Issues

• 3.x.x: backport sqlx's TLS lost-wakeup fix (launchbadge/sqlx#4251) so a large statement cannot stall a ledger connection

  Qbit-Org/qbit-mining-bootstrap#771 · Filter by author RobertClarke opened on Oct 10, 2026

• 3.x.x: import-audits holds SETTLEMENT_LOCK while writing a large canonical audit body (≥90 s for ~112 MB), stalling template refresh and settlement

  Qbit-Org/qbit-mining-bootstrap#765 · Filter by author RobertClarke opened on Oct 10, 2026
  1

• 3.x.x: a submitblock whose deadline passes while its connection is still opening is recorded as unknown, not unsent, so a block that never left the frontend is never offered again

  Qbit-Org/qbit-mining-bootstrap#762 · Filter by author RobertClarke opened on Oct 10, 2026

• 3.x.x: the qbitd RPC client reuses idle connections for 60 s, past qbitd's default 30 s rpcservertimeout, so a call can race the idle close (transport failed; a submitblock caught in it is never offered again)

  Qbit-Org/qbit-mining-bootstrap#759 · Filter by author RobertClarke opened on Oct 9, 2026
  1

• 3.x.x: the revision-work warnings (PrismAcceptedRevisionWorkPending, ...TrackingUnknown) fire on an idle backup frontend: not gated on authorized clients like their critical twin (post-cutover)

  Qbit-Org/qbit-mining-bootstrap#758 · Filter by author RobertClarke opened on Oct 9, 2026

• 3.x.x: five alert-rule UIDs in docs/prism-native-alert-rules.json exceed Grafana's 40-character UID limit (post-cutover)

  Qbit-Org/qbit-mining-bootstrap#757 · Filter by author RobertClarke opened on Oct 9, 2026

• 3.x.x: fill the hashrate series' partial 1d and 1h buckets from finer rollups, and compute its buckets in integer arithmetic (post-cutover, after #754)

  Qbit-Org/qbit-mining-bootstrap#755 · Filter by author RobertClarke opened on Oct 8, 2026

• 3.x.x: after each pool block, the second payout-revision bump makes non-landing frontends refuse shares on just-published new-tip work as stale-job until their next blockpoll tick (post-cutover)

  Qbit-Org/qbit-mining-bootstrap#752 · Filter by author RobertClarke opened on Oct 8, 2026
  1

• 3.x.x: recovery-evidence export: let a frozen 2.x source skip the synthesized share_hashes kind when the migrated side is deferred (post-cutover; about 12 min per source export)

  Qbit-Org/qbit-mining-bootstrap#751 · Filter by author RobertClarke opened on Oct 8, 2026

• 3.x.x: run the admission gate's real-schema tests as a non-superuser schema owner (follow-up to #740)

  Qbit-Org/qbit-mining-bootstrap#741 · Filter by author RobertClarke opened on Oct 7, 2026

• 3.x.x: backfill-ctv refuses every legacy CTV set: 2.x's sorted-key set digest vs 3.x's struct-order digest in persist_fanouts (rc.6; the cutover skips backfill-ctv)

  Qbit-Org/qbit-mining-bootstrap#739 · Filter by author RobertClarke opened on Oct 7, 2026

• 3.x.x: every share append UPDATEs the qbit_prism_cluster singleton (ledger_clock_ms), so any snapshot held on the primary bloats it and saturates ORDER_LOCK in ~20 s at 400 shares/s; derive the clock from append-only state (rc.5)

  Qbit-Org/qbit-mining-bootstrap#738 · Filter by author RobertClarke opened on Oct 7, 2026
  1

• 3.x.x: the carry-forward integrity report takes ~64 s at production size, so self-check, fatal-state clear and the /audit integrity routes fail under the default 15 s statement timeout (rc.5)

  Qbit-Org/qbit-mining-bootstrap#737 · Filter by author RobertClarke opened on Oct 7, 2026
  4

• 3.x.x: /public/v1/artifacts/<sha> answers 500 instead of 404 for an audit whose legacy body is not retrievable yet (artifact_fallback_error) (post-cutover)

  Qbit-Org/qbit-mining-bootstrap#735 · Filter by author RobertClarke opened on Oct 6, 2026

• 3.x.x: production self-check stops at legacy audit completeness and skips its local checks, so it can't gate a deployment while import-audits runs (post-cutover)

  Qbit-Org/qbit-mining-bootstrap#734 · Filter by author RobertClarke opened on Oct 6, 2026

• 3.x.x: a legacy segment_range part is refused when its slot has a share_seq gap outside the referenced range, because select_share_segment_range checks the whole segment's contiguity (post-cutover; sidecars cover the cutover)

  Qbit-Org/qbit-mining-bootstrap#733 · Filter by author RobertClarke opened on Oct 6, 2026

• 3.x.x: 2.x audit-bundle.v2 bodies fail 'window proof share_parts_digest_hex mismatch' because 3.x recomputes the parts digest with serde_json's sorted keys and 2.x hashed insertion order (post-cutover; sidecars cover the cutover)

  Qbit-Org/qbit-mining-bootstrap#731 · Filter by author RobertClarke opened on Oct 6, 2026

• 3.x.x: flaky real-node smoke: load-fe-0 exits during startup (exit 1), and CI keeps no frontend log to say why

  Qbit-Org/qbit-mining-bootstrap#730 · Filter by author RobertClarke opened on Oct 6, 2026

• 3.x.x: flaky test: a_relaunched_frontend_recovers_one_row_at_the_same_statement_cost_at_both_cardinalities counts one extra candidate-poll transaction (65 vs 61 statements)

  Qbit-Org/qbit-mining-bootstrap#729 · Filter by author RobertClarke opened on Oct 6, 2026

• 3.x.x: prism-load faults race node B's keepalive tip: a fault's own scheduled block isn't covered by #638's keepalive hold (faults-short-real-node, rc.4 batch)

  Qbit-Org/qbit-mining-bootstrap#727 · Filter by author RobertClarke opened on Oct 6, 2026

• 3.x.x: test race in node_stopped_before_submitblock_lands_the_block_once_after_resume: SIGSTOP can land after the node runs the released submitblock (rc.3 tag CI)

  Qbit-Org/qbit-mining-bootstrap#726 · Filter by author RobertClarke opened on Oct 6, 2026

• 3.x.x: flaky live test: async_promotion_with_live_miners_accounts_for_every_share_and_lands_a_block answers a share on new-tip work stale-job

  Qbit-Org/qbit-mining-bootstrap#724 · Filter by author RobertClarke opened on Oct 6, 2026
  2

• 3.x.x: a share under a tip-change lease is refused 'commit gate closed' when the authority try_read meets a writer, which can drop a valid share's credit and fires PrismShareAppendFailures

  Qbit-Org/qbit-mining-bootstrap#716 · Filter by author RobertClarke opened on Oct 6, 2026 · 3.x.x: GA (P2)
  1

• 3.x.x: sslmode=verify-ca still checks the host name (sqlx 0.8.6 misses rustls 0.23.45's NotValidForNameContext), so the load harness's delay proxy on 127.0.0.1 is refused

  Qbit-Org/qbit-mining-bootstrap#713 · Filter by author RobertClarke opened on Oct 6, 2026

• 3.x.x: the recovery-evidence export takes 85 min at 65.9M shares; export the large kinds in snapshot-sharing parallel streams and take the source export off the critical path

  Qbit-Org/qbit-mining-bootstrap#712 · Filter by author RobertClarke opened on Oct 6, 2026