fix(replication): drain priority sync queue and recover from routing-event lag#165

Open

mickvandijke wants to merge 42 commits into

mainfrom

fix/neighbor-sync-drain-window

Commits on Jul 22, 2026

fix(replication): harden paid-list verification repair
mickvandijke
committed
fix(replication): tolerate paid-list edge churn
mickvandijke
committed
fix(replication): drain priority sync queue and recover from routing-event lag

mickvandijke
and
claude
committed
fix(replication): preserve verification retry capacity
mickvandijke
committed
fix(replication): eliminate false audit challenge timeouts
mickvandijke
committed
fix(replication): preserve dequeued retry reservations
mickvandijke
committed
chore(replication): fix audit admission clippy
mickvandijke
committed
fix(replication): release cancelled async work
mickvandijke
committed
fix(replication): silence no-logging audit label warnings
mickvandijke
committed
fix(replication): unblock bootstrap when rejected peer leaves
mickvandijke
committed
refactor(replication): prioritize source-aware hints
mickvandijke
committed
refactor(replication): aggregate bootstrap hint batches
mickvandijke
committed
fix(replication): aggregate verification per peer
mickvandijke
committed
fix(replication): drain fresh offers through LMDB writes
mickvandijke
committed
fix(replication): track detached audit work
mickvandijke
committed
fix(replication): stop message handler when event streams close

mickvandijke
and
claude
committed
fix(replication): prune departed peers during DHT lag recovery

mickvandijke
and
claude
committed
fix(replication): penalize rejected singleton replica hints
mickvandijke
committed
fix(replication): keep fresh offers off the serial message loop

mickvandijke
and
claude
committed
fix(replication): gate replica downloads on storage responsibility

mickvandijke
and
claude
committed
refactor(replication): admit hints through a single relevance gate

mickvandijke
and
claude
committed
perf(replication): merge duplicate hints without rebuilding the fetch heap

mickvandijke
and
claude
committed
fix(replication): drain detached LMDB blocking ops on shutdown

mickvandijke
and
claude
committed
fix(replication): expire orphaned capacity-rejection records to unstall bootstrap

mickvandijke
and
claude
committed
fix(replication): recheck storage responsibility at the point of download

mickvandijke
and
claude
committed
docs(adr): record replication repair hardening decisions (PR #165)

mickvandijke
and
claude
committed
fix(replication): address deep review findings
mickvandijke
committed
fix(replication): resolve all-target clippy failures
mickvandijke
committed
fix(docs): avoid private rustdoc link
mickvandijke
committed
fix(storage): guard raw LMDB reads against concurrent map resize

mickvandijke
and
claude
committed
fix(replication): fairly share verification capacity
mickvandijke
committed
fix(replication): generalize bounded responder admission
mickvandijke
committed
fix(replication): isolate fetch responders
mickvandijke
committed
fix(replication): isolate verification responders
mickvandijke
committed
fix(replication): isolate neighbor sync responders
mickvandijke
committed
fix(replication): drop serial queue overflow
mickvandijke
committed
fix(replication): count responder load shedding
mickvandijke
committed
feat(replication): trace audit request origins and latency
mickvandijke
committed
feat(replication): log possession proof failure reasons
mickvandijke
committed

Commits on Jul 23, 2026