fix(sim/soak): disable peer_private + ignore availability gaps so soak passes cross-impl#15
Merged
LeJamon merged 1 commit intoJun 8, 2026
Conversation
…k passes cross-impl
The soak suite failed 0/1 on --client rxrpl,rippled: the oracle saw
rippled stuck on empty ledgers while rxrpl advanced with the submitted
payments. Two harness causes, both fixed here:
1. peer_private. The rippled client defaults to peer_private=1, under
which rippled drops transactions relayed by peers it does not treat
as cluster members. It kept validating ledgers but every one was
empty (anyTransactions=0), so the payments submitted to the rxrpl
node never reached rippled. Starting the nodes with
XRPL_PEER_PRIVATE=0 (the same fix the sync suite already applies)
lets rippled accept and apply the relayed traffic — confirmed in
logs: 'Applied 5 transactions', 'Ledger 9 has 5 transactions'.
2. Oracle conflated availability with divergence. A node that booted or
synced at a higher seq, or pruned a low seq via online_delete,
returns no ledger for that seq — an Errors entry with no Divergences.
The oracle treated that as disagreement and failed the test. Now a
non-empty Divergences (two nodes serving different ledger_hash for a
seq they both have) is the only real fork; a pure availability gap is
logged and skipped. Also adds a per-component (account_hash /
transaction_hash) breakdown on a genuine divergence to speed future
diagnosis.
Result: soak traffic-and-oracle 0/1 -> 1/1 ('all nodes agree on ledgers
5-18 after 15 transactions') against rxrpl carrying the close-time fix
(rxrpl PR #117). Only simulators/soak/main.go changes — no other suite
affected.
3 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
--sim soak --client rxrpl,rippledfailed 0/1: the oracle saw rippled stuck on empty ledgers while rxrpl advanced with the submitted payments. Two harness causes, both fixed insimulators/soak/main.go:1.
peer_privatedropped relayed transactionsThe rippled client defaults to
peer_private=1, under which rippled drops transactions relayed by peers it doesn't treat as cluster members. rippled kept validating ledgers but every one was empty (anyTransactions=0), so payments submitted to the rxrpl node never reached rippled. Starting the nodes withXRPL_PEER_PRIVATE=0(the same fix the sync suite already applies) lets rippled accept and apply the relayed traffic — confirmed in logs:Applied 5 transactions,Ledger 9 has 5 transactions.2. Oracle conflated availability with divergence
A node that booted/synced at a higher seq, or pruned a low seq via
online_delete, returns no ledger for that seq — anErrorsentry with noDivergences. The oracle treated that as disagreement and failed the test. Now:Divergences(two nodes serving a differentledger_hashfor a seq they both have) = real fork → failaccount_hash/transaction_hash) breakdown on a genuine divergence for faster future diagnosisTest plan
--sim soak --client rxrpl,rippled→ 1/1 PASS (all nodes agree on ledgers 5-18 after 15 transactions), against rxrpl carrying the close-time fix (rxrpl PR #117).simulators/soak/main.gochanges — no other suite affected.Note
Branch name (
...consensus-timing) is historical — the consensus-timing config angle was investigated and ruled out; the actual fix ispeer_private+ oracle robustness.