- rpc:
eth_getFilterLogsnow runs a historical query from the criteria stored byeth_newFilter; it neither drains the live-log queue read byeth_getFilterChangesnor refreshes the filter expiry deadline. Historical queries enforcerpc.blockrange.limit,rpc.logs.maxresults, andrpc.logs.querylimit; a filter can return-32602when one of these limits is exceeded, and queries may surface initialization, pruned-history, or not-yet-executed errors. This can be a breaking change (#23296) — by @taratorio - rpc:
eth_newFilterandeth_subscribe("logs")now reject criteria exceeding the per-filterrpc.subscription.filters.maxaddressesorrpc.subscription.filters.maxtopicslimit with-32602instead of silently capping them. This can be a breaking change for nodes that configure either limit above0; both limits default to0(unlimited) (#23296) — by @taratorio - rpc:
eth_newFilterandeth_subscribe("logs")now reject criteria with more than four topic positions with-32602instead of accepting a filter that cannot match any Ethereum log. This can be a breaking change (#23296) — by @taratorio - rpc: the
callTracerwithLogoption ofdebug_traceTransactionanddebug_traceBlock*now numbers a log'sindexover the whole block instead of restarting at0x0on every transaction, matching geth'scallTracerand the index the same log carries ineth_getLogsand in its receipt. Indebug_traceCallManythe calls of one bundle now share that count instead of each restarting, while every bundle still starts at zero, since a bundle is a simulated block of its own;debug_traceCallis unchanged. On a node whose receipt domain has no record for a transaction,eth_getTransactionReceiptandeth_getLogsnow return an error instead of a receipt whose logindexhad underflowed. This can be a breaking change for clients that join trace logs toeth_getLogsindices (#23797) — by @lupin012
--witness.cache.blocks,--witness.cache.head-capture, and--witness.cache.maxmbenable an eager in-memory cache of recent-block legacydebug_executionWitnessresults, keyed by block hash. Head-capture mode lets a minimal node (no commitment-domain history) serve witnesses for the last N head blocks cache-only — a miss returns out-of-window rather than recomputing from history. Newwitness_cache_*metrics track hits, misses, builds, and resident entries. Embedded RPC only — by @awskii
- rpc: live logs returned by
eth_getFilterChangesandeth_subscribe("logs")now includeblockTimestamp, matchingeth_getLogsandeth_getFilterLogs(#23296) — by @taratorio - node: RPC HTTP gzip compression moved from the cgo-based
go-libdeflateto the pure-Goklauspost/compress, and fully buffered (one-shot) responses now compress atBestSpeed(level 1) instead of level 6; streaming responses were already atBestSpeed. Large one-shot responses gain ~10-14% latency and roughly half the compression CPU, in exchange for slightly larger compressed bodies (+2-8% wire size — relevant for operators who pay for egress). Also removes the libdeflate compressor pool and itslibdeflate_pool_*metrics, eliminating the cgo leak class fixed in #22700 (#22882) — by @lupin012 - rpc:
erigon_getLogsanderigon_getLatestLogsrenamed the log fieldtimestamptoblockTimestamp, matching the nameeth_getLogsand the other log-returning methods already use. Clients that read thetimestampkey must be updated (#23337) - by @Sahil-4555
v3.6.1 is a bugfix and security release recommended for all users. It fixes a downloader regression that could silently replace a valid local snapshot (#23859), a peer-ban bypass that let a banned Caplin peer reconnect immediately (#23866), an index-build defect that could leave stale bits in a recsplit index after a salt-collision retry (#23858), and bumps google.golang.org/grpc for two HIGH-severity CVEs (#23888). It is a drop-in upgrade from 3.6.0 — no re-sync required.
Bugfixes
- db/downloader: keep local snapshot data once the initial download is complete (#23859) by @AskAlexSharov — a locally-differing
.segthat no longer matched its.torrentwas silently replaced even afterpreverified.tomlexisted; it is now kept and registered for seeding, with a warn-level log when a mismatched file is still replaced. Fixes #21522. - cl/sentinel: honour the peer ban on inbound connections (#23866) by @lystopad —
BanStatuswas only consulted on outbound dials, so a peer already three-strike banned still received a fresh handshake on every inbound reconnect; one Gnosis archive node logged 13,827 failed handshakes against 1,638 banned peers in ninety minutes. Fixes #23605. - db/recsplit: reset the Golomb-Rice encoder, Elias-Fano structure, and existence filter on
ResetNextSalt(#23858) by @AskAlexSharov — retrying an index build after a salt collision left the failed attempt's encoded bits and existence-filter data in place, corrupting the rebuilt index. - sd: lock
changesetMuon everyDomainPut, not only on field reads (#23680) by @AskAlexSharov — the mutex guarded reads of change-set state but not the writes that mutate it, leaving a data race under concurrent commits. - rpc/jsonrpc: report base-fee sub-pool transactions as pending (#23647) by @lupin012 — transactions parked in the base-fee sub-pool were reported as queued instead of pending, understating the pending count seen by clients and tooling.
- cl/beacon: report a syncing node as 503, not 500 (#23608) by @lystopad — the committee-subscription and duties endpoints answered
500 Internal Server Errorwhile starting up or catching up, where beacon-APIs specifies503 CurrentlySyncing; alerting on 500s no longer fires on ordinary startup or resync. - cl: seed the ENR custody group count instead of an empty entry (#23591) by @lystopad — nodes advertised an empty PeerDAS custody-group-count ENR entry, making them unselectable as custody peers by counterparties that filter on it.
Security
- p2p/discover: require a proven bond before accepting a PING endpoint statement (#23639) by @yperbasis — an unbonded PING could feed an arbitrary claimed address into a node's local endpoint predictor; a statement is now only accepted once a recent PONG has proven return reachability.
- build: bump
google.golang.org/grpcto v1.83.2 (#23888) by @lystopad — CVE-2026-84304: fragmented HTTP/2 DATA frames were each stored as a separaterecvMsg, letting an unauthenticated peer exhaust heap memory with many one-byte frames across concurrent streams. CVE-2026-84445: an xDS gRPC server panicked on a request missing both:authorityandHost; not reachable in Erigon, which never callsxds.NewGRPCServer().mainalready carries v1.83.2 (#23267) and needs no code change, only this note.
Improvements
- cl, cmd/utils: derive the data-column retention window from the chain config instead of a hardcoded constant, guarding the derivation against zero and overflow inputs (#23851, #23863) by @lystopad — mainnet retention grows from 131072 to 131104 slots (one extra epoch of alignment slack); Gnosis and Chiado drop from 131072 to 65552.
Full Changelog: https://github.com/erigontech/erigon/compare/v3.6.0...v3.6.1
Erigon 3.6.0 is headlined by more reliable Caplin block production, pruned nodes reclaiming old snapshots, and
plain commitment snapshots, with leaner state access and maintenance throughout. Most 3.5 users can upgrade without
re-syncing. To benefit immediately from the new plain commitment format, existing users should install 3.6, then run
erigon seg reset --datadir=<path> before starting the node; the next start will re-sync using the new snapshots. Datadirs
created with Erigon 3.3 or earlier that have not already been rebased must use this reset path or run
erigon seg step-rebase --datadir=<path> --new-step-size=390625 before their first normal 3.6 start.
- More reliable Caplin block production. Caplin starts payload building before proposer slots and publishes fork choice earlier, giving blocks more time to reach attesters (#23437, #23172) — by @lystopad
- Pruned nodes now reclaim old state snapshots. State history, indexes, optional commitment history, and the receipts
cache are retired once they fall outside their configured windows; fresh syncs also skip optional snapshots outside
those windows. This fixes unbounded snapshot growth on long-running
minimaland other pruned nodes (#21306, #22123, #21200, #22243, #22349) — by @AskAlexSharov, @JkLondon - Plain commitment snapshots. New commitment files store values directly by default, speeding state reads and merges
at the cost of larger files. Existing referenced files remain readable and convert lazily during merges; the offline
integration commitment convertcommand provides an explicit migration path (#14809, #21452, #21376, #21933) — by @awskii, @AskAlexSharov - Faster state access and snapshot maintenance. Sharded LRU state, code, and commitment caches replace whole-cache
resets; compression during state-snapshot merges is 2–3× faster, dictionary-building merges are about 1.5× faster,
and the mainnet
.btpivot cache uses roughly one quarter of its previous heap (#21386, #21982, #22154, #21625, #22050, #21875) — by @mh0lt, @yperbasis, @sudeepdino008, @AskAlexSharov
- Downgrading after 3.6 writes new snapshot files is unsupported. New 3.6 snapshot and accessor formats are incompatible with Erigon 3.5. Back up the datadir before upgrading if rollback is required (#21452, #21376, #21778).
- Historical receipts are off by default on fresh datadirs in every prune mode. Existing datadirs keep their stored
enable/disable setting. Without the cache, receipt and log RPCs re-execute within the state-history window at higher
latency. Blocks-mode operators who need receipts and logs back to genesis must use
--prune.include-receipts --prune.receipts.distance=keep-all(#22296, #22349) — by @yperbasis, @AskAlexSharov - Idle polling filters expire after five minutes. Clients using
eth_newFilter,eth_newBlockFilter, oreth_newPendingTransactionFiltermust poll within the timeout or recreate the filter. Configure--rpc.subscription.filters.timeout, or set it to0to disable eviction. Newsubscriptions_activeandsubscriptions_*_totalmetrics track the filter lifecycle (#22261) — by @onelapahead - JSON-RPC compatibility changes. Quoted decimal block numbers such as
"3"are rejected; use"0x3", a bare integer, or a named tag. Streaming trace errors no longer includeresult: nullalongsideerror(#21985, #21922) — by @lupin012
- With
--prune.include-commitment-history,eth_getWitnessnow builds and verifies a lean witness for any block within retained history (#22099) — by @awskii eth_fillTransactionfills missing fields and returns the unsigned raw transaction plus its JSON form, compatible with geth; KZG generation from raw blobs is not included (#21870) — by @lupin012eth_getBalance,eth_getCode,eth_getStorageAt,eth_getTransactionCount,eth_getProof, andeth_getStorageValuesnow default an omitted block parameter tolatest(#21586) — by @MysticRyuujintrace_blockandtrace_replayBlockTransactionscan include beacon-withdrawal balance changes through the optionalIncludeWithdrawalstrace setting (#21592) — by @lupin012
- Validation now rejects fork-inconsistent blocks and sidecars, truncated typed transactions, invalid fee relationships, and blob wrappers with trailing data (#22920, #23041, #22902, #23297) — by @yperbasis, @chfast
- Default block graffiti now identifies both the execution and consensus clients (#22303) — by @lystopad
- Caplin archive maintenance repairs truncated validator tables, backs off repeatedly failing retirement work, removes frozen state rows from its indexing DB for reuse, and stores compact effective balances, cutting the measured per-slot copy from about 266 MB to 18 MB (#22385, #22396, #22411, #23408) — by @awskii, @AskAlexSharov, @lystopad
- Caplin advertises its bound ports, skips peers without TCP endpoints, and accepts Chiado libp2p bootstrap nodes (#22273, #22271, #23245) — by @MysticRyuujin, @domiwei
- A deadlock in the transaction pool that could wedge block-producing nodes during high load has been resolved (#23360) — by @lystopad
- Glamsterdam devnet-6 support adds EIP-8282 builder execution requests, EIP-2780, EIP-8038, EIP-8246, and the devnet-6 revisions of EIP-7928/EIP-8037; Caplin tracks the Gloas alpha.11 request format, and BALs can be regenerated throughout retained state history, extending beyond the weak-subjectivity period (#22091, #22093, #22053, #22060, #22122, #22136, #22161, #21764) — by @domiwei, @taratorio. Devnet/testing only — not scheduled on mainnet or any public testnet.
- Pruning flags now use readable policies:
--prune.distanceacceptskeep-all, and--prune.distance.blocksacceptskeep-post-mergeandkeep-all. Receipt and commitment-history caches have independent--prune.*.distancewindows; the former receipt flag names remain aliases (#22119, #21200, #22349) — by @yperbasis, @JkLondon, @AskAlexSharov - Missing E3 accessors are rebuilt automatically on restart.
erigon seg index --rebuildrebuilds every snapshot accessor and index without deleting snapshot data (#22682, #21919) — by @AskAlexSharov, @sudeepdino008 - Commitment preloading no longer stalls when a step budget fills exactly (#23153) — by @lystopad
- The command-line layer now uses
urfave/cli/v3. Normal flag handling remains compatible, and--confignow applies to subcommands; operators with unusual invocation patterns should smoke-test them (#22130, #22546) — by @AskAlexSharov, @yperbasis, @lupin012
Full Changelog: https://github.com/erigontech/erigon/compare/v3.5.5...v3.6.0
v3.5.5 is a bugfix and security release recommended for all users, and especially for block proposers and archive-node operators. It fixes a panic on every Fulu blinded-block submission (#23150), a payload frozen before its slot began under an external consensus layer (#23102), and wrong logIndex values on archive nodes (#22951). It is a drop-in upgrade from 3.5.4 — no re-sync required.
Bugfixes
- cl/beacon: handle an empty Fulu builder response (#23150) by @domiwei — Builder API v2 answers a successful Fulu blinded-block submission with
202 Acceptedand no body, so the builder client returned all-nil andpublishBlindedBlocksdereferenced the nil payload. The beacon router installs nomiddleware.Recoverer, so the validator client saw only a dropped connection and nothing reached the Erigon log. Fixes #22598. - execution: build payloads until the slot they are for (#23102) by @lystopad — the builder's time budget ran from when payload attributes arrived rather than from the payload timestamp, so a consensus layer sending attributes well ahead of the slot got a payload frozen before that slot began. The budget now derives from the timestamp, floored at the old value and capped at two slots. Non-proposing nodes and Caplin are unaffected.
- db/state: route receipt-domain reads through the overlay
DomainReader(#22951) by @Sahil-4555 and @mh0lt — a block admitted through the block overlay read its receipt metadata from the committed tx, where a history miss fell back toGetLatestand returned the previous writing block's final log count. The wronglogIndexwas then cached in the RPC layer. Affected v3.5.1–v3.5.4. Fixes #22106. - db/state: roll back receipt domains on an in-RAM reorg unwind (#23064) by @MoonBoi9001 — the in-memory unwind path restored a hand-written list of the four state kinds that existed when it was written and never gained the receipt counters added later, so it left the abandoned blocks' values in place. It now walks every state kind, as the disk path does.
- db/downloader: allow seeding caplin state snapshots with a nil global type (#22980) by @lystopad — a node upgraded in place from a recent
release/3.4build loggednil ptr after parsing file: caplin/…-NextSyncCommittee.segand silently stopped seeding its caplin state snapshots.
Security
- build: bump
golang.org/x/textto v0.39.0 (#23178) by @lystopad — CVE-2026-56852: infinite loop inunicode/normon invalid UTF-8. Reachable from the downloader's HTTP/3 webseed path, and the only advisory here thatgovulncheckreports as called. - build: bump
golang.org/x/netto v0.56.0 (#23178) by @lystopad — CVE-2026-46600: panic parsing a malformed SVCB or HTTPS DNS resource record. - build: bump
github.com/quic-go/webtransport-goto v0.11.1 (#23178) by @lystopad — CVE-2026-57497: memory exhaustion from unbounded buffering of unknown capsules. Not reachable in Erigon: no WebTransport libp2p transport is registered. - build: bump
github.com/go-chi/chi/v5to v5.3.1 (#23178) by @lystopad — CVE-2025-69725 (open redirect inRedirectSlashes) plus threemiddleware.RealIPIP-spoofing advisories. Not reachable in Erigon: none of chi's middleware is used.
Improvements
- build: revert the
go-eth-kzgverifier optimization (#23177) by @yperbasis —release/3.5pinnedcrate-crypto/go-eth-kzgto an Erigon fork carrying an upstream change that is still an unmerged draft. This restores the released upstreamv1.5.0while the correctness concerns are investigated.
Full Changelog: https://github.com/erigontech/erigon/compare/v3.5.4...v3.5.5
v3.5.4 is a bugfix release recommended for all users, and especially for operators running the RPC daemon with response compression enabled — archive nodes and high-traffic RPC endpoints — where a native-memory leak in the gzip path could grow by ~9-15 GiB/day (#22700). It is a drop-in upgrade from 3.5.3 — no re-sync required.
Bugfixes
- node: fix a native (C-allocated) memory leak in the RPC gzip path (#22700) by @AskAlexSharov — the
libdeflate.Compressorwas pooled in async.Poolwhose GC-evicted entries never hadClose()called, leaking the C context (~9-15 GiB/day on an archive node). Replaced with a bounded channel pool that closes compressors on overflow. Fixes #22672. - cmd: expand a leading
~/(or~\on Windows) and$VARin--datadirfor the cobra-based binaries (#22785) by @lystopad — rpcdaemon and the other cobra commands previously used the raw path, so a tilde/env-prefixed--datadirwas not resolved the way the main erigon binary resolves it. Fixes #14629.
Improvements
- build: update
google.golang.org/grpcto v1.82.1 (#22690) by @AskAlexSharov — bringsrelease/3.5in line withrelease/3.6andmain.
Full Changelog: https://github.com/erigontech/erigon/compare/v3.5.3...v3.5.4
v3.5.3 is a bugfix release recommended for all users. It is a drop-in upgrade from 3.5.2 — no re-sync required.
Bugfixes
- execution/stagedsync: restore log/receipt notifications for mid-block resumed blocks (#22648) by @lupin012 — backports the notification-completeness half of #22235. After the earlier
ReceiptDomainfix, both executors still gatedRecentReceipts.Addon a staleisPartial/startTxIndex==0flag, so a block resumed mid-block never emitted a correct receipt/log notification overeth_subscribe("logs"/"newReceipts"). Both executors now reconstruct the resumed block's prefix receipts and gate on areceiptsCompletefield. - rpc: return a JSON-RPC error instead of panicking on a null transaction in the trace path (#22668) by @AskAlexSharov —
trace_transaction/trace_filtercould dereference a nil txn whenTxnByIdxInBlockresolves a txIndex whose body hasn't materialized yet at the chain tip. Fixes #22643. - rpc: fix the
handleBatchdeadlock (#22459) by @yperbasis — a filtered JSON-RPC batch could wedge onwg.Wait()and time the request out. Fixes #22424. - types: reject legacy transactions wrapped in a typed (EIP-2718) envelope (#22525) by @taratorio.
- types: reject an empty-string element in RLP transaction-list decoding (#22524) by @taratorio.
- cl/phase1/stages: fix
uint64underflow in the "Downloading Execution History" progress log (#22462) by @lystopad — once the live EL head advanced past the frozen initial progress, thetoprocesssubtraction wrapped to ~2⁶⁴ and produced a garbage ETA. Fixes #22455. - cl/phase1/stages: guard the "[Caplin] Forward Sync" progress log against slot under/overflow (#22465) by @lystopad — the same unguarded-subtraction /
time.Duration-overflow pattern on the forward-sync line.
Improvements
- rpc: make the
eth_getLogsper-position address/topic limit configurable via--rpc.logs.querylimit(#22477) by @lupin012 — replaces the hardcoded 1000-entry limit; default 1000 preserves current behaviour, 0 = unlimited. - cl/phase1/stages: route the history-download and forward-sync progress ETAs through the shared, overflow-safe
utils.ETAhelper (#22493, #22512) by @lystopad.
Full Changelog: https://github.com/erigontech/erigon/compare/v3.5.2...v3.5.3
v3.5.2 is a bugfix release recommended for all users, and especially for anyone running 3.5.1 — it fixes a sync-halting trie-root regression introduced there (#22399). It is a drop-in upgrade from 3.5.1 — no re-sync required.
Bugfixes
- db/state: clear the StateCache on
SharedDomainsunwind below the reorg window (#22402) by @Sahil-4555 — after the v3.5.1 changeset-isolation backport, a block that failed execution within the reorg window did a disk-noop overlay unwind that left dirty, uncommitted writes in the state cache; on the next run execution read those stale values instead of the database, producing a deterministic trie-root mismatch that halted sync. Closes #22399. - rpc, node: fix the nil-pointer panic in the gzip batch flush race (#22383) by @lupin012 — a gzipped JSON-RPC batch with two or more streamable methods invoked the shared gzip-streaming flush hook concurrently from per-call goroutines;
gzipResponseWriter.Flushis not concurrency-safe, so the calls raced on the underlying gzip writer and could dereference a nil flate compressor, crashing the node. Closes #22334.
Full Changelog: https://github.com/erigontech/erigon/compare/v3.5.1...v3.5.2
v3.5.1 is a bugfix release recommended for all users. It is a drop-in upgrade from 3.5.0 — no re-sync required.
Bugfixes
- execution/stagedsync: fix parallel-execution commitment consistency at step boundaries (#22111, #22135, #22147, #22094) by @awskii, @sudeepdino008 — a block straddling a step boundary left that step's commitment inconsistent with the account/storage/code domains, so published snapshots could mis-serve state-read RPCs (
eth_getProof,debug_executionWitness) and, on a later run, wedge the Execution stage in a zero-progress loop with nothing logged. Closes #21992, #22101. - execution/stagedsync: fix log index reset and missing WebSocket notifications in parallel execution (#22155) by @Sahil-4555 — restores
logs/newHeadssubscription notifications and correct log indexing when executing in parallel. - execution/stagedsync: fix pruning in
stage_custom_trace(#22052) by @sudeepdino008 — regenerated domains were never pruned and accumulated unbounded in the DB. Closes #22013. - cl/persistence: fix caplin historical state reconstruction loop at the Bellatrix transition (#22370) by @Sahil-4555 — a Caplin node could stop advancing at the Merge (Bellatrix) slot; pre-Merge blocks carry an all-zero execution payload header, and reconstruction now detects the zero block hash and skips the EL transaction lookup. Closes #22337.
- cl/phase1/forkchoice: replace latestMessagesStore interning with a flat per-validator slice (#22355) by @lystopad — the prune scan walked the whole message map on every update and stalled
GetHeadpast the attestation deadline. Closes #22351. - cl/antiquary: commit reconstructed state in bounded batches (#22348) by @awskii — committing up to 30 minutes of replayed beacon state in a single MDBX transaction could overflow libmdbx and crash on large (100+ GB) archive databases.
- cl: bound caplin archive blob-column backfill so it can't wedge on Fulu (#22318) by @awskii — prevents an archive node from getting stuck while backfilling blob (data-column) sidecars.
- cl, db/snapshotsync: caplin snapshot correctness — don't freeze empty block/state roots (#22323) and remove overlapping state snapshots on retire (#22317) by @awskii.
- cl: allow boundary attestations while head state lags (#22251) by @domiwei — fixes attestation validation at epoch boundaries when the wall-clock slot has advanced but the head state briefly lags behind.
- rpc/jsonrpc: gate
debug_executionWitnessonkeys[]completeness and keep preimages for in-block-deleted accounts (#22320) by @awskii — avoids returning an incomplete witness. - rpc: remove the state-history check from block-data-only endpoints (#22073) by @Sahil-4555 — endpoints that only need block data no longer error on nodes pruned below the requested block's state history.
- cmd/utils: allow snapshot reset on upgraded datadirs by restricting the table config (#22291) by @Sahil-4555 — snapshot reset failed to resolve the chain name on upgraded datadirs. Closes #22275.
Improvements
- execution: disable gzip compression for the Engine API (#22369) by @taratorio — removes compression overhead on
engine_*responses, loweringengine_getPayload/getBlobslatency. - p2p: re-resolve the STUN external IP at runtime (#22188) by @lystopad — with
--nat=stun, a node whose public IP changes while running now re-resolves and re-advertises it instead of keeping the stale startup value. - db/version: relax the minor-version check in
Supports(#22205) by @sudeepdino008 — a newer, backward-readable minor version within a supported major is now accepted, so the binary won't refuse newer minor-versioned files.
Full Changelog: https://github.com/erigontech/erigon/compare/v3.5.0...v3.5.1
Erigon 3.5.0 is a major release headlined by parallel block execution becoming the default and initial support for Ethereum's upcoming Glamsterdam hardfork. It is a drop-in upgrade for 3.4.x users — no re-sync required; existing datadirs upgrade their prune configuration automatically (see Breaking Changes).
- Parallel block execution, on by default. Erigon now executes EVM transactions across multiple cores by default, using the Block-STM (software transactional memory) design pioneered by Aptos: transactions run optimistically in parallel and are re-validated against a multi-version state, re-executing only on conflict (#21591 by @mh0lt, closes #17630). Revert to serial with
EXEC3_PARALLEL=falseor--exec.serial. - Glamsterdam devnet support. Initial implementation of Ethereum's next hardfork: Block-Level Access Lists (EIP-7928), enshrined Proposer-Builder Separation / "Gloas" (EIP-7732) in Caplin, gas repricings (EIP-8037, EIP-7976, EIP-7981), larger contracts (EIP-7954), transfer logs (EIP-7708), and the
eth/71Block Access List wire protocol (EIP-8159). Devnet/testing only — not scheduled on mainnet or any public testnet. debug_executionWitness. Stateless execution-witness generation (EIP-7928/8025) with reth-compatible output, for zkEVM and stateless clients (#20205 by @antonis19, #21629 by @awskii).- More aggressive history pruning by default.
--prune.mode=fullnow follows the EIP-8252 reorg-retention window (~36 days / 262,144 blocks) — see Breaking Changes. - GraphQL API revival. Broad resolver coverage restored — queries, logs,
call,sendRawTransaction,estimateGas,gasPrice, storage, and EIP-4844 fields.
Full mode now retains state and block data for the last 262,144 blocks (~36.4 days), matching EIP-8252's REORG_RETENTION_WINDOW (#21342). Previously full mode pruned only pre-merge block data (EIP-4444 history-expiry) and kept the last 100,000 blocks of state history.
What changed:
| Before | After | |
|---|---|---|
| State history retention | last 100,000 blocks | last 262,144 blocks |
| Block data retention | pre-merge pruned, all post-merge kept (EIP-4444) | last 262,144 blocks |
Migration: existing datadirs upgrade automatically and silently. To keep the old "retain all post-merge block data" behavior, set --prune.distance.blocks=18446744073709551615.
Note: physical deletion of frozen snapshot files is not implemented yet (see #21306), so existing on-disk historical blocks persist for now, though the new cutoff is already recorded at the config level.
In practice, this means only freshly synced full nodes will have a reduced disk footprint.
--prune.mode=blocks keeps the same shape as before (all block data retained), but its state history retention also bumps from 100,000 to 262,144 blocks. --prune.mode=minimal is unchanged — both block and state history retain the 100,000-block window, deliberately sub-EIP-8252 for disk-constrained operators. See #21342 for details.
Erigon now opens a single TCP listener on --port (default 30303) carrying every configured eth protocol version, instead of one listener per protocol on 30303/30304/30305. This fixes a discovery-DHT race that left inbound peers stuck at a fraction of --maxpeers for multi-protocol deployments: per-protocol ENRs collided under one Node ID, so only one survived in the DHT and peers dialed the wrong listener (#21335).
What changed:
| Aspect | Before | After |
|---|---|---|
| Inbound peer ports | 30303, 30304, 30305, … (one per eth version) |
30303 only |
--p2p.allowed-ports flag |
Picked one port per protocol from this list | Removed — passing it now errors |
--maxpeers semantics |
Per-protocol cap; actual ceiling ≈ N × maxpeers | Honest total cap |
Default --maxpeers |
32 |
64 (compensates for the now-honest cap) |
| Enode database directory | <datadir>/nodes/eth68, <datadir>/nodes/eth69, … |
<datadir>/nodes/eth |
Migration:
- Remove
--p2p.allowed-ports=...from CLI args / config files; it is no longer recognised. - Firewall, Kubernetes Service, and monitoring rules that explicitly opened 30304/30305 can drop those entries — only
--portis bound now. - If you previously lowered
--maxpeersbecause you knew the per-protocol multiplication inflated the real ceiling, raise it back to the target total (the cap is now what the flag says). - First run after upgrade loses the warm peer cache in
nodes/eth{68,69,…}— nothing on disk is deleted, the directories are simply no longer read; discovery rebuilds the peer set from bootnodes within a few minutes.
Standalone sentry binary (cmd/sentry) and --sentry.api.addr (remote sentry over gRPC) are unaffected — neither had the bug.
Aligns Erigon with the execution-apis specification (ethereum/execution-apis#762) and Geth behavior.
What changed:
| Field | Before (Erigon) | After (Erigon / Geth / Spec) |
|---|---|---|
| Memory in trace | disableMemory (default: included) |
enableMemory (default: excluded) |
| Return data in trace | disableReturnData (default: included) |
enableReturnData (default: excluded) |
Both the key and its default changed: disable* → enable*, and memory and return data are now excluded unless explicitly enabled — matching the spec and Geth.
Migration: memory and return data are now excluded by default. To include them, add the new opt-in key (omit it to keep the default):
- Memory:
{ "enableMemory": true } - Return data:
{ "enableReturnData": true }
Affected RPC methods: debug_traceTransaction, debug_traceBlockByHash, debug_traceBlockByNumber, debug_traceCall.
The legacy Clique proof-of-authority engine has been removed (#20532 by @yperbasis). --chain=dev now runs on an embedded proof-of-stake consensus instead of Clique (#20451 by @mh0lt), matching how all live networks operate post-Merge. Networks or tooling that still depended on Clique are no longer supported.
The optional Silkworm C++ execution-backend integration and its --silkworm.* flags have been removed (#19662 by @canepat). Erigon uses its native Go execution engine exclusively.
3.5.0 adds an initial implementation of Ethereum's next hardfork — Glamsterdam (consensus-layer "Gloas" + execution-layer "Amsterdam") — for devnet testing and validation. It is not scheduled on mainnet or any public testnet, and these code paths are inert on production networks until an activation time is configured.
- EIP-7928 — Block-Level Access Lists (BAL): records every account and storage slot a block touches, enabling deterministic parallel validation. Full builder, validator, and strict-validation support (#19627, #19656, #20602, #20776), plus the
eth_getBlockAccessListRPC method (#19929) — by @mh0lt, @yperbasis, @Sahil-4555 - EIP-7732 — Enshrined Proposer-Builder Separation (ePBS / "Gloas"): implemented in Caplin — execution-payload envelope, PTC, and builder payments (#18956) — with follow-up audit and fork-choice fixes (#21248, #21228) — by @domiwei
- Gas repricings: EIP-8037 State Creation Gas Cost Increase (#19596), EIP-7976 calldata floor cost (#20613), EIP-7981 access-list cost (#20671) — by @taratorio
- EIP-7954 — Increase Maximum Contract Size (#19624) — by @yperbasis
- EIP-7843 — slot-number opcode (
SLOTNUM), wired into Caplin block production andengine_forkchoiceUpdatedV4(#20175) — by @yperbasis - Networking:
eth/71Block Access List exchange (EIP-8159, #20793, #20794, #20795) — by @mh0lt
debug_executionWitness: generate stateless execution witnesses (EIP-7928/8025), withlegacyandcanonicaloutput modes — thelegacyformat is reth-compatible — for zkEVM and stateless clients (#20205, #21371, #21518, #21629) — by @antonis19, @lupin012, @awskiieth_capabilities: report the set of supported RPC methods (#20951) — by @lupin012debug_setHead: rewind the chain head (#19577) — by @canepat- GraphQL substantially revived — transaction, logs,
call,sendRawTransaction,estimateGas,gasPrice, and storage resolvers, plus EIP-4844 fields (#20389, #20916, #21219, #21379, #21060) — by @lupin012 testing_namespace exposed via--http.apifor engine/spec test harnesses (#20482) — by @lupin012eth_simulateV1: per-call gas and result limits (#20232) — by @Sahil-4555
--exec.no-prune(disable all DB pruning),--exec.serial(force single-threaded execution), and--exec.*executor-tuning flags (#20915, #20853, #20797) — by @mh0ltseg du(snapshot disk-usage analysis, #20104) andseg rm-blocks(remove latest block snapshots, #20554) — by @awskii, @sudeepdino008
- WebSocket transport rewritten on
coder/websocket, with overload protection, clean close frames, and bounded write timeouts (#20097, #20446, #20788, #20923) — by @lystopad, @lupin012, @Sahil-4555 - Admission control: uniform
503responses under load (#20303); optional response compression via libdeflate (#20665) — by @lupin012 - Geth compatibility:
debug_traceTransactionindex format (#20210),trace_rawTransaction(#20448),debug_accountRange(#20057), nullv,r,sfor unsigned transactions (#21321) — by @lupin012 - Performance: faster
eth_getLogs(#20561),trace_block(#20182),eth_gasPrice(#19678), canonical-hash cache (#19173);engine_getPayload~2.4× andgetBlobs~10× faster (#21615, #21606) — by @lupin012, @taratorio trace_*returns an explicit error when an unsupported custom tracer is supplied (#21544) — by @lupin012
- New
eth/70wire protocol: partial block receipt lists (EIP-7975, #19755) — by @yperbasis - All eth protocol versions now multiplex on a single TCP listener (see Breaking Changes, #21335) — by @lystopad
- Peer hygiene / DoS hardening: cap and rate-limit inbound
NewBlockHashes(#21557), enforce the 4096-hash limit onNewPooledTransactionHashes(#20577), drop peers failing blob KZG verification (#21421), and bound fan-out stream buffers (#20783) — by @yperbasis - Skip chain-specific bootnodes on genesis-hash mismatch (#19807); honour an explicitly empty
--bootnodes(#20630) — by @yperbasis
- Proactive dormancy-based eviction of stale queued transactions (#19862) — by @lystopad
- Transaction parsing migrated onto the shared
execution/typestransaction types (#19757); malformed EIP-7702 authorization tuples are now tolerated rather than rejected wholesale (#20809) — by @yperbasis
- Unified Engine API client for standalone mode (#20035) — by @mh0lt
- Fork-choice and ENR-stability fixes — recovery from a post-Gloas fork-choice stall and a persistent node key for stable ENR across restarts (#21228, #21276) — by @domiwei
- Block production: give the EL builder a build window before stopping it, fixing near-empty proposed blocks (~0–2% gas) on otherwise-healthy validators (#21990) — by @lystopad
- Off-heap Elias-Fano index building (#20640) and parallel commitment computation (#20805) — by @AskAlexSharov, @mh0lt
- Transient-storage zero-write fast path (#20568) and opcode-scoped intern cache to eliminate duplicate
unique.Make()(#20552) — by @Sahil-4555, @AskAlexSharov
- Clique PoA engine (#20532 by @yperbasis) and Silkworm integration (#19662 by @canepat) — see Breaking Changes.
- Unused
hack(#20412),state(#20420), anddiag(#21351) helper binaries — by @awskii, @AskAlexSharov
--ethstatscredentials are redacted from the startup command log (#20890) — by @MysticRyuujin- DoS-resistance limits on inbound P2P message volume (#20577, #21557) and bounded RPC/stream buffers (#20446, #20783) — by @yperbasis, @lupin012
Full Changelog: https://github.com/erigontech/erigon/compare/v3.4.4...v3.5.0
v3.4.4 is a bugfix release recommended for all users.
Bugfixes
- execution/stagedsync: prune in-RAM overlay when execution unwind is a no-op (#21824, #21847) by @JkLondon — third fix for the post-reorg
gas used mismatch. - caplin: serialize uint64 beacon API fields as JSON strings (#21805) by @BitWonka - Per the beacon-APIs spec, Uint64/Gwei fields must be serialized as JSON strings. Several Caplin response types were emitting them as JSON numbers, breaking spec-compliant clients. Fixes #20562.
Full Changelog: https://github.com/erigontech/erigon/compare/v3.4.3...v3.4.4
v3.4.3 is a bugfix release recommended for all users.
Bugfixes
- db/state: prune
TemporalMemBatchoverlay entries past the unwind point (#21538) by @JkLondon — second fix for the post-reorggas used mismatch/ state-leak some users still hit on v3.4.2. After a tip reorg a stale read in the in-memory overlay could return a write made inside the unwoundtxNumrange, flipping anSSTOREfrom cold to warm gas pricing. Complements the #21157 diffset fix shipped in v3.4.2. - rpc: match Geth semantics in
debug_getModifiedAccountsByHash/debug_getModifiedAccountsByNumber(#21507) by @lupin012 — corrects the block-range convention (exclusive start), now also reports contracts whose storage changed without an account change, and excludes touched-but-unchanged precompiles and self-destructed accounts. - node/cli: register
--rpc.logs.maxresultsinDefaultFlagsso it takes effect via the CLI (#21389) by @lupin012 — the limit was documented in 3.4.0 but never wired into the flag set, so setting it on the command line had no effect; it now applies.
Improvements
- execution/p2p, execution/engineapi: fail-fast
engine_newPayloadbackward download when the gap exceeds the reorg limit (#21502) by @yperbasis — when a payload's parent is more thanMaxReorgDepthblocks from the local head, the download short-circuits instead of fetching a header batch every slot, and logs the expected gap at INFO instead of WARN. The gap is still closed by the following fork-choice update.
Full Changelog: https://github.com/erigontech/erigon/compare/v3.4.2...v3.4.3
v3.4.2 is a bugfix release recommended for all users.
Bugfixes
- execution/stagedsync: find diffset by actually-executed hash on unwind (#21157) by @JkLondon — fixes a
state-leak bug in
unwindExec3that surfaces asgas used mismatch/Cannot update chain headafter a tip reorg whose unwound block deployed a contract viaCREATE/CREATE2to a counterfactual address (safe-wallets, EIP-1167 clone factories, ERC-4337 accounts, deterministic deployers). The unwind now walks every header at the height to find the diffset of the block actually executed, instead of assuming the (already-flipped) canonical hash matches. - rawdb: ignore invalid receipt cache transaction indexes (#21262) by @Sahil-4555
Improvements
- db/state, ethconfig: bound domain merge; add
--erigondb.domain.steps-in-frozen-file(#21148) by @wmitsuda
Full Changelog: https://github.com/erigontech/erigon/compare/v3.4.1...v3.4.2
Bugfixes
- commitment: segfault fix - caused by branch slice returned by TrieContext.Branch (#21044) by @awskii
Full Changelog: https://github.com/erigontech/erigon/compare/v3.4.0...v3.4.1
Erigon 3.4.0 is a major update for node operators and validators, focused on stability, performance, and efficiency at ChainTip. It is a drop-in upgrade for 3.3.x users — no data migration or re-sync required.
- Fast restart on ChainTip: no more blocking indexing or pruning at startup; Caplin doesn't lose its download progress.
- 4x smaller Chaindata (~20 GB): improves ChainTip speed. Available via re-sync, or by running
./build/bin/erigon seg step-rebase --datadir=<your_path> --new-step-size=390625(takes ~10 seconds). - Historical
eth_getProof: is no longer experimental. Recommended: 32 GB+ RAM. Re-sync with--prune.include-commitment-historyto apply the latest data fixes. - New RPC endpoints:
trace_rawTransaction,eth_getStorageValues,admin_addTrustedPeer,admin_removeTrustedPeer, flat tracers,engine_getBlobsV3. - Reduced impact on ChainTip performance: from RPC, background file merging, pruning, and optional
heavy flags (
--persist.receipts,--prune.include-commitment-history).
- Minimum Go version: 1.25
--rpc.blockrange.limit=1_000new limit. Maximum block range (end - begin) allowed for range queries over RPC. 0 - means unlimited. Default: 1_000--rpc.logs.maxresults=20_000new limit. Maximum number of logs returned by eth_getLogs, erigon_getLogs, erigon_getLatestLogs. 0 - means unlimited. Default: 20_000--rpc.max.concurrency=0new limit. Maximum number of concurrent HTTP RPC requests (HTTP admission control). 0 = use db.read.concurrency, -1 = unlimited (no admission control). Default: 0- p2p: switched to
discv5.discv4disabled by default.
trace_rawTransaction: execute and trace a raw signed transaction without broadcasting it (#19524)eth_getStorageValues: batch fetch of multiple storage slots in a single call (#19442)admin_addTrustedPeer/admin_removeTrustedPeer: manage trusted peers at runtime (#19413)- Call flat tracers (
trace_callfamily): flat trace output format support (#18556) engine_getBlobsV3: Engine API blob retrieval v3 (#18512)trace_call:StateOverridesprecompile support (#18401, #18492)
- Fusaka scheduled for Chiado (Gnosis Chain testnet) at slot 21 651 456, epoch 1 353 216, timestamp 1 773 653 580 (Mon 16 Mar 2026 09:33:00 UTC) (#19682)
- Chiado bootstrap nodes updated to match the Lighthouse built-in Chiado network config (#19693)
- Balancer hard fork for Gnosis Chain mainnet (#18122)
- Amsterdam signer support and BAL non-determinism fix (#19434)
- BAL selfdestruct net-zero fix (#19528)
- Parallel execution fixes for block-access-list (BAL) workloads (#17319)
- execution/vm: EIP-8024 (
SWAPN,DUPN,EXCHANGE) opcodes implemented (#18670) - Pectra requests-hash validation: fix partial block receipt reconstruction when execution resumes from
a snapshot boundary mid-block — resolves
invalid requests root hash in headeron mainnet re-sync at block 24966723 (#20452)
- Persistent historical download — Caplin now persists and resumes historical beacon block downloads across restarts (#18320)
- Discovery v5 enabled by default —
discv5is now the default peer discovery protocol (#18578) - cl/p2p, cl/sentinel: fix DISCV5 ENR missing IP when the discovery address is unspecified (#19585)
- Fix missing attestations by using GossipSub for subnet peer coverage (#19523)
- cl/gossip: fix conditions forwarding, ENR lifecycle, and epoch-mismatch — prevents false peer banning, reduces log flooding from redundant ENR updates, and guards against stale RANDAO committee computation (#20777)
- cl/beacon: add
single_attestationevent topic support (#18142) - cl/beacon: set
Eth-Consensus-Versionheader when versioned (#18377)
eth_getBlockReceipts: limit over-concurrency — prevents latency growth and out-of-memory (OOM) at high request rates (#19725)debug_traceCallMany: fix globalBlockOverridesandStateOverridesnot being applied (#19547)debug_traceCall: fix state and block overrides interaction (#18480)eth_blobBaseFee: fix incorrect value returned (#18506)eth_getBalanceand others: fix block-not-found error for certain historical queries (#18457)- Block-hash canonicality check added to APIs that accept a block hash and state (#19356)
- rpc: fix batch limit exceeded error to comply with the JSON-RPC spec (#18260)
trace_replayTransaction(): add stack info forTLOADopcode (#19550)eth_feeHistory: performance optimisation and pending block support (#19526, #19455)debug_trace*: zero-alloc memory word encoding inJsonStreamLogger— eliminates per-word heap allocations and prevents OOM on large block traces (#20754)- prestateTracer: fix diff mode missing deleted accounts to match geth behaviour (#20775)
- rpc/mcp: fix
tools/callhanging indefinitely under DB load by switching the SSE context to non-blocking read-tx acquire (#20778)
- Zombie queued transactions: transactions exceeding
MaxNonceGapare now evicted (#19449) - txnprovider/shutter: fix premature encrypted txn pool cleanup and peer drops (#18351)
- txpool: cache
pendingBaseFeefor queue comparisons to reduce recomputation (#18341)
- p2p/sentry: fix wrong OR in case statement for
witprotocol messages (#19580) - p2p: better handling of RLP errors in
wit(#19569) - discv4 disabled by default on mainnet (discv5 preferred) (#18640)
- merge: prioritize Domain merge over History — 2x less disk space required, and less impact to ChainTip from history-merge (#19441)
- merge: fix O(n²) InvertedIndex re-merge — eliminates quadratic time complexity during re-merge (#19680)
SequenceBuilder: avoids an intermediate Elias-Fano representation during sequence building and merge (#19552, #19567)- Faster startup: state-file index building deferred to reduce restart latency (#19583, #19407)
- Graceful restart in history download:
SpawnStageHistoryDownloadnow honours the stage context, so Ctrl+C during history download exits promptly instead of waiting for the download to finish (#20766) - DB integrity checks: time budget added so checks no longer run unbounded (#20714)
- Commitment-history integrity: skip the integrity check when commitment history is disabled (#20835)
seg ls: report dictionary size and memory usage in the output (#20790)- anacrolix/torrent: fix peerconn panic on bad reads while serving peer requests (#20748)
- github.com/buger/jsonparser: bump 1.1.1 → 1.1.2 to address CVE-2026-32285 (HIGH, CVSS 7.5 — out-of-bounds read) (#20018)
Full Changelog: https://github.com/erigontech/erigon/compare/v3.3.10...v3.4.0
This release schedules Fusaka on Gnosis Chain mainnet at Tue 14 April 2026, 12:06:20 UTC and thus is mandatory for all Gnosis users. It is also recommended for all users in general.
Changes
- Schedule Gnosis Fusaka (#20090) by @domiwei
- txpool: fully discard pre-Osaka blob txns in best() (#20120) by @yperbasis
- rpc: auto-convert legacy blob sidecar to v1 cell proofs after Osaka (#20087) by @domiwei
- rpc: add blockRangeLimit param for API works on blocks range (#19614) by @lupin012
- cl: fall back to local head state when remote checkpoint sync fails (#20003) by @domiwei
- cl: fix fork choice block validation and error propagation (#19927) by @domiwei
- cl: re-enable voluntary_exit gossip topic subscription (#19764) by @Giulio2002
Full Changelog: https://github.com/erigontech/erigon/compare/v3.3.9...v3.3.10
This release schedules Fusaka on Chiado on Mon 16 March 2026, 09:33:00 UTC and thus is mandatory for all Chiado users.
Changes
- fix(txpool): evict zombie queued txns exceeding MaxNonceGap (cherry-pick #19449 → release/3.3) (#19591) by @Giulio2002
- cl/sentinel: fix DISCV5 ENR missing IP when discovery address is unspecified (#19647) by @lystopad
- rpc: accept only canonical hash for block receipts (#19649) by @canepat
- Schedule Fusaka for Chiado (#19681) by @yperbasis
- cl: amend Chiado bootstrap nodes (#19692) by @yperbasis
Full Changelog: https://github.com/erigontech/erigon/compare/v3.3.8...v3.3.9
Changes
- New Chiado boot nodes (cherry-pick #18867) (#19241) by @sudeepdino008
- eth_getLogs: receipts availability check to be aware about
--persist.receiptsand--prune.mode=minimal(#19226) by @AskAlexSharov - txnprovider/shutter: fix decryption keys processing when keys do not follow txnIndex order (#18951) (#18959) by @taratorio
- execution: fix Chiado re-exec from genesis (#18887) by @yperbasis
Bugfixes
- execution/tests: minor fix chainmaker add withdrawals in shanghai (#18886) by @taratorio
- Reduce impact of background merge/compress to ChainTip (#18995) by @AskAlexSharov
- fix(caplin): Fixes for DataColumnSidecar (#18268) (#19003) by @taratorio
- rpc: bound checks in receipts cache V2 and generator (#19046) by @canepat
- execution/execmodule: fix unwinding logic when side forks go back in height (#18993) (#19063) by @taratorio
- p2p: fix nil pointer crash with --nodiscover (#19056) by @sudeepdino008
- rpc: add check on latest executed block (#19133) by @canepat
- protect History from events duplication (#19230) by @AskAlexSharov
Full Changelog: https://github.com/erigontech/erigon/compare/v3.3.7...v3.3.8
P2P stability. Prune performance.
Changes
- prune: remove early-exit based on DirtySpace() (#18787) by @AskAlexSharov
- execution: fix commitment state key txNum when last block tx is at step boundary (#18858) by @taratorio
- p2p deadlock fix (#18862) by @Copilot
- Forward compatibility: Fixed index building for v0 snapshot format (#18824) by @eastorski
- Backward compatibility: Add --v5disc alias (#18785) by @anacrolix
- Show the default P2P discovery bools in --help (#18819) by @anacrolix
Full Changelog: https://github.com/erigontech/erigon/compare/v3.3.4...v3.3.7
v3.3.4 is a bugfix release recommended for all users.
Changes
- Fix for missing rcache at stepBoundary (#18698) by @sudeepdino008
- execution/stagedsync: port --experimental.always-generate-changesets flag for gas repricing benchmarks (#18733) by @taratorio
- execution: add --fcu.timeout and --fcu.background.prune flags (#18723) by @taratorio
- Not needed collector erased in prune domains (#18665) by @JkLondon
- Forward compatible snapshots format (#18746) by @eastorski
Full Changelog: https://github.com/erigontech/erigon/compare/v3.3.3...v3.3.4
v3.3.3 is a bugfix release recommended for all users.
Fixes
- CL: Fix incorrect committee count in process_attestation (#18316) by @domiwei
- CL: Fix for periodic retry of not-ready response (#18376) by @taratorio
- Shutter: fix premature encrypted txn pool cleanup bug and peer drops (#18264) by @taratorio
- Half block exec fix in receipts (#18426, #18505) by @sudeepdino008
- p2p: resolve security vulnerability (#18650) by @yperbasis
Improvements
- p2p: Turn on discovery v5 by default (#18639) by @anacrolix
Full Changelog: https://github.com/erigontech/erigon/compare/v3.3.2...v3.3.3
Gnosis hardfork support.
Changes
- rpc: fix panic on
GetEthV1BeaconBlobs(#17992) (#18232) by @domiwei - rpc: fix invalid key error code in
eth_getStorageAt(#18238) by @canepat - cl/beacon: add single_attestation event topic support (#18142) (#18277) by @domiwei
- Gnosis Balancer (#18282) by @mh0lt
- bittorrent: torrent excessive wantPeers event fix (#18256) by @anacrolix
Full Changelog: https://github.com/erigontech/erigon/compare/v3.3.1...v3.3.2
- We have new Docs and HelpCenter: https://docs.erigon.tech/
- Support of historical
eth_getProof(erigontech#12984). It requires--prune.include-commitment-historyflag.
RPC
- rpc: fix txpool_content crash: unknown (#18120) by @canepat
CL
- Caplin: Simplified peer refreshing (#18123) (#18152) by @domiwei
TxPool
- txpool: remove double PopWorst() in pending pool overflow (#18159) by @AskAlexSharov
Sync
- Torrent client fixes (#18179) by @anacrolix
Full Changelog: https://github.com/erigontech/erigon/compare/v3.3.0...v3.3.1
- Support of historical
eth_getProof(erigontech#12984). It requires--prune.experimental.include-commitment-historyflag. - Look our new Docs and HelpCenter: https://docs.erigon.tech/
eth_simulateV1: Complete implementation of Ethereum simulation API with support for state overrides, blob transactions, block overrides, and historical state roots (#15771)eth_createAccessList: StateOverrides parameter support (#17653)- Support for
eth_callwith blockOverrides (#17261) trace_filter: Block tags support (#17238)debug_traceTransaction: Self-destruct operation validation (EIP 6780) (#17728)
- EIP-7928: BlockAccessList type support (#17544)
- EIP-7934: EstimateGas capped by MaxTxnGasLimit in Osaka (#17251)
- EIP-7702 transaction support in
(r *Receipt) decodeTyped(#17412) - Rewrite bytecode support for post-Merge blocks (#17770)
- Get blobs support (Fusaka compatibility) (#17829)
eth_getTransactionReceipt: Pre-Byzantium transaction handling (#17479, #17509)eth_estimateGas: Improved handling with StateOverrides (#17914, #17295)debug_traceCall: System contract execution support (#17339)- Blob transaction and blob base fee override support (#17313)
- Experimental Parallel Exec (#16922)
- MAX_PENDING_PARTIALS_PER_WITHDRAWALS_SWEEP for Gnosis (#17501)
- Reduce goroutines amount produced by BitTorrent library (#17765)
- Up base image to
Go 1.25-trixie(#17837)
- PoW mining was removed in #17813. The
--chain=devmode now uses an embedded PoS consensus layer (Caplin) with deterministic validators instead of Clique. Seedocs/DEV_CHAIN.mdfor usage. - Holesky network support removed (#17685)
- eth/67 protocol support removed (#17318)
- SkipAnalysis VM optimization removed (#17217)
Full Changelog: https://github.com/erigontech/erigon/compare/v3.2.2...v3.3.0
Changes
- Caplin: add getBlobs support (fusaka) (#17840) by @Giulio2002
Full Changelog: https://github.com/erigontech/erigon/compare/v3.2.2...v3.2.3
v3.2.2 schedules Fusaka on Ethereum mainnet on December 3, 2025 at 09:49:11pm UTC. Thus it is a mandatory update for all Ethereum mainnet users.
New features
- Schedule Fusaka on Ethereum mainnet in #17736 by @yperbasis
- Tool to fetch and recover blobs from a remote beacon API in #17611 by @Giulio2002
Full Changelog: https://github.com/erigontech/erigon/compare/v3.2.1...v3.2.2
v3.2.1 is a bugfix release recommended for all users, especially validators.
Fixes
- Fix validators producing bad blocks on Hoodi in #17487 by @mh0lt
- RPC: fix "insufficient funds for gas * price + value" error in traces retrieval for a specific block (Issues #16909, #17232) in #17523 by @antonis19
- RPC: fix no changes and filter not found in eth_getFilter* (Issue #17246) in #17350 by @canepat
- RPC: debug_traceCall fix avoid to trace sysContract (Issue #17220) in #17360 by @lupin012
- CL: fix initial previous_version in fork_schedule (Issue #17262) in #17331 by @domiwei
Improvements
- Ethereum mainnet default block gas limit is raised to 60M in #17321 by @yperbasis
- CL: Allow blob requests after Fusaka in #17500 by @domiwei
Full Changelog: https://github.com/erigontech/erigon/compare/v3.2.0...v3.2.1
Erigon 3.2.0 has a complete implementation of Fusaka and schedules it on the test nets (#17197):
- Holesky on Wednesday, 1 October 2025 08:48:00 UTC
- Sepolia on Tuesday, 14 October 2025 07:36:00 UTC
- Hoodi on Tuesday, 28 October 2025 18:53:12 UTC
Fixes
- Re-org/unwind fixes (#17105, #17165) by @taratorio
- RPC: Fixes to eth_getProof (#16220, #16251, #16564, #16606, #16687) by @awskii
- tracer: fix prestates for EIP7702 transactions (#16497) by @nebojsa94
Improvements
- New EL block downloader (#16270, #16673) by @taratorio
- Caplin p2p improvements (#16719, #16995) by @Giulio2002
- EVM: MODEXP precompile performance improvements (#16579, #16583, #16396, #17151) by @chfast & @yperbasis
- execution: more accurate bad block responses (#16994) by @taratorio
- Block builder: improve txpool polling (#16412) by @taratorio
- execution/stagedsync: handle sync loop block limit exhaustion (#16268) by @taratorio
- RPC: Apply batch limit to WebSocket/IPC connections (#16255) by @grandchildrice
- RPC: Estimate gas align to geth (#16397) by @lupin012
- snapshotsync: add support for
--snap.download.to.block(#16938) by @taratorio
New features
- Complete Fusaka implementation (#16183, #16185, #16186, #16187, #16184, #16391, #16401, #16207, #16420, #16428, #16494, #16457, #16644, #16928, #16945, #17060, #16989, #17076, #17169) by @taratorio, @yperbasis, @Giulio2002 and @domiwei
- Implement eth/69 (#15279, #17186, #17171) by @shohamc1
- RPC: implement new eth_config spec (#16218, #16410) by @taratorio
- RPC: impl admin_RemovePeer (#16292) by @lupin012
Full Changelog: https://github.com/erigontech/erigon/compare/v3.1.0...v3.2.0