Joseph Doherty
ec3b835e4a
M1 IsPoison classified every SQL Server transient as poison: Microsoft.Data.SqlClient never sets DbException.IsTransient, so a deadlock (1205), timeout (-2 / Win32 258), dropped connection or AG failover was counted from memory and the node's row left Applying. Poison is now an ALLOWLIST: a DbUpdateException (not concurrency) carrying a SqlException whose number is 8152, 2628, 547, 515, 2627 or 2601. Everything else retries, bounded by the deadline deferral cap. IsPoison is public for tests. Acks counted only from memory (poison branch and the deferral cap) are kept in _memoryOnlyAcks and re-sent by PreRestart alongside _unpersisted, so a restart before the terminal write does not lose them. M2 Revisions adopted without an outcome-reporting apply are unconfirmed, so a same-revision dispatch re-applies instead of short-circuiting a clean Applied over an empty address space: TryRecoverFromStale (restores only the revision marker) and RestoreApplied when the artifact read fails (null blob, LoadArtifactBlob swallows the error) or the restore throws. L1 The re-drive comment documents the CreatedAtUtc recompute after a second restart / failover. L2 _currentRevisionDegraded renamed _currentRevisionUnconfirmed; the stale TryBootFromCache remark corrected. L3 docs/Redundancy.md: a Direct node running from cache that still cannot reach the ConfigDb acks a same-revision re-dispatch Failed (PartiallyFailed, not Sealed); Stale recovery and failed restores are unconfirmed too. Test harness: SqlExceptionFactory builds real SqlExceptions (reflection on SqlError / SqlErrorCollection / SqlException.CreateException); ConstraintEnforcingDbFactory now throws DbUpdateException over SQL 2628 (over-length), 547 (poison) and 1205 (transient), plus FailNextContexts. Tests (all fail on the previous code): poison allowlist theory (1205, 40613, 10054, 4060 not poison) + timeout/concurrency/providerless not poison; a deadlocked ack write is retried and recorded; a memory-counted ack survives a coordinator restart; Stale-recovered and failed-restore revisions re-apply on re-dispatch. Claude-Session: https://claude.ai/code/session_01GT67AD99zQQqpLCVm3p7xk
64 lines
3.0 KiB
C#
64 lines
3.0 KiB
C#
using System.ComponentModel;
|
|
using System.Reflection;
|
|
using Microsoft.Data.SqlClient;
|
|
|
|
namespace ZB.MOM.WW.OtOpcUa.ControlPlane.Tests.Harness;
|
|
|
|
/// <summary>
|
|
/// Builds real <see cref="SqlException"/>s for tests. The type has no public constructor, so this goes through its
|
|
/// internal factory by reflection — the same objects SQL Server failures produce, carrying a real error number.
|
|
/// </summary>
|
|
public static class SqlExceptionFactory
|
|
{
|
|
private const BindingFlags NonPublic = BindingFlags.NonPublic | BindingFlags.Instance;
|
|
|
|
/// <summary>A <see cref="SqlException"/> with one error of the given number.</summary>
|
|
/// <param name="number">The SQL Server error number (e.g. 2628, 1205, -2).</param>
|
|
/// <param name="message">The error text.</param>
|
|
/// <param name="inner">An inner exception (e.g. a <see cref="Win32Exception"/> for a timeout).</param>
|
|
/// <returns>The exception.</returns>
|
|
public static SqlException Create(int number, string message, Exception? inner = null)
|
|
{
|
|
var errorCtor = typeof(SqlError).GetConstructors(NonPublic)
|
|
.OrderByDescending(c => c.GetParameters().Length)
|
|
.First();
|
|
var args = errorCtor.GetParameters().Select(p => p.Name switch
|
|
{
|
|
"infoNumber" => (object?)number,
|
|
"errorMessage" => message,
|
|
"server" => "test-sql",
|
|
"procedure" => string.Empty,
|
|
"exception" => inner,
|
|
"errorClass" => (byte)16,
|
|
_ => p.ParameterType == typeof(string) ? string.Empty
|
|
: p.ParameterType.IsValueType ? Activator.CreateInstance(p.ParameterType) : null,
|
|
}).ToArray();
|
|
var error = (SqlError)errorCtor.Invoke(args);
|
|
|
|
var collection = (SqlErrorCollection)Activator.CreateInstance(typeof(SqlErrorCollection), nonPublic: true)!;
|
|
typeof(SqlErrorCollection).GetMethod("Add", NonPublic)!.Invoke(collection, [error]);
|
|
|
|
var create = typeof(SqlException)
|
|
.GetMethods(BindingFlags.NonPublic | BindingFlags.Static)
|
|
.Where(m => m.Name == "CreateException")
|
|
.Select(m => (Method: m, Params: m.GetParameters()))
|
|
.Where(x => x.Params.Length >= 2 && x.Params[0].ParameterType == typeof(SqlErrorCollection)
|
|
&& x.Params[1].ParameterType == typeof(string))
|
|
.OrderBy(x => x.Params.Length)
|
|
.First();
|
|
var createArgs = create.Params.Select((p, i) => i switch
|
|
{
|
|
0 => collection,
|
|
1 => (object?)"16.0.0",
|
|
_ => p.ParameterType == typeof(Exception) ? inner
|
|
: p.ParameterType.IsValueType ? Activator.CreateInstance(p.ParameterType) : null,
|
|
}).ToArray();
|
|
return (SqlException)create.Method.Invoke(null, createArgs)!;
|
|
}
|
|
|
|
/// <summary>A command timeout as SQL Server reports it: number -2 over a Win32 258 (WAIT_TIMEOUT).</summary>
|
|
/// <returns>The exception.</returns>
|
|
public static SqlException Timeout() =>
|
|
Create(-2, "Execution Timeout Expired.", new Win32Exception(258, "The wait operation timed out."));
|
|
}
|