TraceDB: Snapshot-backed state for the trace baker

[Kbhat1]

Describe your changes and provide context

• Adds optional snapshot-backed trace baking so the baker can replay from in-memory memiavl state instead of SS-pebble.
• Refcounts memiavl snapshots and releases trace leases through geth's existing StateReleaseFunc, avoiding GC finalizers.
• Opt-in via trace_bake_use_snapshot; falls back when the backend cannot provide a snapshot.

Testing performed to validate your change

• Verified on node
• Unit tests
• Long-running node on mainnet
• go test ./sei-db/state_db/sc/memiavl -run 'TreeCopy|Snapshot' -count=1

---

Note

Medium Risk
Introduces optional snapshot-backed state for debug_trace* baking and changes memiavl snapshot lifecycle/refcounting, which can affect memory usage and correctness of trace/state replay if mishandled (though it is opt-in with fallbacks).

Overview
Adds an opt-in path (evm.trace_bake_use_snapshot, evm.trace_bake_snapshot_window) for the trace baker to replay blocks against in-memory memiavl snapshots instead of SS-pebble, including wiring snapshot capture at EndBlock and closing retained snapshots on app shutdown.

Plumbs a new TraceContextProvider through the EVM RPC servers and tracing backend so debug_trace* can build contexts from leased snapshots and return a release function via geth’s StateReleaseFunc.

Extends storev2/rootmulti and SeiDB committers to support O(1) snapshot Copy() and safe release of snapshot refs, and adds refcounting to memiavl snapshot mmap lifecycles with regression/unit tests covering rewrite/reload and snapshot-store eviction behavior.

^{Reviewed by Cursor Bugbot for commit 0e3215d. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

The latest Buf updates on your PR. Results from workflow Buf / buf (pull_request).

Build	Format	Lint	Breaking	Updated (UTC)
✅ passed	✅ passed	✅ passed	✅ passed	May 11, 2026, 2:37 PM

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 59.22%. Comparing base (654d40b) to head (0e3215d).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3360      +/-   ##
==========================================
- Coverage   59.25%   59.22%   -0.03%     
==========================================
  Files        2110     2110              
  Lines      174181   174058     -123     
==========================================
- Hits       103210   103094     -116     
+ Misses      62044    62032      -12     
- Partials     8927     8932       +5     

Flag	Coverage Δ
sei-db	70.41% <ø> (ø)

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

Files with missing lines	Coverage Δ
app/app.go	69.77% <ø> (+0.08%)	⬆️
evmrpc/config/config.go	74.31% <ø> (ø)
evmrpc/server.go	88.46% <ø> (ø)
evmrpc/simulate.go	74.58% <ø> (+0.21%)	⬆️
evmrpc/tracers.go	67.95% <ø> (ø)
sei-cosmos/storev2/rootmulti/store.go	65.30% <ø> (ø)
sei-db/state_db/sc/composite/store.go	68.78% <ø> (-5.34%)	⬇️
sei-db/state_db/sc/memiavl/db.go	66.66% <ø> (+0.44%)	⬆️
sei-db/state_db/sc/memiavl/snapshot.go	62.42% <ø> (-0.18%)	⬇️
sei-db/state_db/sc/memiavl/store.go	90.69% <ø> (-2.80%)	⬇️
... and 3 more

... and 28 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit aacf715. Configure here.}

[blindchaser]

only the first element is ever read. A non-variadic *TraceContextProvider parameter (or a small options struct) avoids the "what if someone passes two" ambiguity and keeps the signature self-documenting.

[blindchaser]

TraceSnapshotStore.Close() returns nothing, but inside it ignores per-snapshot release errors:

x/evm/keeper/trace_snapshot.go:97-112

_ = releaser.ReleaseSnapshotRefs()

Same as the WARN above: refcount mismatches are real bugs. Either return error or log at WARN level on close to keep ops visibility.

[blindchaser]

EndBlock snapshot semantics are subtle — comment is dense, easy to mis-read

The off-by-one here is correct but non-obvious: storev2/rootmulti.flush() doesn't run until Commit(), so at EndBlock(N) the SC tree state is state_after_commit_of_(N-1) and snap.Version() == N-1. The baker then traces H=N-1, whose initializeBlock calls ctxProvider(H-1) = ctxProvider(N-2), which finds snap[N-2] Put at the previous EndBlock(N-1). Worth a one-line "lined up because rs.flush is called from Commit, not from EndBlock" in the comment to save the next reader a half-hour.

Also: initializeBlock calls the provider twice — once for prevBlockHeight (H-1) and once for blockNumber (H) (for WithNextMs). That means a single trace leases both snap[H-1] and snap[H]. As long as TraceBakeSnapshotWindow >= 2 this is fine, but if an operator misconfigures window=1 the second lease will miss and silently fall through to SS-pebble for oracle_mem/WithNextMs. Consider clamping window to >= 2 (or whatever the documented minimum is) at config-load time.

[blindchaser]

_ = will silently hide the new "Snapshot over-close" path. Since over-close is a real bug (it means refcounts are unbalanced), at least:

defer func() {
if err := cloned.MultiTree.Close(); err != nil {
logger.Error("failed to release cloned snapshot refs after rewrite", "err", err)
}
}()