID +0.8 s; disable with
Init.PipelineWarmupEnabled=false). The ~2 s KZG trusted-setup load is no longer on the startup critical path. (#13931, #13998) - JSON-RPC: generated writers for blocks/transactions/logs (−13–27% time, up to −93% allocations), fewer passes over request and batch bodies, allocation-free filter binding,
newHeadsserialized once for all subscribers. (#14299, #14187, #14197, #14198, #14195, #13894) - RLP: built-in transaction types encode/decode without virtual dispatch (legacy encode −29%, EIP-1559 encode −11%). (#14455)
- EVM: a long series of interpreter micro-optimisations (frameless MLOAD/MSTORE, padded code copy without end-of-code checks, cheaper DUP/EXP/SUB/SIGNEXTEND, per-depth call frame reuse, cached EIP-7702 delegation targets). (#14028, #13896, #13893, #13878, #13802, #14139, #13799 and others)
- Flat backend reads: a bloom filter lets read-only execution (
eth_call,eth_estimateGas,eth_createAccessList,eth_simulateV1, traces) skip in-memory layers for slots none of them wrote (FlatDb.InMemorySnapshotBloomBitsPerKey, default 14). The carry-forward slot cache becomes a fixed 64 MB set-associative table. (#14094, #14330, #14027) - Tracing: streamed struct logs flush every 1 MiB so a slow reader no longer holds hundreds of MB per request;
trace_filterstops oncecounttraces are collected. (#14461, #13890) - Networking: inbound transaction admission is budgeted before decoding (16 MiB per peer, 64 MiB total), oversized broadcast transactions are skipped before decoding, eth/snap response request IDs are checked before decoding, and discovery runs on a plain UDP socket instead of DotNetty. (#14434, #13506, #14184, #14346)
- Transaction pool: an underpriced replacement of a set-code, blob or frame transaction is rejected before its authorization signatures are recovered or its blob proofs checked. (#14577)
- zkVM guests: dozens of ZisK/SP1 guest optimisations (precompile-backed 256-bit arithmetic, Keccak and SHA-256, register-based dispatch, cheaper trie and witness handling) and tests-zkevm v21.0.1. (#14232, #14217, #13975, #14402, #14118 and others)
⚠️ Breaking changes
If you run a node (config and startup)
JsonRpc.Host,JsonRpc.EngineHostandAdditionalRpcUrlshost names are resolved and bound to their addresses. Before, any host that was not an IP literal orlocalhostmade Kestrel listen on all interfaces. Now the node listens only on the addresses the name resolves to; if it cannot resolve or resolves to a wildcard address, JSON-RPC does not start (the node keeps running and logs the host). Migration: behind NAT or with public DNS, configure a local interface IP; with Docker service names, point health checks at the bind address; to keep listening on every IPv4 interface, set the host to0.0.0.0explicitly. IP literals,localhost,*and+are unchanged. (#13746)Init.DiagnosticMode=VerifyTrieandFlatDb.ImportFromPruningTrieState=truenow run only their step and exit, instead of doing the work mid-boot and continuing into a normal start. Equivalent commands:nethermind verify-trieandnethermind import-flat-db.verify-triereports its verdict through the exit code (and fails on a DB with no head); the import fails if there is nothing to import. Migration: run the import/verification as a one-off, then remove the setting before starting the node normally. (#13715)- **
KeyStore.BlockAuthorAccountwhose key cannot be loaded is now a configura...
v2.1.0
Release notes
359 merged pull requests since 2.0.0, from 29 contributors.
Important
Sepolia: Glamsterdam activated on Tue, 06 Oct 2026, 13:53:36 UTC. Sepolia nodes must run 2.1.0 or later.
Note
Upgrading from 1.39.x? The breaking changes below cover 2.0.0 → 2.1.0 only. Apply the 2.0.0 breaking changes and upgrade checklist first; they are the complete 1.39.x → 2.0 list.
For most nodes this is an in-place upgrade with no resync. Read Breaking changes first if you run on the flat backend with a custom sync config, run poacore, run a Flashbots relay, parse Parity-style traces, or build plugins against Nethermind.
Highlights
- Glamsterdam on Sepolia. Final EIP-8037 state-gas rules, the
TX_MAX_TOTAL_GAS_LIMITtransaction cap, and Engine API alignment for Amsterdam. (#13140, #13484, #13627, #13554, #13583, #13586, #13755) - Faster EVM. The interpreter loop is replaced by tail-call chain dispatch, stack words stay in
UInt256limb layout, storage values stay asUInt256until they are persisted, and JUMPDEST analysis is rewritten. (#13128, #13323, #13420, #13189, #13810) - Lower
engine_newPayloadlatency. The response goes out as soon as the block has executed and its roots match, before the commit.engine_forkchoiceUpdatedstill waits for that head's commit. GC pauses no longer stall payloads the way they did. (#13647, #13711, #13476, #13334) - Per-transaction changeset index for flat archives (opt-in). A single-transaction trace starts at its target transaction instead of replaying the whole block before it:
debug_traceTransactionwent from ~7.1 s to ~28 ms in the PR's measurements. Whole-block traces run their transactions in parallel (6–10× faster). Enable withFlatDb.HistoryTransactionIndexEnabled=true. (#13454, #13426) - BAL healing after snap sync (flat backend). On the flat backend, the state left by snap sync is healed by replaying block access lists instead of downloading trie nodes. It is on by default (
Sync.BalHealing=true); set it tofalseto go back to trie-node healing. snap/2 is advertised only when BAL healing can run. (#12945) - Dual-stack IPv4/IPv6 networking. RLPx, discv4 and discv5 listen on both families, ENRs advertise both, and there is a new
Network.ExternalIpV6override. (#12679, #12982, #12983, #13731) - Flat backend safety rails. A flat DB that RocksDB had to auto-repair is now re-synced instead of trusted. Unsafe flat activations are refused at start instead of silently corrupting or discarding state. (#13645, #13644)
- Reclaim the patricia trie after a trie→flat conversion.
FlatDb.DropPruningTrieState=truedeletes the old trie store on the next start once the flat DB owns the state (~227 GB on mainnet). Off by default. (#13572) - RocksDB 11 through Nethermind's own bindings. The RocksDB package is replaced with
Nethermind.RocksDbBindings(RocksDB v10 → v11): source-generated P/Invoke with no marshalling stubs, and every native handle held in aSafeHandle. (#12853) - New EF bootnodes. The EF bootnode entries for mainnet, Sepolia and Hoodi are replaced with the new NodeOps fleet: five dual-stack nodes running both discv4 and discv5. (#13313)
- More zkVM targets. Stateless guests for SP1 and OpenVM ship alongside ZisK, and the ZisK toolchain moves to 1.3.0-alpha. (#13435, #13769, #13743)
Performance
- EVM dispatch and call frames: 3.8–4.5% faster block processing on the superblocks, realblocks and fusaka benchmark sets. (#13189)
- Storage values held as
UInt256: SLOAD 9–18%, SSTORE 42% and TSTORE ~50% faster, with no allocations on those paths. (#13420) - EVM stack in limb layout: 45–53% faster ADD, 16–36% faster SUB, and 9–18% faster shifts. (#13323)
- JUMPDEST analysis: worst-case analysis of 64 KiB initcode (a 200 MGas block of PUSH1/JUMPDEST patterns) drops from 0.4–3.2 s to 22–153 ms, depending on the SIMD tier. Initcode is analysed on first use. (#13810)
- Trie and storage-key hashing: 35–80% less time on transaction and receipt-root workloads. (#13552)
- Prewarming: a single fan-out while senders arrive gives superblocks processing a median −4.4%, P99 −27.9%, and +7.6% MGas/s. Prewarming and trie concurrency caps are removed, and parallel work is budgeted across the state root, blooms, the receipts root and prewarming (2–4% faster processing). (#13640, #13388, #13770)
- Carried caches: pre-block account and storage caches are carried across blocks. (#13100)
- Discovery traffic: on a warm-started Hoodi node, discovery traffic fell by 90% (152 KiB/s → 15 KiB/s), and the standalone bootnode uses ~44% less working set. (#13122)
- Snap sync: downloads are bounded and overlap with response processing. Whale-storage verification in
FlatTrieVerifieris parallelised. (#12623, #12615) - RPC:
eth_call,eth_estimateGasandeth_createAccessListno longer hash the transaction on every run (~0.2 ms on large calldata) (#13804). The stream interpreter is off by default because it lost on every measuredeth_callshape.eth_getFilterChangesis served from shared block logs. (#12999, #13742) - Intel hybrid CPUs: block processing is pinned to the performance cores by default (
Blocks.ProcessingCores). Nothing changes on other CPUs. (#13717)
⚠️ Breaking changes
If you run a node (config and startup)
- poacore / poacore_validator now ship
FlatDb.Enabled=falsebecause they use FastSync without SnapSync (#13644). A node on those configs that synced on 2.0.0 is on the flat backend, since flat is 2.0.0's default. On 2.1.0 it refuses to start (Refusing --FlatDb.Enabled=false on an existing flat DB). Either pass--FlatDb.Enabled=true, or delete theflat,flatHistoryandpersistedSnapshotdirectories and resync the state. --FlatDb.Enabled=falseon an existing flat DB is refused. In 2.0.0 the node silently discarded the state and resynced (#13428). It now stops with an error. (#13644)- Flat +
Sync.FastSync=true+Sync.SnapSync=falseis refused on a fresh node. Legacy tree sync on the flat backend can leave holes in the state. An already-synced node keeps running and logs a warning. The exception is importing from patricia (FlatDb.ImportFromPruningTrieState=trueon a datadir that holds patricia state). (#13644) - New
FlatDb.OnRepairsetting, defaultResync. If RocksDB auto-repairs the flat DB, the node now wipes the flat state and runs state sync again. Headers, bodies and receipts are kept. Archive nodes synced from genesis cannot snap-resync. SetFlatDb.OnRepair=Ignoreand plan a resync yourself if a repair is ever reported. (#13645) Receipt.MaxBlockDepthdefault raised back from 1,000 to 10,000, and the limit is now also enforced oneth_getFilterLogs. On nodes with the log index, only blocks the index cannot serve count towards the limit. A filter created byeth_newFilterover a wider range is still accepted, buteth_getFilterLogson it fails with-32602. (#13131)- Removed config keys:
FlatDb.EnablePreimageRecording,Db.PreimageDbRocksDbOptions,Db.PreimageDbAdditionalRocksDbOptions(#13690),LogIndex.VerifyRpcResponse(#13131), andGrpc.Enabled,Grpc.Host,Grpc.Portalong with the gRPC endpoint (#13330). A JSON config that still sets them logsInvalid configuration settingsat startup. Remove them from configs, environment variables and command-line flags. Merge.CompactMemorydefault changed fromYestoNo. This removes GC compaction stalls fromengine_newPayload. Steady-state RSS may be somewhat higher. (#13476)Init.LogRulesno longer routes to Seq, andSeq.MinLevelis now the floor for the Seq target. To send a namespace's Trace output to Seq, add an explicitwriteTo="seq"rule inNLog.config. (#13419)- Dual-stack by default. With
Network.LocalIpunset, the P2P TCP and discovery UDP listeners now try to bind[::]on Linux and Windows, so the P2P port may now be reachable over IPv6. Check your firewall. macOS stays IPv4-only. SetNetwork.LocalIp=0.0.0.0to keep IPv4-only. (#12982) Pruning.FullPruningTrigger=VolumeFreeSpacewith a threshold too low to ever trigger now logs a startup warning. Automatic pruning was already not running in that case, so nothing new is broken. (#13485)
If you run a Flashbots relay
Flashbots.EnableValidationnow defaults totrue.flashbots_validateBuilderSubmissionV3checks the submitted block's state root, receipts root, gas used and logs bloom against its execution. Set it tofalseexplicitly to get the old behaviour. (#13452)
If you integrate over JSON-RPC / Engine API
eth_getHeaderByNumber("pending")returnsnull, matching go-ethereum and execution-apis#877. (#13082)proof_getTransactionByHash/proof_getTransactionReceiptreturnresult: nullfor a transaction they cannot serve, instead of a-32000error.proof_getTransactionReceiptreturns-32002when the block can no longer be retraced. (#13366)- Parity
vmTracestack pushes are hex quantities (0x0,0x1), not byte strings. Word operations now report their pushes. (#13750, #13487) - Other trace output changes. Empty trace selections are accepted.
trace_getkeeps lookup errors instead of replacing them.trace_filterreturns an error when the requested history is unavailable, and bufferedtrace_filternow applies a singleJsonRpc.Timeoutto the whole request. Deleted-account fields carry deletion markers.callTraceromitstoon a failed top-level CREATE. (#13667, #13676, #13677, #13668, #13078) - **EIP-3155 /
structLogstraces are corrected...
zkVM guests v2.1.0-unstable+5c2a2768
Release notes
SP1 and OpenVM support
This release introduces initial support for SP1 and OpenVM zkVMs.
Updates and improvements for ZisK
In addition to ZisK v1.3.0-alpha support, this release brings impressive performance improvements:
| Block 25,532,382 | Old guest | This release | Change |
|---|---|---|---|
| Total cost | 75.66B | 36.46B | −51.8% (2.1× cheaper) |
| Steps | ~704M | 305.7M | −56.6% (2.3× fewer) |
Guest binaries
Built from 5c2a2768fdac1930af35aadf609b61bc6f4af8ee with the following SHA-256 checksums:
| zkVM | SHA-256 |
|---|---|
| ZisK | 69b7975c8fcbd3604984018870632336b412c9772c374d2e593298c928c9348d |
| SP1 | efddc2291fdfa64949ce115791b74ce454f7bd6bf1248a018e5b68bb8d5e66e2 |
| OpenVM | 79a7cd7509b10b339d372711186764024f97be1f307b892351d0fc7614ac35b6 |
Build signature
The packages are signed with the following OpenPGP key: AD12 7976 5093 C675 9CD8 A400 24A7 7461 6F1E 617E
v2.0.0
Release notes
669 merged pull requests since 1.39.3, from 39 contributors.
Nethermind 2.0.0 is a hardened, faster, archive-capable execution client, built for Ethereum, its L2s, and beyond. It consolidates everything shipped in 2.0.0-rc and 2.0.0-rc2.
Most nodes upgrade in place with no resync. There are breaking changes — read Breaking changes before upgrading, especially if you run an archive node, set custom config keys, or build against Nethermind as a library or plugin.
Important
Upgrading from 1.39.x? This page is the complete list of breaking changes from 1.39.x to 2.0. It replaces the 2.0.0-rc and 2.0.0-rc2 notes; the 2.0.0-rc release page has been removed. Go through Breaking changes and the Upgrade checklist below.
Going straight to 2.1.0 or later? Apply this list first, then the 2.1.0 breaking changes.
⚠️ Breaking changes
If you run a node (config)
Db.FlushOnExitis now an enum, not a boolean. Values:None,WalOnly,Full. The default changes from full-flush toWalOnly. If your config, env var, or CLI still setstrue/false, the node fails to start — changetrue→Full,false→None.Blocks.PreWarmStateOnBlockProcessingwas replaced byBlocks.PreWarming. New enum values:None,Block,BlockAndMempool(defaultBlockAndMempool). The old key is now silently ignored (you'll get a warning, not a crash) — rename it or you lose your setting.- Numeric config keys are now unsigned. Negative values are rejected and the node fails to start. The one common case:
Receipt.TxLookupLimit: -1(the old "never index" sentinel) must become18446744073709551615. Receipt.MaxBlockDepthdefault lowered from 10000 → 1000. A singleeth_getLogsspanning more than 1000 blocks is now rejected unless the log index is enabled. If you rely on wide ranges: raiseReceipt.MaxBlockDepth, split the query, or enable the log index. 2.1.0 restores the 10,000 default (#13131).FlatDb.HistoryRetentionis required whenever you set a window size or a since-block.FlatDb.HistoryRetentionBlockson its own fails to start. A window size without a mode,Rollingwithout a size,SinceBlockwithout a block, or both numbers at once, are all refused.- A configured
FinalTotalDifficultyis no longer treated as "TTD reached" (#12555). Custom-FTD configs that relied on that inference will wait for the real TTD. - sepolia and hoodi default configs are now discv5-only (hoodi's static bootnode list was removed). Stock-config nodes on those networks discover peers via discv5 + DNS. Custom
DiscoveryConfigis unaffected. - Mainnet ancient-history barrier raised (block 15537394 → 24600000). Existing nodes keep all their data. Only new or resyncing nodes retain less pre-24.6M history by default.
If you run an archive node
- Archive configs now enable the flat-database backend with history (see New archive layout below). Your existing patricia archive DB keeps working (the new flag is ignored on an existing DB — nothing corrupts). To actually run the new flat archive you must resync, or set
FlatDb.ImportFromPruningTrieState=truefor an in-place trie→flat import. Enabling flat history is forward-only — there is no backfill below the point you enable it.
If you integrate over JSON-RPC / Engine API
engine_newPayloadV5: a malformed block-access-list envelope now returns{"status":"INVALID", ...}(HTTP 200) instead of JSON-RPC error-32602. (#13012)trace_simulateV1: the response fieldstateChangeswas renamed tostateDiff. (#12892)- Receipt JSON on the
t8n/ direct-serialization path now emitsrootorstatus(mutually exclusive), not both. The standardeth_getTransactionReceiptpath is unchanged. (#12867) debug_traceCallnow uses a base fee of 0 for unpriced calls, matching go-ethereum. Reported gas figures change accordingly. (#12804)eth_maxPriorityFeePerGasreturned a gas price; it now returns a priority fee. (#13129)eth_getStorageAtposition parsing is stricter — spec-compliant hex-quantity clients are fine, but a client sending a bare number or un-prefixed string may now get a parse error. (Verify against your integrations.)- Amsterdam-gated (not active on mainnet yet, relevant if you test against devnets):
eth_callnow returns errors (out-of-gas, address collision, EF-prefix) where it previously returned0x;eth_estimateGasandeth_createAccessListreturn updated values. (#12806, #12971, #12970)
If you build against Nethermind (plugin / library authors)
- Many
IConfignumeric properties changed fromlong/inttoulong. This is source-breaking — recompile your plugins against 2.0.0.
Consensus / protocol
- EIP-7610 support removed. (#12935) No change on any live network — the storage-collision check is superseded by EIP-684 and was never activated as a fork. Only affects state-test / collision fixtures.
- Stricter RLPx/Snappy frame validation. (#12896) Malformed frames are now rejected/disconnected. Conformant peers (Geth, Besu, older Nethermind, CL relays) are unaffected — this is DoS hardening.
- Optimism
engine_signalSuperchainV1V0 byte layout corrected (7→8 bytes) to match op-geth/op-node. (#12805) OP-Stack operators only.
Dropped
ethernity-sepolianetwork removed (ethernity-mainnetis retained).
Upgrade checklist
From 1.39.x
- Stop the node. No resync is needed for a normal full node, and your existing patricia database stays patricia — it is detected at start.
- Audit your config for the breaking keys above. The ones that crash on start are
Db.FlushOnExit(bool→enum), any negative numeric value,Receipt.TxLookupLimit: -1, andFlatDb.HistoryRetentionBlockswithoutFlatDb.HistoryRetention. - Rename
Blocks.PreWarmStateOnBlockProcessing→Blocks.PreWarming(otherwise it is silently ignored). eth_getLogsusers: check whetherReceipt.MaxBlockDepth(now 1000) affects your queries.- Archive operators: decide between keeping your existing patricia archive (works as-is) or resyncing/importing to the new flat archive with history.
- Plugin/library authors: recompile against 2.0.0 (
IConfignumeric types changed). - Fresh nodes will serve snap. Turn it off with
Sync.SnapServingEnabled=falseif that is not what you want. - Start the node and confirm
2.0.0in the startup logs.
From 2.0.0-rc / 2.0.0-rc2
No database-format change and no resync. Do step 2 above (the FlatDb.HistoryRetention requirement arrived in rc2), then start.
Do not set --FlatDb.Enabled=false on an existing flat datadir — see Known issues.
Known issues
These were open at release against the flat backend, which is the default for a new database. A node that upgraded in place from 1.39.x keeps its patricia backend and is not exposed to any of them. Where a later release addresses an item, it is noted.
- A snap sync on the flat backend cannot resume across a restart. Every run clears the flat columns, so restarting mid-sync discards the state downloaded so far and starts again from zero. Plan an uninterrupted initial sync, or sync on
--FlatDb.Enabled=false(patricia/half-path), which does resume. A node that has reached the tip never re-enters snap, so this affects bring-up only. (#13601) - Copy-style full pruning is not available on the flat backend. A flat node accepts
Pruning.Mode=Hybrid,Pruning.FullPruningTrigger, andPruning.FullPruningThresholdMbwithout warning but never prunes, andadmin_prunereportsdisabled. If you depend on periodic full pruning, stay on the patricia backend for now. In-place upgrades from 1.39.x keep patricia and are unaffected. (#13600) From 2.1.0 the node logs a warning when an automatic full-pruning trigger is set on flat (#13644); pruning itself is still patricia-only. - Legacy (non-snap) state sync on the flat backend can produce a bad state. With
Sync.SnapSync=false, a restart during state sync can leave permanent account/storage holes (the trie is complete, the flat leaf is missing), and pivot changes can leave stale storage subtrees. Both surface later as a deterministicHeaderGasUsedMismatch. Use snap sync (the default), or resync on the patricia backend. (#13575, #13576) From 2.1.0 a fresh node refuses flat +Sync.FastSync=true+Sync.SnapSync=false(#13644). - A RocksDB auto-repair of a corrupted flat DB can silently drop leaves. The node then resumes as "already have state" and diverges from consensus. If RocksDB reports that it repaired your database, resync rather than continue. (#13577) Fixed in 2.1.0: a repaired flat DB is re-synced by default (
FlatDb.OnRepair=Resync, #13645). --FlatDb.Enabled=falseon an existing flat datadir silently discards the state and full-resyncs. The node does not refuse. The inverse (a patricia DB with the new flat default) is detected and protected. Do not pass that flag to an existing flat node. (#13428) Fixed in 2.1.0: the node refuses to start instead (#13644).
Two smaller JSON-RPC rough edges were also open at release: guard-rail rejections (-32002, -32005, 4444) still log one WARN per request (#13602), and "no state for that block" is answered with different error codes depending on the method (#13603). Both fixed in 2.1.0 (#13643).
Highlights
- Flat DB is the default state backend (#11931). `Fla...