fix(site-secrets): report every site node's secret mirror failures on the health report (M1)

A DeployedSecretsMirror failure on the STANDBY node (e.g. a master-key mismatch
that stops a deployed secret decrypting there) was only logged on that node:
site health reports are sent by the active node only, so central never saw it.

- New 12th LocalDb-replicated table secret_mirror_status (node_name PK,
  failures_json of {SecretName, Message}, checked_at; no FKs). Each node's
  mirror writes only its own row, and only when its failure set changes plus
  once per start-up (every write replicates via CDC). Best-effort: a failed
  write is a Warning and is retried next pass. Node key = INodeIdentityProvider
  NodeName, falling back to the machine name.
- SecretMirrorStatusReporter (SiteRuntime, every node, 30 s) reads every row
  into ISiteHealthCollector.SetSecretMirrorStatuses -> additive nullable
  SiteHealthReport.SecretMirrorStatuses (null = no data).
- Wire: SiteHealthReportDto.secret_mirror_statuses = 33 (presence wrapper so
  null != empty); vendored CentralControl.cs regenerated; mapper both ways.
- Central UI Health card: 'Secret mirror failing' badge + compact per-node
  node / secret / message list.
- Tests: mirror status write-on-change/first-pass/failure isolation, storage
  round-trip, reporter/collector, DTO mapper round-trip incl. null/empty,
  bUnit, replicated-table set (twelve) + FK guard wording.
This commit is contained in:
Joseph Doherty
2026-09-28 21:33:10 -04:00
parent 23a139a46c
commit 0289d51137
24 changed files with 1718 additions and 100 deletions
@@ -275,6 +275,14 @@
No self-heal
</span>
}
@if (HasSecretMirrorFailures(state))
{
<span class="badge bg-warning text-dark"
data-test="secret-mirror-badge"
title="A site node could not apply and/or decrypt one or more deployed secrets. A standby node's failures surface here before a failover would expose them — see the details in the card body.">
Secret mirror failing
</span>
}
@if (!state.IsOnline && state.LastStatusChangeAt is { } changedAt)
{
<small class="text-muted">offline since @changedAt.ToString("u")</small>
@@ -302,6 +310,26 @@
@if (state.LatestReport != null)
{
var report = state.LatestReport;
@if (HasSecretMirrorFailures(state))
{
@* Review finding M1: every site node's deployed-secrets mirror failures,
the standby's included (carried on the active node's report). *@
<div class="alert alert-warning py-2 px-3 mb-3 small" data-test="secret-mirror-failures" role="alert">
<strong>Deployed secrets failing on a site node</strong>
<ul class="mb-0 ps-3">
@foreach (var node in report.SecretMirrorStatuses!.Where(n => n.Failures.Count > 0))
{
@foreach (var failure in node.Failures)
{
<li>
<span class="font-monospace">@node.NodeName</span> &middot;
<span class="font-monospace">@failure.SecretName</span>: @failure.Message
</li>
}
}
</ul>
</div>
}
<div class="row g-3">
@* Each logical group below sits in its own card, with the group's
head promoted to a real card-header, instead of h6 rules stacked
@@ -590,6 +618,15 @@
state.IsOnline
&& state.LatestReport?.StartupReconcilerActive == false;
// Review finding M1: warn when any site node (the standby included) reports deployed
// secrets it cannot apply/decrypt. Null ("no data": central self-report, pre-M1 site) and
// nodes with an empty failure list are healthy. Offline sites are skipped like the other
// warning badges — the Offline badge is the signal that matters there.
private static bool HasSecretMirrorFailures(SiteHealthState state) =>
state.IsOnline
&& state.LatestReport?.SecretMirrorStatuses is { } statuses
&& statuses.Any(n => n.Failures.Count > 0);
// Online central nodes, from the same ClusterNodes list the Nodes column renders.
// Fewer than two means no standby, which disables the manual-failover control. This is
// a display-side guard only — AkkaManualFailoverService re-checks against live cluster
@@ -0,0 +1,26 @@
namespace ZB.MOM.WW.ScadaBridge.Commons.Messages.Health;
/// <summary>
/// One site node's latest deployed-secrets mirror outcome, as carried on
/// <see cref="SiteHealthReport.SecretMirrorStatuses"/> (review finding M1).
/// </summary>
/// <remarks>
/// Every site node runs its own <c>DeployedSecretsMirror</c> and records its current
/// failure set in the LocalDb-replicated <c>secret_mirror_status</c> table, keyed by node
/// name; the ACTIVE node reads every row and ships them on its health report, so a
/// STANDBY node's failures (most importantly a master-key mismatch that stops a deployed
/// secret decrypting there) reach central before a failover exposes them. Carries names
/// and human-readable messages only — never a secret value or ciphertext.
/// </remarks>
/// <param name="NodeName">The reporting node's semantic name (<c>node-a</c>/<c>node-b</c>), or its host name when unconfigured.</param>
/// <param name="CheckedAtUtc">When that node last recorded a CHANGE in its failure set (or its first pass after start-up) — not the time of its latest pass.</param>
/// <param name="Failures">That node's current mirror failures; empty when the node's mirror is healthy.</param>
public sealed record SecretMirrorNodeStatus(
string NodeName,
DateTimeOffset CheckedAtUtc,
IReadOnlyList<SecretMirrorFailure> Failures);
/// <summary>A single deployed secret that failed to apply and/or decrypt on one site node.</summary>
/// <param name="SecretName">The secret's name.</param>
/// <param name="Message">A human-readable failure message — never includes the secret's value.</param>
public sealed record SecretMirrorFailure(string SecretName, string Message);
@@ -173,6 +173,20 @@ public record SiteHealthReport(
/// </para>
/// </remarks>
public bool? StartupReconcilerActive { get; init; }
/// <summary>
/// Every site node's latest deployed-secrets mirror outcome (review finding M1) — one
/// entry per node that has recorded a row in the replicated <c>secret_mirror_status</c>
/// table, including the STANDBY node, whose own health report is never sent.
/// </summary>
/// <remarks>
/// <b>Null = no data</b>: a central self-report, a pre-M1 site build, a site whose
/// reporter has not run yet, or a failed read. An EMPTY list means the feature is wired
/// but no node has recorded a status yet. A node entry with an empty
/// <see cref="SecretMirrorNodeStatus.Failures"/> list is a healthy mirror. Names and
/// messages only — never secret values or ciphertext.
/// </remarks>
public IReadOnlyList<SecretMirrorNodeStatus>? SecretMirrorStatuses { get; init; }
}
/// <summary>
File diff suppressed because it is too large Load Diff
@@ -559,6 +559,32 @@ public static class CentralControlDtoMapper
dto.SiteAuditBacklog = snapshot;
}
// Review finding M1: wrapper presence keeps null ("no data") distinct from empty.
if (msg.SecretMirrorStatuses is { } mirrorStatuses)
{
var wrapper = new SecretMirrorStatusListDto();
foreach (var node in mirrorStatuses)
{
var nodeDto = new SecretMirrorNodeStatusDto
{
NodeName = node.NodeName,
CheckedAtUtc = Timestamp.FromDateTimeOffset(node.CheckedAtUtc),
};
foreach (var failure in node.Failures)
{
nodeDto.Failures.Add(new SecretMirrorFailureDto
{
SecretName = failure.SecretName,
Message = failure.Message,
});
}
wrapper.Nodes.Add(nodeDto);
}
dto.SecretMirrorStatuses = wrapper;
}
return dto;
}
@@ -620,6 +646,19 @@ public static class CentralControlDtoMapper
OnDiskBytes: snapshot.OnDiskBytes);
}
List<SecretMirrorNodeStatus>? mirrorStatuses = null;
if (dto.SecretMirrorStatuses is { } mirrorWrapper)
{
mirrorStatuses = new List<SecretMirrorNodeStatus>(mirrorWrapper.Nodes.Count);
foreach (var node in mirrorWrapper.Nodes)
{
mirrorStatuses.Add(new SecretMirrorNodeStatus(
node.NodeName,
node.CheckedAtUtc?.ToDateTimeOffset() ?? DateTimeOffset.MinValue,
node.Failures.Select(f => new SecretMirrorFailure(f.SecretName, f.Message)).ToList()));
}
}
return new SiteHealthReport(
SiteId: dto.SiteId,
SequenceNumber: dto.SequenceNumber,
@@ -657,6 +696,7 @@ public static class CentralControlDtoMapper
DetachedScriptThreads = dto.DetachedScriptThreads,
ScriptRunShedCount = dto.ScriptRunShedCount,
SiteStreamAlarmDropCount = dto.SiteStreamAlarmDropCount,
SecretMirrorStatuses = mirrorStatuses,
};
}
@@ -193,6 +193,25 @@ message NodeStatusListDto {
repeated NodeStatusDto nodes = 1;
}
// Review finding M1: one site node's deployed-secrets mirror failures. Names and
// human-readable messages only — never a secret value or ciphertext.
message SecretMirrorFailureDto {
string secret_name = 1;
string message = 2;
}
message SecretMirrorNodeStatusDto {
string node_name = 1;
google.protobuf.Timestamp checked_at_utc = 2;
repeated SecretMirrorFailureDto failures = 3;
}
// Presence wrapper (same reason as NodeStatusListDto): absent = null ("no data"),
// present-but-empty = the feature is wired and no node has recorded a status yet.
message SecretMirrorStatusListDto {
repeated SecretMirrorNodeStatusDto nodes = 1;
}
// Site -> Central: periodic site health report. Mirrors Commons SiteHealthReport.
// Additive-only evolution: field numbers are never reused.
message SiteHealthReportDto {
@@ -245,6 +264,10 @@ message SiteHealthReportDto {
// dedicated alarm publish queue because it was at capacity. Point-in-time
// (not reset on collect), like site_event_log_write_failures.
int64 site_stream_alarm_drop_count = 32;
// Review finding M1: every site node's deployed-secrets mirror status (the standby's
// included — it is read from a LocalDb-replicated table on the active node).
// Absent means null ("no data": central self-report, pre-M1 site, or not read yet).
SecretMirrorStatusListDto secret_mirror_statuses = 33;
}
// Central -> Site: health report ack, so a lost report is observable.
@@ -198,6 +198,19 @@ public interface ISiteHealthCollector
// Default no-op so test fakes do not need to be updated.
}
/// <summary>
/// Replace the latest per-node deployed-secrets mirror statuses (review finding M1) used
/// by the <c>SecretMirrorStatuses</c> field of the next <see cref="CollectReport"/> call.
/// Refreshed periodically by the site-side <c>SecretMirrorStatusReporter</c> from the
/// LocalDb-replicated <c>secret_mirror_status</c> table, so the ACTIVE node's report
/// carries the STANDBY node's mirror failures too. Point-in-time: NOT reset on collect.
/// </summary>
/// <param name="statuses">Every node's row, or <see langword="null"/> for "no data".</param>
void SetSecretMirrorStatuses(IReadOnlyList<SecretMirrorNodeStatus>? statuses)
{
// Default no-op so test fakes do not need to be updated.
}
/// <summary>
/// Replace the latest script-execution-scheduler gauges (queue depth, busy
/// thread count, and age in seconds of the oldest in-flight script) used by
@@ -32,6 +32,10 @@ public class SiteHealthCollector : ISiteHealthCollector
// -1 = reconciler SKIPPED (no IDeploymentConfigFetcher)
// Null must stay distinct from -1: "we don't know" is not "self-heal is missing".
private volatile int _startupReconcilerActive;
// Review finding M1: every site node's deployed-secrets mirror status, refreshed by the
// SiteRuntime SecretMirrorStatusReporter. Null = the reporter has not run (or its read
// failed before it ever succeeded) — "no data", distinct from an empty list.
private volatile IReadOnlyList<SecretMirrorNodeStatus>? _secretMirrorStatuses;
private readonly ConcurrentDictionary<string, ConnectionHealth> _connectionStatuses = new();
private readonly ConcurrentDictionary<string, TagResolutionStatus> _tagResolutionCounts = new();
private readonly ConcurrentDictionary<string, string> _connectionEndpoints = new();
@@ -130,6 +134,12 @@ public class SiteHealthCollector : ISiteHealthCollector
_startupReconcilerActive = active ? 1 : -1;
}
/// <inheritdoc />
public void SetSecretMirrorStatuses(IReadOnlyList<SecretMirrorNodeStatus>? statuses)
{
_secretMirrorStatuses = statuses?.ToList();
}
/// <inheritdoc />
public void UpdateConnectionHealth(string connectionName, ConnectionHealth health)
{
@@ -333,7 +343,9 @@ public class SiteHealthCollector : ISiteHealthCollector
SiteStreamAlarmDropCount = Interlocked.Read(ref _siteStreamAlarmDropCount),
// Tri-state: 0 stays null ("not reported"), which must never collapse into
// false ("self-heal is missing"). Register row 73.
StartupReconcilerActive = startupReconciler == 0 ? null : startupReconciler > 0
StartupReconcilerActive = startupReconciler == 0 ? null : startupReconciler > 0,
// M1: point-in-time, not drained; null stays null ("no data").
SecretMirrorStatuses = _secretMirrorStatuses
};
}
}
@@ -48,7 +48,7 @@ namespace ZB.MOM.WW.ScadaBridge.Host;
public static class SiteLocalDbSetup
{
/// <summary>
/// The eleven tables that participate in replication — registered when this node has a peer,
/// The twelve tables that participate in replication — registered when this node has a peer,
/// and deregistered (stale triggers dropped) when it does not.
/// </summary>
/// <remarks>
@@ -69,6 +69,13 @@ public static class SiteLocalDbSetup
/// <c>deployed_secrets</c> row, and replicating the ledger would make one node's progress
/// masquerade as the other's.
/// </para>
/// <para>
/// <c>secret_mirror_status</c> is the twelfth entry (review finding M1). It is per-node
/// DATA — one row per node, keyed by <c>node_name</c>, each written only by its own node —
/// but it is replicated precisely so the active node can read the standby's row and put
/// it on the site health report (only the active node sends one). Explicit primary key,
/// no FK, no BLOB columns.
/// </para>
/// </remarks>
internal static readonly string[] ReplicatedTables =
[
@@ -83,6 +90,7 @@ public static class SiteLocalDbSetup
"data_connection_definitions",
"native_alarm_state",
"deployed_secrets",
"secret_mirror_status",
];
/// <summary>
@@ -140,8 +148,9 @@ public static class SiteLocalDbSetup
// ELEVENTH entry to ReplicatedTables on an already-replicating site streams a
// snapshot of only that one table, where through 0.2.x the flag was a single
// per-database bit and the same edit would have re-streamed all eleven in full, in
// both directions. Nothing about the first boot changes: enabling replication on a
// site that has been running without it seeds all eleven at once, the flagged set is
// both directions (the same holds for secret_mirror_status, the twelfth, added by
// review finding M1). Nothing about the first boot changes: enabling replication on a
// site that has been running without it seeds all twelve at once, the flagged set is
// therefore every registered table, and the library sends that as an ordinary full
// snapshot. A hypothetical "register the whole list in one call" API would silently
// give the coarse behaviour back.
@@ -3,8 +3,8 @@ using Microsoft.Data.Sqlite;
namespace ZB.MOM.WW.ScadaBridge.SiteRuntime.Persistence;
/// <summary>
/// DDL for the site's eleven configuration tables (nine original plus
/// <c>deployed_secrets</c> and <c>mirrored_secrets</c>), extracted from
/// DDL for the site's twelve configuration tables (nine original plus
/// <c>deployed_secrets</c>, <c>mirrored_secrets</c> and <c>secret_mirror_status</c>), extracted from
/// <see cref="SiteStorageService"/> so it can be applied by whoever owns the database
/// file.
/// <para>
@@ -23,7 +23,7 @@ namespace ZB.MOM.WW.ScadaBridge.SiteRuntime.Persistence;
public static class SiteStorageSchema
{
/// <summary>
/// Creates the eleven site configuration tables when absent, and additively upgrades
/// Creates the twelve site configuration tables when absent, and additively upgrades
/// tables created by an older build. Idempotent — safe to run on every startup.
/// </summary>
/// <param name="connection">An open connection to the database that should hold the tables.</param>
@@ -136,6 +136,20 @@ public static class SiteStorageSchema
revision INTEGER NOT NULL,
mirrored_at TEXT NOT NULL
);
-- Per-node deployed-secrets mirror status (review finding M1): one row per site
-- node, keyed by node name, holding that node's CURRENT mirror failures as a JSON
-- array of {SecretName, Message} (names and messages only — never values or
-- ciphertext). Unlike mirrored_secrets this IS replicated
-- (SiteLocalDbSetup.ReplicatedTables): each node writes only its own row, and the
-- replication is exactly what lets the ACTIVE node's health report carry the
-- STANDBY node's failures to central. Written only when a node's failure set
-- changes (plus once per start-up), never per mirror tick. No foreign keys.
CREATE TABLE IF NOT EXISTS secret_mirror_status (
node_name TEXT PRIMARY KEY,
failures_json TEXT NOT NULL,
checked_at TEXT NOT NULL
);
";
command.ExecuteNonQuery();
}
@@ -3,6 +3,7 @@ using Microsoft.Data.Sqlite;
using Microsoft.Extensions.Logging;
using ZB.MOM.WW.LocalDb;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Artifacts;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Health;
namespace ZB.MOM.WW.ScadaBridge.SiteRuntime.Persistence;
@@ -924,6 +925,87 @@ public class SiteStorageService
await command.ExecuteNonQueryAsync();
}
// ── Secret mirror status (review finding M1 — LocalDb-replicated, one row per node) ──
/// <summary>
/// Records <paramref name="nodeName"/>'s current deployed-secrets mirror failures in the
/// replicated <c>secret_mirror_status</c> table (UPSERT on the node name). An empty
/// <paramref name="failures"/> list records a healthy mirror. Every call is a replicated
/// write, so callers write only when the failure set changes.
/// </summary>
/// <param name="nodeName">This node's name — the row key.</param>
/// <param name="failures">The node's current failures (names and messages only, never values).</param>
/// <param name="checkedAtUtc">When the recorded failure set was observed.</param>
/// <returns>A task that completes when the row has been stored.</returns>
public async Task SetSecretMirrorStatusAsync(
string nodeName, IReadOnlyList<SecretMirrorFailure> failures, DateTimeOffset checkedAtUtc)
{
ArgumentException.ThrowIfNullOrWhiteSpace(nodeName);
ArgumentNullException.ThrowIfNull(failures);
await using var connection = OpenConnection();
await using var command = connection.CreateCommand();
command.CommandText = @"
INSERT INTO secret_mirror_status (node_name, failures_json, checked_at)
VALUES (@nodeName, @failures, @checkedAt)
ON CONFLICT(node_name) DO UPDATE SET
failures_json = excluded.failures_json,
checked_at = excluded.checked_at";
command.Parameters.AddWithValue("@nodeName", nodeName);
command.Parameters.AddWithValue("@failures", JsonSerializer.Serialize(failures));
command.Parameters.AddWithValue("@checkedAt", checkedAtUtc.ToUniversalTime().ToString("O"));
await command.ExecuteNonQueryAsync();
}
/// <summary>
/// Returns every node's row from <c>secret_mirror_status</c>, ordered by node name. A row
/// whose <c>failures_json</c> cannot be parsed is returned with a single synthetic failure
/// saying so, rather than failing the whole read or silently reading as healthy.
/// </summary>
/// <returns>A task that resolves to every node's recorded mirror status.</returns>
public async Task<IReadOnlyList<SecretMirrorNodeStatus>> GetSecretMirrorStatusesAsync()
{
await using var connection = OpenConnection();
await using var command = connection.CreateCommand();
command.CommandText =
"SELECT node_name, failures_json, checked_at FROM secret_mirror_status ORDER BY node_name";
var results = new List<SecretMirrorNodeStatus>();
await using var reader = await command.ExecuteReaderAsync();
while (await reader.ReadAsync())
{
var nodeName = reader.GetString(0);
IReadOnlyList<SecretMirrorFailure> failures;
try
{
failures = JsonSerializer.Deserialize<List<SecretMirrorFailure>>(reader.GetString(1))
?? [];
}
catch (Exception ex) when (ex is JsonException or NotSupportedException)
{
failures =
[
new SecretMirrorFailure(
"(unreadable)",
$"Mirror status row for node '{nodeName}' could not be parsed ({ex.GetType().Name})"),
];
}
var checkedAt = DateTimeOffset.TryParse(
reader.GetString(2), null, System.Globalization.DateTimeStyles.RoundtripKind, out var parsed)
? parsed
: DateTimeOffset.MinValue;
results.Add(new SecretMirrorNodeStatus(nodeName, checkedAt, failures));
}
return results;
}
// ── Shared Script CRUD ──
/// <summary>
@@ -1,7 +1,9 @@
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;
using Microsoft.Extensions.Options;
using ZB.MOM.WW.ScadaBridge.Commons.Interfaces.Services;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Artifacts;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Health;
using ZB.MOM.WW.ScadaBridge.SiteRuntime.Persistence;
using ZB.MOM.WW.Secrets.Abstractions;
@@ -153,6 +155,18 @@ public sealed record MirrorResult(
/// so a LATER failure — even one with an identical fingerprint — logs at Error again. This state
/// is in-memory only (per node process), which matches the ledger's own node-local scope.
/// </para>
/// <para>
/// <b>Status row for central (review finding M1).</b> A failure on the STANDBY node used to be
/// visible only in that node's own log — site health reports are sent by the active node only.
/// After each pass this mirror records its node's current failure set (secret name + message,
/// never a value) in the LocalDb-replicated <c>secret_mirror_status</c> table, keyed by node
/// name; <c>SecretMirrorStatusReporter</c> on the active node reads every node's row onto the
/// health report. Because every write replicates via CDC, the row is written only when the
/// failure set (compared as sorted name+message pairs) differs from what this process last
/// wrote — plus once on the first pass after start-up, so a restart refreshes a stale row. The
/// write is best-effort: a failure logs a Warning (repeats at Debug) and is retried on the next
/// pass, and never affects the pass's own <see cref="MirrorResult"/>.
/// </para>
/// </remarks>
public sealed class DeployedSecretsMirror : BackgroundService
{
@@ -165,6 +179,17 @@ public sealed class DeployedSecretsMirror : BackgroundService
private readonly IOptions<SiteRuntimeOptions> _options;
private readonly ILogger<DeployedSecretsMirror> _logger;
private readonly TimeProvider _time;
private readonly string _nodeName;
/// <summary>
/// The failure set (sorted name+message pairs) this process last wrote successfully to
/// <c>secret_mirror_status</c>, or <see langword="null"/> before the first successful write
/// — which is what forces the once-per-start-up write (review finding M1).
/// </summary>
private IReadOnlyList<SecretMirrorFailure>? _lastWrittenStatus;
/// <summary>Whether a failed status write has already been logged at Warning (de-spam).</summary>
private bool _statusWriteFailureLogged;
private readonly SemaphoreSlim _gate = new(1, 1);
private bool _disposed;
@@ -199,6 +224,11 @@ public sealed class DeployedSecretsMirror : BackgroundService
/// <param name="options">Supplies <see cref="SiteRuntimeOptions.DeployedSecretsMirrorInterval"/>.</param>
/// <param name="logger">Logger instance. Never logs secret values — names, revisions, and KEK ids only.</param>
/// <param name="time">Clock driving the reconcile loop's <see cref="PeriodicTimer"/> and the tombstone-recovery stamp.</param>
/// <param name="nodeIdentity">
/// Supplies this node's name (<c>node-a</c>/<c>node-b</c>), the key of its
/// <c>secret_mirror_status</c> row. Optional: when absent or unconfigured, the machine name
/// is used instead.
/// </param>
public DeployedSecretsMirror(
SiteStorageService storage,
ISecretStore store,
@@ -206,7 +236,8 @@ public sealed class DeployedSecretsMirror : BackgroundService
ISecretCacheInvalidator? invalidator,
IOptions<SiteRuntimeOptions> options,
ILogger<DeployedSecretsMirror> logger,
TimeProvider time)
TimeProvider time,
INodeIdentityProvider? nodeIdentity = null)
{
_storage = storage ?? throw new ArgumentNullException(nameof(storage));
_store = store ?? throw new ArgumentNullException(nameof(store));
@@ -215,6 +246,7 @@ public sealed class DeployedSecretsMirror : BackgroundService
_options = options ?? throw new ArgumentNullException(nameof(options));
_logger = logger ?? throw new ArgumentNullException(nameof(logger));
_time = time ?? throw new ArgumentNullException(nameof(time));
_nodeName = nodeIdentity?.NodeName ?? Environment.MachineName;
}
/// <summary>
@@ -314,6 +346,8 @@ public sealed class DeployedSecretsMirror : BackgroundService
rows.Count(r => !r.IsRemoved) + unreadable.Count(u => !u.IsRemoved), failed.Count);
}
await RecordStatusIfChangedAsync(failedNames, failed).ConfigureAwait(false);
return result;
}
finally
@@ -322,6 +356,42 @@ public sealed class DeployedSecretsMirror : BackgroundService
}
}
/// <summary>
/// Writes this node's <c>secret_mirror_status</c> row when the pass's failure set differs
/// from the last one this process wrote (or nothing has been written yet since start-up).
/// Best-effort: never throws, so the pass result is unaffected by a status-write fault.
/// See the class remarks (review finding M1).
/// </summary>
private async Task RecordStatusIfChangedAsync(IReadOnlyList<string> failedNames, IReadOnlyList<string> failed)
{
var current = failedNames
.Select((name, i) => new SecretMirrorFailure(name, failed[i]))
.OrderBy(f => f.SecretName, StringComparer.Ordinal)
.ThenBy(f => f.Message, StringComparer.Ordinal)
.ToList();
if (_lastWrittenStatus is not null && _lastWrittenStatus.SequenceEqual(current))
{
return;
}
try
{
await _storage.SetSecretMirrorStatusAsync(_nodeName, current, _time.GetUtcNow()).ConfigureAwait(false);
_lastWrittenStatus = current;
_statusWriteFailureLogged = false;
}
catch (Exception ex)
{
_logger.Log(
_statusWriteFailureLogged ? LogLevel.Debug : LogLevel.Warning,
ex,
"Failed to record the deployed secrets mirror status for node {NodeName}; retrying next pass",
_nodeName);
_statusWriteFailureLogged = true;
}
}
/// <summary>
/// Handles a tombstoned <c>deployed_secrets</c> row: deletes it from this node's store, but
/// only if this node had previously ledgered it as mirrored — an out-of-band, same-named
@@ -0,0 +1,95 @@
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;
using ZB.MOM.WW.ScadaBridge.HealthMonitoring;
using ZB.MOM.WW.ScadaBridge.SiteRuntime.Persistence;
namespace ZB.MOM.WW.ScadaBridge.SiteRuntime.Secrets;
/// <summary>
/// Review finding M1: site-side hosted service that periodically reads EVERY node's row from
/// the LocalDb-replicated <c>secret_mirror_status</c> table and pushes the list into
/// <see cref="ISiteHealthCollector.SetSecretMirrorStatuses"/>, so the next health report
/// carries each site node's deployed-secrets mirror failures — the standby's included.
/// </summary>
/// <remarks>
/// <para>
/// Runs on every site node (the read is a cheap local SELECT), but only the active node's
/// report is ever sent (<c>HealthReportSender</c>), which is exactly why the status lives in a
/// REPLICATED table: the standby writes its row, CDC carries it to the active node, and this
/// reporter on the active node forwards it. A site pair WITHOUT LocalDb replication reports
/// only the active node's row — the same prerequisite the deployed secrets themselves have.
/// </para>
/// <para>
/// Lives in SiteRuntime (which already references HealthMonitoring and owns
/// <see cref="SiteStorageService"/>), following <c>SiteStreamAlarmDropReporter</c>: immediate
/// first probe, fixed cadence, exceptions logged and swallowed. A failed read leaves the
/// previous snapshot on the collector rather than clearing it.
/// </para>
/// </remarks>
public sealed class SecretMirrorStatusReporter : BackgroundService
{
/// <summary>Default poll cadence (30 s), matching the site health report interval.</summary>
internal static readonly TimeSpan DefaultPollInterval = TimeSpan.FromSeconds(30);
private readonly ISiteHealthCollector _collector;
private readonly SiteStorageService _storage;
private readonly ILogger<SecretMirrorStatusReporter> _logger;
private readonly TimeSpan _pollInterval;
/// <summary>Initializes a new instance of <see cref="SecretMirrorStatusReporter"/>.</summary>
/// <param name="collector">The site health collector receiving the statuses.</param>
/// <param name="storage">Site storage exposing <c>secret_mirror_status</c>.</param>
/// <param name="logger">Logger instance.</param>
/// <param name="pollInterval">Poll interval override; defaults to <see cref="DefaultPollInterval"/> (30 s).</param>
public SecretMirrorStatusReporter(
ISiteHealthCollector collector,
SiteStorageService storage,
ILogger<SecretMirrorStatusReporter> logger,
TimeSpan? pollInterval = null)
{
_collector = collector ?? throw new ArgumentNullException(nameof(collector));
_storage = storage ?? throw new ArgumentNullException(nameof(storage));
_logger = logger ?? throw new ArgumentNullException(nameof(logger));
_pollInterval = pollInterval ?? DefaultPollInterval;
}
/// <inheritdoc />
protected override async Task ExecuteAsync(CancellationToken stoppingToken)
{
await ProbeAsync().ConfigureAwait(false);
while (!stoppingToken.IsCancellationRequested)
{
try
{
await Task.Delay(_pollInterval, stoppingToken).ConfigureAwait(false);
}
catch (OperationCanceledException)
{
break;
}
await ProbeAsync().ConfigureAwait(false);
}
}
/// <summary>
/// Reads every node's mirror status and pushes it onto the collector; never throws.
/// Public so a test can drive one deterministic probe.
/// </summary>
/// <returns>A task that completes when the probe has finished.</returns>
public async Task ProbeAsync()
{
try
{
var statuses = await _storage.GetSecretMirrorStatusesAsync().ConfigureAwait(false);
_collector.SetSecretMirrorStatuses(statuses);
}
catch (Exception ex)
{
// Catch-all is deliberate: health reporting must never take the node down, and
// the loop must survive so the next tick retries.
_logger.LogWarning(ex, "SecretMirrorStatusReporter probe failed; next tick will retry.");
}
}
}
@@ -140,6 +140,13 @@ public static class ServiceCollectionExtensions
services.AddSingleton<DeployedSecretsMirror>();
services.AddHostedService(sp => sp.GetRequiredService<DeployedSecretsMirror>());
// Review finding M1: forward every node's replicated secret_mirror_status row onto the
// site health report, so the ACTIVE node's report carries the STANDBY's mirror failures.
services.AddHostedService(sp => new SecretMirrorStatusReporter(
sp.GetRequiredService<HealthMonitoring.ISiteHealthCollector>(),
sp.GetRequiredService<SiteStorageService>(),
sp.GetRequiredService<ILogger<SecretMirrorStatusReporter>>()));
return services;
}
@@ -559,6 +559,70 @@ public class HealthPageTests : BunitContext
StartupReconcilerActive = active,
};
// ----- Review finding M1: per-node deployed-secrets mirror failures ------------------
//
// The standby's failures ride the active node's report; the card must name the node,
// the secret and the message. Null ("no data") and all-healthy nodes render nothing.
[Fact]
public void Renders_SecretMirrorFailures_LabelledByNode()
{
SeedStates(new SiteHealthState
{
SiteId = "site-a",
IsOnline = true,
LastHeartbeatAt = DateTimeOffset.UtcNow,
LatestReport = ReportWithReconciler("site-a", active: true) with
{
SecretMirrorStatuses =
[
new SecretMirrorNodeStatus("node-a", DateTimeOffset.UtcNow, []),
new SecretMirrorNodeStatus(
"node-b",
DateTimeOffset.UtcNow,
[new SecretMirrorFailure("OPC-PWD", "did not decrypt on this node")]),
],
},
});
var cut = RenderHealthPage();
cut.WaitForAssertion(() =>
{
Assert.Contains("data-test=\"secret-mirror-badge\"", cut.Markup);
var panel = cut.Find("[data-test=secret-mirror-failures]");
Assert.Contains("node-b", panel.TextContent);
Assert.Contains("OPC-PWD", panel.TextContent);
Assert.Contains("did not decrypt on this node", panel.TextContent);
Assert.DoesNotContain("node-a", panel.TextContent);
});
}
[Theory]
[InlineData(false)] // null = "no data" (central self-report, pre-M1 site)
[InlineData(true)] // every node healthy
public void DoesNotRender_SecretMirrorWarning_WhenNoNodeIsFailing(bool reported)
{
SeedStates(new SiteHealthState
{
SiteId = "site-a",
IsOnline = true,
LastHeartbeatAt = DateTimeOffset.UtcNow,
LatestReport = ReportWithReconciler("site-a", active: true) with
{
SecretMirrorStatuses = reported
? [new SecretMirrorNodeStatus("node-a", DateTimeOffset.UtcNow, [])]
: null,
},
});
var cut = RenderHealthPage();
cut.WaitForAssertion(() => Assert.Contains("site-a", cut.Markup));
Assert.DoesNotContain("data-test=\"secret-mirror-badge\"", cut.Markup);
Assert.DoesNotContain("data-test=\"secret-mirror-failures\"", cut.Markup);
}
[Fact]
public void Renders_OfflineSince_ForOfflineSiteWithStatusChange()
{
@@ -369,6 +369,7 @@ public class CentralControlDtoMapperTests
Assert.Null(roundTripped.LocalDbReplicationConnected);
Assert.Null(roundTripped.LocalDbOplogBacklog);
Assert.Null(roundTripped.StartupReconcilerActive);
Assert.Null(roundTripped.SecretMirrorStatuses);
}
[Fact]
@@ -384,10 +385,14 @@ public class CentralControlDtoMapperTests
DataConnectionEndpoints = new Dictionary<string, string>(),
DataConnectionTagQuality = new Dictionary<string, TagQualityCounts>(),
ClusterNodes = [],
SecretMirrorStatuses = [],
};
var roundTripped = CentralControlDtoMapper.FromDto(Wire(CentralControlDtoMapper.ToDto(original)));
Assert.NotNull(roundTripped.SecretMirrorStatuses);
Assert.Empty(roundTripped.SecretMirrorStatuses);
Assert.NotNull(roundTripped.DataConnectionEndpoints);
Assert.Empty(roundTripped.DataConnectionEndpoints);
Assert.NotNull(roundTripped.DataConnectionTagQuality);
@@ -559,6 +564,18 @@ public class CentralControlDtoMapperTests
DetachedScriptThreads = 2,
ScriptRunShedCount = 13,
SiteStreamAlarmDropCount = 41L,
// Review finding M1: every node's secret mirror status, standby included.
SecretMirrorStatuses =
[
new SecretMirrorNodeStatus("node-a", SampleInstant, []),
new SecretMirrorNodeStatus(
"node-b",
SampleInstant.AddMinutes(-3),
[
new SecretMirrorFailure("OPC-PWD", "did not decrypt on this node"),
new SecretMirrorFailure("MX-KEY", "could not be confirmed applied"),
]),
],
};
/// <summary>
@@ -611,6 +628,17 @@ public class CentralControlDtoMapperTests
Assert.Equal(expected.DetachedScriptThreads, actual.DetachedScriptThreads);
Assert.Equal(expected.ScriptRunShedCount, actual.ScriptRunShedCount);
Assert.Equal(expected.SiteStreamAlarmDropCount, actual.SiteStreamAlarmDropCount);
Assert.Equal(expected.SecretMirrorStatuses is null, actual.SecretMirrorStatuses is null);
if (expected.SecretMirrorStatuses is not null)
{
Assert.Equal(expected.SecretMirrorStatuses.Count, actual.SecretMirrorStatuses!.Count);
for (var i = 0; i < expected.SecretMirrorStatuses.Count; i++)
{
Assert.Equal(expected.SecretMirrorStatuses[i].NodeName, actual.SecretMirrorStatuses[i].NodeName);
Assert.Equal(expected.SecretMirrorStatuses[i].CheckedAtUtc, actual.SecretMirrorStatuses[i].CheckedAtUtc);
Assert.Equal(expected.SecretMirrorStatuses[i].Failures, actual.SecretMirrorStatuses[i].Failures);
}
}
}
/// <summary>
@@ -0,0 +1,45 @@
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Health;
namespace ZB.MOM.WW.ScadaBridge.HealthMonitoring.Tests;
/// <summary>
/// Review finding M1 — <c>SiteHealthReport.SecretMirrorStatuses</c> carries every site node's
/// deployed-secrets mirror failures (the standby's included) on the active node's report.
/// </summary>
public class SecretMirrorStatusHealthFieldTests
{
[Fact]
public void CollectReport_NeverTold_LeavesTheFieldNull()
{
Assert.Null(new SiteHealthCollector().CollectReport("TestSite").SecretMirrorStatuses);
}
[Fact]
public void CollectReport_CarriesTheLatestSnapshot_AndDoesNotDrainIt()
{
var collector = new SiteHealthCollector();
collector.SetSecretMirrorStatuses(
[
new SecretMirrorNodeStatus("node-a", DateTimeOffset.UtcNow, []),
new SecretMirrorNodeStatus(
"node-b", DateTimeOffset.UtcNow, [new SecretMirrorFailure("S1", "did not decrypt")]),
]);
var first = collector.CollectReport("TestSite").SecretMirrorStatuses!;
var second = collector.CollectReport("TestSite").SecretMirrorStatuses!;
Assert.Equal(["node-a", "node-b"], first.Select(s => s.NodeName));
Assert.Equal(2, second.Count); // point-in-time, not interval-drained
Assert.Equal("S1", Assert.Single(first[1].Failures).SecretName);
}
[Fact]
public void SetNull_ResetsToNoData()
{
var collector = new SiteHealthCollector();
collector.SetSecretMirrorStatuses([]);
collector.SetSecretMirrorStatuses(null);
Assert.Null(collector.CollectReport("TestSite").SecretMirrorStatuses);
}
}
@@ -212,7 +212,7 @@ public class SiteLocalDbCdcRegistrationTests : IDisposable
// ---- helpers ----------------------------------------------------------------------
/// <summary>
/// The eleven replicated tables, plus the two deliberately-unregistered central-only
/// The twelve replicated tables, plus the two deliberately-unregistered central-only
/// notification tables, plus the node-local (also deliberately-unregistered)
/// <c>mirrored_secrets</c> ledger.
/// </summary>
@@ -221,7 +221,7 @@ public class SiteLocalDbCdcRegistrationTests : IDisposable
"OperationTracking", "site_events", "sf_messages", "deployed_configurations",
"static_attribute_overrides", "shared_scripts", "external_systems",
"database_connections", "data_connection_definitions", "native_alarm_state",
"deployed_secrets",
"deployed_secrets", "secret_mirror_status",
"notification_lists", "smtp_configurations", "mirrored_secrets",
];
@@ -230,7 +230,7 @@ public class SiteLocalDbCdcRegistrationTests : IDisposable
"OperationTracking", "site_events", "sf_messages", "deployed_configurations",
"static_attribute_overrides", "shared_scripts", "external_systems",
"database_connections", "data_connection_definitions", "native_alarm_state",
"deployed_secrets",
"deployed_secrets", "secret_mirror_status",
];
private static void AssertCaptureTriggersCoverTheReplicatedTables(ILocalDb db)
@@ -15,8 +15,9 @@ namespace ZB.MOM.WW.ScadaBridge.Host.Tests;
/// <c>RegisterReplicated</c> time, which on a site node is a BOOT failure of the replicating
/// pair. This test moves that discovery from the first replicated boot to the build: it walks
/// the real site schema (every table <c>SiteLocalDbSetup.OnReady</c> creates) and asserts none
/// of the eleven <see cref="SiteLocalDbSetup.ReplicatedTables"/> (incl. <c>deployed_secrets</c>,
/// added for site-secrets artifact deploy) has such an inbound FK.
/// of the twelve <see cref="SiteLocalDbSetup.ReplicatedTables"/> (incl. <c>deployed_secrets</c>,
/// added for site-secrets artifact deploy, and <c>secret_mirror_status</c>, added by review
/// finding M1) has such an inbound FK.
/// </summary>
/// <remarks>
/// The schema is built with an UNREPLICATED config on purpose: a replicated one would make
@@ -254,7 +254,7 @@ public class SiteLocalDbWiringTests : IDisposable
"deployed_configurations", "static_attribute_overrides", "shared_scripts",
"external_systems", "database_connections", "notification_lists",
"data_connection_definitions", "smtp_configurations", "native_alarm_state",
"deployed_secrets", "mirrored_secrets",
"deployed_secrets", "mirrored_secrets", "secret_mirror_status",
})
{
Assert.Contains(table, tables);
@@ -262,9 +262,10 @@ public class SiteLocalDbWiringTests : IDisposable
}
[Fact]
public void Site_LocalDb_RegistersExactlyTheElevenReplicatedTables()
public void Site_LocalDb_RegistersExactlyTheTwelveReplicatedTables()
{
// The Phase 2 cutover, extended by site-secrets artifact deploy's deployed_secrets.
// The Phase 2 cutover, extended by site-secrets artifact deploy's deployed_secrets
// and review finding M1's per-node secret_mirror_status.
// This is an EXACTLY check rather than a Contains in both directions, and both
// directions are load-bearing.
//
@@ -285,8 +286,8 @@ public class SiteLocalDbWiringTests : IDisposable
// data_connection_definitions precedes database_connections.
"OperationTracking", "data_connection_definitions", "database_connections",
"deployed_configurations", "deployed_secrets", "external_systems",
"native_alarm_state", "sf_messages", "shared_scripts", "site_events",
"static_attribute_overrides",
"native_alarm_state", "secret_mirror_status", "sf_messages", "shared_scripts",
"site_events", "static_attribute_overrides",
],
db.ReplicatedTables.Keys.OrderBy(k => k, StringComparer.Ordinal).ToArray());
}
@@ -334,6 +335,7 @@ public class SiteLocalDbWiringTests : IDisposable
["instance_unique_name", "source_canonical_name", "source_reference"],
db.ReplicatedTables["native_alarm_state"].PkColumns);
Assert.Equal(["name"], db.ReplicatedTables["deployed_secrets"].PkColumns);
Assert.Equal(["node_name"], db.ReplicatedTables["secret_mirror_status"].PkColumns);
}
private static HashSet<string> TableNames(ILocalDb db)
@@ -1,6 +1,7 @@
using System.Text.Json;
using Microsoft.Extensions.Logging.Abstractions;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Artifacts;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Health;
using ZB.MOM.WW.ScadaBridge.SiteRuntime.Persistence;
using ZB.MOM.WW.ScadaBridge.TestSupport;
@@ -303,6 +304,54 @@ public class DeployedSecretsStorageTests : IAsyncLifetime, IDisposable
Assert.Equal("A", mirrored.Name);
}
// ── secret_mirror_status (review finding M1) ──
[Fact]
public async Task SecretMirrorStatus_RoundTrips_OneRowPerNode_OrderedByNodeName()
{
var at = new DateTimeOffset(2026, 9, 28, 6, 0, 0, TimeSpan.Zero);
await _storage.SetSecretMirrorStatusAsync(
"node-b", [new SecretMirrorFailure("A", "did not decrypt")], at);
await _storage.SetSecretMirrorStatusAsync("node-a", [], at.AddMinutes(1));
var rows = await _storage.GetSecretMirrorStatusesAsync();
Assert.Equal(["node-a", "node-b"], rows.Select(r => r.NodeName));
Assert.Empty(rows[0].Failures);
Assert.Equal(at.AddMinutes(1), rows[0].CheckedAtUtc);
var failure = Assert.Single(rows[1].Failures);
Assert.Equal(new SecretMirrorFailure("A", "did not decrypt"), failure);
Assert.Equal(at, rows[1].CheckedAtUtc);
}
[Fact]
public async Task SecretMirrorStatus_Set_UpsertsTheNodesRow()
{
var at = new DateTimeOffset(2026, 9, 28, 6, 0, 0, TimeSpan.Zero);
await _storage.SetSecretMirrorStatusAsync("node-a", [new SecretMirrorFailure("A", "x")], at);
await _storage.SetSecretMirrorStatusAsync("node-a", [], at.AddHours(1));
var row = Assert.Single(await _storage.GetSecretMirrorStatusesAsync());
Assert.Empty(row.Failures);
Assert.Equal(at.AddHours(1), row.CheckedAtUtc);
}
[Fact]
public async Task SecretMirrorStatus_UnparseableRow_ReadsAsAFailure_NotAsHealthy()
{
await using (var connection = _storage.CreateConnection())
await using (var insert = connection.CreateCommand())
{
insert.CommandText =
"INSERT INTO secret_mirror_status (node_name, failures_json, checked_at) VALUES ('node-b', '{not json', '2026-09-28T00:00:00.0000000+00:00')";
await insert.ExecuteNonQueryAsync();
}
var row = Assert.Single(await _storage.GetSecretMirrorStatusesAsync());
Assert.Equal("node-b", row.NodeName);
Assert.Single(row.Failures);
}
private static SecretArtifact Art(string name, long revision) => new(
Name: name,
ContentType: "application/octet-stream",
@@ -23,14 +23,14 @@ public class SiteStorageSchemaTests
SiteStorageSchema.Apply(connection);
SiteStorageSchema.Apply(connection); // second run must not throw
// Every table is named explicitly on purpose — a loop asserting "11 tables"
// Every table is named explicitly on purpose — a loop asserting "12 tables"
// would still pass if one were renamed.
foreach (var table in new[]
{
"deployed_configurations", "static_attribute_overrides", "shared_scripts",
"external_systems", "database_connections", "notification_lists",
"data_connection_definitions", "smtp_configurations", "native_alarm_state",
"deployed_secrets", "mirrored_secrets",
"deployed_secrets", "mirrored_secrets", "secret_mirror_status",
})
{
Assert.True(TableExists(connection, table), $"missing table: {table}");
@@ -2,6 +2,7 @@ using System.Collections.Concurrent;
using Microsoft.Extensions.Logging;
using Microsoft.Extensions.Logging.Abstractions;
using Microsoft.Extensions.Options;
using ZB.MOM.WW.ScadaBridge.Commons.Interfaces.Services;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Artifacts;
using ZB.MOM.WW.ScadaBridge.SiteRuntime.Persistence;
using ZB.MOM.WW.ScadaBridge.SiteRuntime.Secrets;
@@ -655,6 +656,132 @@ public class DeployedSecretsMirrorTests : IDisposable
}
}
// ── Per-node status row for central (review finding M1) ──
[Fact]
public async Task Status_FirstPassAfterStartup_WritesThisNodesRow_EvenWhenHealthy()
{
var time = new FakeTimeProvider(new DateTimeOffset(2026, 9, 28, 6, 0, 0, TimeSpan.Zero));
var (mirror, storage, _, _, _) = await CreateAsync(time: time, nodeIdentity: new FixedNodeIdentity("node-b"));
await mirror.ReconcileAsync(CancellationToken.None);
var status = Assert.Single(await storage.GetSecretMirrorStatusesAsync());
Assert.Equal("node-b", status.NodeName);
Assert.Empty(status.Failures);
Assert.Equal(time.Now, status.CheckedAtUtc);
}
[Fact]
public async Task Status_UnchangedFailureSet_IsNotRewrittenOnLaterPasses()
{
var time = new FakeTimeProvider(new DateTimeOffset(2026, 9, 28, 6, 0, 0, TimeSpan.Zero));
var (mirror, storage, _, resolver, _) = await CreateAsync(time: time, nodeIdentity: new FixedNodeIdentity("node-b"));
resolver.ThrowFor("a");
await storage.UpsertDeployedSecretsAsync([Art("a", 1, kek: "kek-central")], "d1");
await mirror.ReconcileAsync(CancellationToken.None);
var firstWrite = Assert.Single(await storage.GetSecretMirrorStatusesAsync());
// Same failure on every later pass: the row is replicated, so it must not be rewritten.
time.Now = time.Now.AddMinutes(5);
await mirror.ReconcileAsync(CancellationToken.None);
await mirror.ReconcileAsync(CancellationToken.None);
var after = Assert.Single(await storage.GetSecretMirrorStatusesAsync());
Assert.Equal(firstWrite.CheckedAtUtc, after.CheckedAtUtc);
var failure = Assert.Single(after.Failures);
Assert.Equal("a", failure.SecretName);
Assert.Contains("kek-central", failure.Message);
}
[Fact]
public async Task Status_FailureSetChange_RewritesTheRow_AndClearsWhenHealed()
{
var time = new FakeTimeProvider(new DateTimeOffset(2026, 9, 28, 6, 0, 0, TimeSpan.Zero));
var (mirror, storage, _, resolver, _) = await CreateAsync(time: time, nodeIdentity: new FixedNodeIdentity("node-a"));
await storage.UpsertDeployedSecretsAsync([Art("a", 1)], "d1");
await mirror.ReconcileAsync(CancellationToken.None);
Assert.Empty(Assert.Single(await storage.GetSecretMirrorStatusesAsync()).Failures);
// A decrypt failure appears -> row rewritten with the failure.
resolver.ThrowFor("a");
time.Now = time.Now.AddMinutes(1);
await mirror.ReconcileAsync(CancellationToken.None);
var failing = Assert.Single(await storage.GetSecretMirrorStatusesAsync());
Assert.Equal("a", Assert.Single(failing.Failures).SecretName);
Assert.Equal(time.Now, failing.CheckedAtUtc);
// Healed -> row rewritten empty.
resolver.ClearFailures();
time.Now = time.Now.AddMinutes(1);
await mirror.ReconcileAsync(CancellationToken.None);
var healed = Assert.Single(await storage.GetSecretMirrorStatusesAsync());
Assert.Empty(healed.Failures);
Assert.Equal(time.Now, healed.CheckedAtUtc);
}
[Fact]
public async Task Status_RowCarriesNamesAndMessagesOnly_NeverCiphertext()
{
var (mirror, storage, _, resolver, _) = await CreateAsync(nodeIdentity: new FixedNodeIdentity("node-a"));
resolver.ThrowFor("a");
await storage.UpsertDeployedSecretsAsync([Art("a", 1)], "d1");
await mirror.ReconcileAsync(CancellationToken.None);
await using var connection = storage.CreateConnection();
await using var command = connection.CreateCommand();
command.CommandText = "SELECT failures_json FROM secret_mirror_status WHERE node_name = 'node-a'";
var json = (string)(await command.ExecuteScalarAsync())!;
Assert.Contains("\"SecretName\":\"a\"", json);
Assert.DoesNotContain("Ciphertext", json);
Assert.DoesNotContain(Convert.ToBase64String(new byte[] { 1, 2, 3, 4, 5 }), json);
}
[Fact]
public async Task Status_WriteFailure_DoesNotBreakThePass_AndIsRetriedNextPass()
{
var logger = new CapturingLogger<DeployedSecretsMirror>();
var (mirror, storage, store, _, _) = await CreateAsync(logger: logger, nodeIdentity: new FixedNodeIdentity("node-a"));
await storage.UpsertDeployedSecretsAsync([Art("a", 1)], "d1");
await using (var connection = storage.CreateConnection())
await using (var drop = connection.CreateCommand())
{
drop.CommandText = "DROP TABLE secret_mirror_status";
await drop.ExecuteNonQueryAsync();
}
var r = await mirror.ReconcileAsync(CancellationToken.None);
Assert.Equal(["a"], r.Applied);
Assert.True(store.Contains("a"));
Assert.Contains(logger.Entries, e => e.Level == LogLevel.Warning && e.Message.Contains("mirror status"));
// Table back (e.g. transient fault cleared): the unchanged failure set is still
// written, because the earlier write never landed.
await storage.InitializeAsync();
await mirror.ReconcileAsync(CancellationToken.None);
Assert.Equal("node-a", Assert.Single(await storage.GetSecretMirrorStatusesAsync()).NodeName);
}
[Fact]
public async Task Status_WithoutNodeIdentity_FallsBackToMachineName()
{
var (mirror, storage, _, _, _) = await CreateAsync();
await mirror.ReconcileAsync(CancellationToken.None);
Assert.Equal(Environment.MachineName, Assert.Single(await storage.GetSecretMirrorStatusesAsync()).NodeName);
}
private sealed class FixedNodeIdentity(string? name) : INodeIdentityProvider
{
public string? NodeName { get; } = name;
}
// ── Harness ──
private async Task<(
@@ -665,7 +792,8 @@ public class DeployedSecretsMirrorTests : IDisposable
FakeSecretCacheInvalidator Invalidator)> CreateAsync(
CapturingLogger<DeployedSecretsMirror>? logger = null,
TimeSpan? interval = null,
TimeProvider? time = null)
TimeProvider? time = null,
INodeIdentityProvider? nodeIdentity = null)
{
var storage = new SiteStorageService(_localDb.Db, NullLogger<SiteStorageService>.Instance);
await storage.InitializeAsync();
@@ -685,7 +813,8 @@ public class DeployedSecretsMirrorTests : IDisposable
invalidator,
options,
(ILogger<DeployedSecretsMirror>?)logger ?? NullLogger<DeployedSecretsMirror>.Instance,
time ?? TimeProvider.System);
time ?? TimeProvider.System,
nodeIdentity);
return (mirror, storage, store, resolver, invalidator);
}
@@ -0,0 +1,95 @@
using Microsoft.Extensions.Logging.Abstractions;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Health;
using ZB.MOM.WW.ScadaBridge.HealthMonitoring;
using ZB.MOM.WW.ScadaBridge.SiteRuntime.Persistence;
using ZB.MOM.WW.ScadaBridge.SiteRuntime.Secrets;
using ZB.MOM.WW.ScadaBridge.TestSupport;
namespace ZB.MOM.WW.ScadaBridge.SiteRuntime.Tests.Secrets;
/// <summary>
/// Review finding M1: <see cref="SecretMirrorStatusReporter"/> forwards every node's
/// replicated <c>secret_mirror_status</c> row onto the site health report via the collector.
/// </summary>
public class SecretMirrorStatusReporterTests : IAsyncLifetime, IDisposable
{
private readonly TestLocalDb _localDb = TestLocalDb.CreateTemp("secret-mirror-status");
private SiteStorageService _storage = null!;
public async Task InitializeAsync()
{
_storage = new SiteStorageService(_localDb.Db, NullLogger<SiteStorageService>.Instance);
await _storage.InitializeAsync();
}
public Task DisposeAsync() => Task.CompletedTask;
public void Dispose()
{
var path = _localDb.Path;
_localDb.Dispose();
TestLocalDb.DeleteFiles(path);
}
[Fact]
public async Task Probe_PushesEveryNodesRow_OntoTheNextReport()
{
var at = new DateTimeOffset(2026, 9, 28, 6, 0, 0, TimeSpan.Zero);
await _storage.SetSecretMirrorStatusAsync("node-a", [], at);
await _storage.SetSecretMirrorStatusAsync(
"node-b", [new SecretMirrorFailure("OPC-PWD", "did not decrypt on this node")], at);
var collector = new SiteHealthCollector();
var reporter = new SecretMirrorStatusReporter(
collector, _storage, NullLogger<SecretMirrorStatusReporter>.Instance);
await reporter.ProbeAsync();
var report = collector.CollectReport("site-a");
Assert.NotNull(report.SecretMirrorStatuses);
Assert.Equal(["node-a", "node-b"], report.SecretMirrorStatuses!.Select(s => s.NodeName));
Assert.Equal("OPC-PWD", Assert.Single(report.SecretMirrorStatuses[1].Failures).SecretName);
}
[Fact]
public void BeforeAnyProbe_TheReportFieldIsNull()
{
var report = new SiteHealthCollector().CollectReport("site-a");
Assert.Null(report.SecretMirrorStatuses);
}
[Fact]
public async Task Probe_NoRowsYet_PushesAnEmptyList_NotNull()
{
var collector = new SiteHealthCollector();
var reporter = new SecretMirrorStatusReporter(
collector, _storage, NullLogger<SecretMirrorStatusReporter>.Instance);
await reporter.ProbeAsync();
Assert.Empty(collector.CollectReport("site-a").SecretMirrorStatuses!);
}
[Fact]
public async Task Probe_ReadFailure_KeepsThePreviousSnapshot_AndDoesNotThrow()
{
await _storage.SetSecretMirrorStatusAsync("node-a", [], DateTimeOffset.UtcNow);
var collector = new SiteHealthCollector();
var reporter = new SecretMirrorStatusReporter(
collector, _storage, NullLogger<SecretMirrorStatusReporter>.Instance);
await reporter.ProbeAsync();
await using (var connection = _storage.CreateConnection())
await using (var drop = connection.CreateCommand())
{
drop.CommandText = "DROP TABLE secret_mirror_status";
await drop.ExecuteNonQueryAsync();
}
await reporter.ProbeAsync();
// Point-in-time: not drained on collect, and not cleared by a failed read.
Assert.Single(collector.CollectReport("site-a").SecretMirrorStatuses!);
Assert.Single(collector.CollectReport("site-a").SecretMirrorStatuses!);
}
}