feat(audit): SHA-256 hash chain over the central AuditLog (ScadaBridge#41)

Audit roadmap T1 (alog.md §11.4): one chain per UTC calendar month of
OccurredAtUtc. Each chained row carries ChainSeq (gap-free within its month)
and RowHash = SHA-256(prev.RowHash || canonical(row)); the v1 canonical form
(AuditChainHasher) is pinned by a known-answer vector computed independently.

- Migration AddAuditLogHashChain: nullable ChainSeq/RowHash on dbo.AuditLog
  (metadata-only add), AuditLogChainHead (per-month chain end + partition-purge
  checkpoint anchor), AuditLogChainTombstone, two partition-aligned filtered
  indexes, scadabridge_audit_chainer role with a column-level UPDATE grant.
  Production script docs/plans/sql/AddAuditLogHashChain.sql.
- AuditLogHashChain chains unchained rows oldest month first under an
  exclusive sp_getapplock (zero timeout, so a dual-active central pair still
  yields one chain), with an optimistic head check and a ChainSeq IS NULL
  row-count check that roll a pass back instead of leaving a gap. The ingest
  path is untouched.
- Purges keep the chain verifiable: the partition switch records the month's
  checkpoint anchor and drops its tombstones; the per-channel DELETE writes a
  tombstone per chained row in the same transaction (OUTPUT INTO). Both, and
  the SourceNode backfill, hold the chain lock; the backfill skips chained
  rows. The switch staging table gains the two new columns.
- AuditChainVerifier reports the first break (partition, ChainSeq, EventId,
  kind) for edited, deleted, reordered rows and tail truncation.
- AuditLogHashChainService (central, default OFF via AuditLog:HashChain).

Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
This commit is contained in:
Joseph Doherty
2026-11-09 16:28:43 -05:00
parent 3cf97f6dda
commit fd4be80913
24 changed files with 4596 additions and 6 deletions
@@ -0,0 +1,346 @@
using Microsoft.Data.SqlClient;
using Microsoft.EntityFrameworkCore;
using ZB.MOM.WW.Audit;
using ZB.MOM.WW.ScadaBridge.Commons.Types.Audit;
using ZB.MOM.WW.ScadaBridge.Commons.Types.Enums;
using ZB.MOM.WW.ScadaBridge.ConfigurationDatabase.Maintenance;
using ZB.MOM.WW.ScadaBridge.ConfigurationDatabase.Repositories;
using ZB.MOM.WW.ScadaBridge.ConfigurationDatabase.Tests.Migrations;
namespace ZB.MOM.WW.ScadaBridge.ConfigurationDatabase.Tests.Maintenance;
/// <summary>
/// SQL Server integration tests for the AuditLog hash chain (ScadaBridge#41) against the
/// real migrated, partitioned schema: continuity under concurrent writers and chainers,
/// detection of an edited / deleted / reordered row, and survival of both purge paths.
/// </summary>
/// <remarks>
/// The fixture database is shared by the class and the chain is per month across the whole
/// table, so every test owns a distinct month (xunit runs one class's tests sequentially).
/// Each test drains the chain before tampering so no other month is left unchained.
/// </remarks>
public class AuditLogHashChainTests : IClassFixture<MsSqlMigrationFixture>
{
private readonly MsSqlMigrationFixture _fixture;
/// <summary>Creates the test class with the shared migrated SQL Server fixture.</summary>
/// <param name="fixture">The shared migration fixture.</param>
public AuditLogHashChainTests(MsSqlMigrationFixture fixture) => _fixture = fixture;
/// <summary>
/// Many concurrent INSERTs (the two central nodes + site ingest) racing two concurrent
/// chainers (a dual-active central pair) still produce one gap-free, intact chain.
/// </summary>
/// <returns>A task that represents the asynchronous operation.</returns>
[SkippableFact]
public async Task ConcurrentWritersAndChainers_ProduceOneIntactGapFreeChain()
{
Skip.IfNot(_fixture.Available, _fixture.SkipReason);
var month = Month(2026, 2);
const int writers = 8;
const int rowsPerWriter = 40;
using var stop = new CancellationTokenSource();
var lockBusySeen = 0;
async Task ChainLoop()
{
while (!stop.IsCancellationRequested)
{
await using var ctx = CreateContext();
var r = await new AuditLogHashChain(ctx).ChainPendingAsync(17);
if (r.LockBusy) Interlocked.Increment(ref lockBusySeen);
if (r.RowsChained == 0) await Task.Delay(5);
}
}
var chainers = new[] { Task.Run(ChainLoop), Task.Run(ChainLoop) };
var writerTasks = Enumerable.Range(0, writers).Select(w => Task.Run(async () =>
{
for (var i = 0; i < rowsPerWriter; i++)
{
await using var ctx = CreateContext();
// Out-of-order OccurredAtUtc on purpose: late-arriving site rows.
var at = month.AddDays((w * 7 + i * 13) % 27).AddSeconds(i);
await new AuditLogRepository(ctx).InsertIfNotExistsAsync(NewEvent(at, sourceNode: $"node-{w}"));
}
})).ToArray();
await Task.WhenAll(writerTasks);
await Task.Delay(200);
await stop.CancelAsync();
await Task.WhenAll(chainers);
await DrainAsync();
var result = await VerifyAsync(month);
Assert.Equal(AuditChainStatus.Intact, result.Status);
Assert.Equal(writers * rowsPerWriter, result.RowsVerified);
Assert.Equal(writers * rowsPerWriter, result.HeadSeq);
Assert.Equal(0, result.UnchainedRows);
var seqs = await ChainSeqsAsync(month);
Assert.Equal(Enumerable.Range(1, writers * rowsPerWriter).Select(i => (long)i), seqs);
}
/// <summary>Editing a chained row's content is reported at that row.</summary>
/// <returns>A task that represents the asynchronous operation.</returns>
[SkippableFact]
public async Task EditedRow_IsDetected_WithRowIdAndPartition()
{
Skip.IfNot(_fixture.Available, _fixture.SkipReason);
var month = Month(2026, 3);
var events = await SeedAndChainAsync(month, 6);
Assert.Equal(AuditChainStatus.Intact, (await VerifyAsync(month)).Status);
var victim = events[3];
await ExecAsync($"UPDATE dbo.AuditLog SET Target = N'forged' WHERE EventId = '{victim}';");
var result = await VerifyAsync(month);
Assert.Equal(AuditChainStatus.Broken, result.Status);
Assert.Equal(AuditChainBreakKind.HashMismatch, result.FirstBreak!.Kind);
Assert.Equal("2026-03", result.FirstBreak.Partition);
Assert.Equal(victim, result.FirstBreak.EventId);
Assert.Equal(await SeqOfAsync(victim), result.FirstBreak.ChainSeq);
}
/// <summary>Deleting a chained row outside the purge path is reported as a missing sequence number.</summary>
/// <returns>A task that represents the asynchronous operation.</returns>
[SkippableFact]
public async Task DeletedRow_IsDetected()
{
Skip.IfNot(_fixture.Available, _fixture.SkipReason);
var month = Month(2026, 4);
var events = await SeedAndChainAsync(month, 6);
var victim = events[2];
var victimSeq = await SeqOfAsync(victim);
await ExecAsync($"DELETE FROM dbo.AuditLog WHERE EventId = '{victim}';");
var result = await VerifyAsync(month);
Assert.Equal(AuditChainStatus.Broken, result.Status);
Assert.Equal(AuditChainBreakKind.MissingRow, result.FirstBreak!.Kind);
Assert.Equal(victimSeq, result.FirstBreak.ChainSeq);
Assert.Equal("2026-04", result.FirstBreak.Partition);
}
/// <summary>Swapping two rows' chain positions is reported.</summary>
/// <returns>A task that represents the asynchronous operation.</returns>
[SkippableFact]
public async Task ReorderedRows_AreDetected()
{
Skip.IfNot(_fixture.Available, _fixture.SkipReason);
var month = Month(2026, 5);
var events = await SeedAndChainAsync(month, 6);
var a = events[1];
var b = events[4];
var seqA = await SeqOfAsync(a);
var seqB = await SeqOfAsync(b);
await ExecAsync(
$"UPDATE dbo.AuditLog SET ChainSeq = CASE WHEN EventId = '{a}' THEN {seqB} ELSE {seqA} END " +
$"WHERE EventId IN ('{a}', '{b}');");
var result = await VerifyAsync(month);
Assert.Equal(AuditChainStatus.Broken, result.Status);
Assert.Equal(AuditChainBreakKind.HashMismatch, result.FirstBreak!.Kind);
Assert.Equal(Math.Min(seqA, seqB), result.FirstBreak.ChainSeq);
}
/// <summary>
/// The per-channel retention DELETE removes chained rows from the middle of a month and
/// the chain still verifies (tombstones); a later chaining pass continues the chain.
/// </summary>
/// <returns>A task that represents the asynchronous operation.</returns>
[SkippableFact]
public async Task PerChannelPurge_LeavesAVerifiableChain()
{
Skip.IfNot(_fixture.Available, _fixture.SkipReason);
var month = Month(2026, 6);
var rows = new List<AuditEvent>();
for (var i = 0; i < 10; i++)
{
rows.Add(NewEvent(month.AddDays(i + 1),
channel: i % 2 == 0 ? AuditChannel.DbOutbound : AuditChannel.ApiOutbound,
kind: i % 2 == 0 ? AuditKind.DbWrite : AuditKind.ApiCall));
}
await InsertAsync(rows);
await DrainAsync();
await using (var ctx = CreateContext())
{
var purged = await new AuditLogRepository(ctx).PurgeChannelOlderThanAsync(
nameof(AuditChannel.DbOutbound), month.AddDays(8), batchSize: 2);
Assert.Equal(4, purged); // days 1,3,5,7
}
var result = await VerifyAsync(month);
Assert.Equal(AuditChainStatus.Intact, result.Status);
Assert.Equal(4, result.TombstonesApplied);
Assert.Equal(6, result.RowsVerified);
// A row arriving after the purge chains on from the existing head.
await InsertAsync([NewEvent(month.AddDays(20))]);
await DrainAsync();
result = await VerifyAsync(month);
Assert.Equal(AuditChainStatus.Intact, result.Status);
Assert.Equal(11, result.HeadSeq);
}
/// <summary>
/// The monthly partition-switch purge removes the whole month; the head's checkpoint
/// anchor keeps the month verifiable, including for a late row chained afterwards.
/// </summary>
/// <returns>A task that represents the asynchronous operation.</returns>
[SkippableFact]
public async Task PartitionSwitchPurge_LeavesAnAnchoredVerifiableChain()
{
Skip.IfNot(_fixture.Available, _fixture.SkipReason);
var month = Month(2026, 7);
await SeedAndChainAsync(month, 5);
await using (var ctx = CreateContext())
{
await new AuditLogRepository(ctx).SwitchOutPartitionAsync(month);
}
var result = await VerifyAsync(month);
Assert.Equal(AuditChainStatus.Intact, result.Status);
Assert.Equal(5, result.PurgedThroughSeq);
Assert.Equal(0, result.RowsVerified);
await InsertAsync([NewEvent(month.AddDays(3))]);
await DrainAsync();
result = await VerifyAsync(month);
Assert.Equal(AuditChainStatus.Intact, result.Status);
Assert.Equal(1, result.RowsVerified);
Assert.Equal(6, result.HeadSeq);
}
/// <summary>
/// The SourceNode backfill (the other sanctioned AuditLog UPDATE) must not touch a chained
/// row — SourceNode is inside its hash.
/// </summary>
/// <returns>A task that represents the asynchronous operation.</returns>
[SkippableFact]
public async Task SourceNodeBackfill_SkipsChainedRows()
{
Skip.IfNot(_fixture.Available, _fixture.SkipReason);
var month = Month(2026, 8);
await InsertAsync([NewEvent(month.AddDays(2), sourceNode: null)]);
await DrainAsync();
var unchained = NewEvent(month.AddDays(3), sourceNode: null);
// A row written after the drain stays unchained until the next pass.
await InsertAsync([unchained]);
await using (var ctx = CreateContext())
{
await new AuditLogRepository(ctx).BackfillSourceNodeAsync("unknown", month.AddMonths(1), batchSize: 10);
}
await DrainAsync();
var result = await VerifyAsync(month);
Assert.Equal(AuditChainStatus.Intact, result.Status);
Assert.Equal(2, result.RowsVerified);
Assert.Equal("unknown", await ScalarAsync<string>($"SELECT SourceNode FROM dbo.AuditLog WHERE EventId = '{unchained.EventId}';"));
}
// ── helpers ──────────────────────────────────────────────────────────────
private static DateTime Month(int year, int month) => new(year, month, 1, 0, 0, 0, DateTimeKind.Utc);
private async Task<List<Guid>> SeedAndChainAsync(DateTime month, int count)
{
var events = Enumerable.Range(1, count).Select(i => NewEvent(month.AddDays(i).AddMinutes(i))).ToList();
await InsertAsync(events);
await DrainAsync();
return events.Select(e => e.EventId).ToList();
}
private async Task InsertAsync(IEnumerable<AuditEvent> events)
{
await using var ctx = CreateContext();
var repo = new AuditLogRepository(ctx);
foreach (var e in events)
{
await repo.InsertIfNotExistsAsync(e);
}
}
private async Task DrainAsync()
{
await using var ctx = CreateContext();
var chain = new AuditLogHashChain(ctx);
for (var i = 0; i < 1000; i++)
{
var r = await chain.ChainPendingAsync(500);
if (r.RowsChained == 0 && !r.LockBusy)
{
return;
}
}
throw new InvalidOperationException("Chain did not drain.");
}
private async Task<AuditChainVerificationResult> VerifyAsync(DateTime month)
{
await using var ctx = CreateContext();
return await new AuditLogHashChain(ctx).VerifyMonthAsync(month);
}
private async Task<long> SeqOfAsync(Guid eventId)
=> await ScalarAsync<long>($"SELECT ChainSeq FROM dbo.AuditLog WHERE EventId = '{eventId}';");
private async Task<List<long>> ChainSeqsAsync(DateTime month)
{
await using var conn = new SqlConnection(_fixture.ConnectionString);
await conn.OpenAsync();
await using var cmd = conn.CreateCommand();
cmd.CommandText = "SELECT ChainSeq FROM dbo.AuditLog WHERE OccurredAtUtc >= @s AND OccurredAtUtc < @e ORDER BY ChainSeq;";
cmd.Parameters.Add(new SqlParameter("@s", System.Data.SqlDbType.DateTime2) { Value = month });
cmd.Parameters.Add(new SqlParameter("@e", System.Data.SqlDbType.DateTime2) { Value = month.AddMonths(1) });
var list = new List<long>();
await using var reader = await cmd.ExecuteReaderAsync();
while (await reader.ReadAsync())
{
list.Add(reader.GetInt64(0));
}
return list;
}
private async Task ExecAsync(string sql)
{
await using var conn = new SqlConnection(_fixture.ConnectionString);
await conn.OpenAsync();
await using var cmd = conn.CreateCommand();
cmd.CommandText = sql;
await cmd.ExecuteNonQueryAsync();
}
private async Task<T> ScalarAsync<T>(string sql)
{
await using var conn = new SqlConnection(_fixture.ConnectionString);
await conn.OpenAsync();
await using var cmd = conn.CreateCommand();
cmd.CommandText = sql;
return (T)(await cmd.ExecuteScalarAsync())!;
}
private ScadaBridgeDbContext CreateContext()
=> new(new DbContextOptionsBuilder<ScadaBridgeDbContext>().UseSqlServer(_fixture.ConnectionString).Options);
private static AuditEvent NewEvent(
DateTime occurredAtUtc,
AuditChannel channel = AuditChannel.ApiOutbound,
AuditKind kind = AuditKind.ApiCall,
string? sourceNode = "central-a") =>
ScadaBridgeAuditEventFactory.Create(
channel: channel,
kind: kind,
status: AuditStatus.Delivered,
occurredAtUtc: occurredAtUtc,
target: "erp",
sourceNode: sourceNode,
sourceSiteId: "hash-chain-test");
}