Issue #10: implement worker process launcher #60
@@ -0,0 +1,62 @@
|
||||
# Worker Process Launcher
|
||||
|
||||
The gateway uses `WorkerProcessLauncher` to validate and start one worker
|
||||
process for a gateway session. The launcher owns process start semantics only;
|
||||
pipe handshaking and `WorkerReady` validation remain part of the worker client
|
||||
startup path.
|
||||
|
||||
## Launch Inputs
|
||||
|
||||
`WorkerProcessLaunchRequest` carries the per-session bootstrap values:
|
||||
|
||||
- `SessionId`,
|
||||
- `PipeName`,
|
||||
- `ProtocolVersion`,
|
||||
- `Nonce`,
|
||||
- optional `PipeReservation` cleanup handle.
|
||||
|
||||
The launcher passes `SessionId`, `PipeName`, and `ProtocolVersion` as command
|
||||
line arguments:
|
||||
|
||||
```text
|
||||
--session-id <sessionId> --pipe-name <pipeName> --protocol-version <version>
|
||||
```
|
||||
|
||||
The launcher sets the nonce through the `MXGATEWAY_WORKER_NONCE` environment
|
||||
variable. The nonce is not included in `WorkerProcessCommandLine` so logs and
|
||||
diagnostics can report the launch command without exposing the secret.
|
||||
|
||||
## Validation And Cleanup
|
||||
|
||||
Before starting the process, the launcher validates that the configured worker
|
||||
path exists, has a `.exe` extension, contains a valid Windows Portable
|
||||
Executable header, and matches the configured `RequiredArchitecture`.
|
||||
|
||||
After the process starts, `IWorkerStartupProbe` waits for startup readiness.
|
||||
The default probe only verifies that the worker did not exit immediately. The
|
||||
worker client replaces this probe when pipe connection, hello, and
|
||||
`WorkerReady` handling are implemented.
|
||||
|
||||
If startup fails or exceeds `WorkerOptions.StartupTimeoutSeconds`, the launcher
|
||||
kills the worker process tree, disposes the process handle, disposes the
|
||||
optional pipe reservation, records a worker kill metric, and reports a
|
||||
`WorkerProcessLaunchException`.
|
||||
|
||||
## Verification
|
||||
|
||||
Run the focused launcher tests after changing process launch behavior:
|
||||
|
||||
```bash
|
||||
dotnet test src/MxGateway.Tests/MxGateway.Tests.csproj --filter WorkerProcessLauncherTests
|
||||
```
|
||||
|
||||
Run the gateway build because the launcher is part of `MxGateway.Server`:
|
||||
|
||||
```bash
|
||||
dotnet build src/MxGateway.Server/MxGateway.Server.csproj
|
||||
```
|
||||
|
||||
## Related Documentation
|
||||
|
||||
- [Gateway Process Detailed Design](./gateway-process-design.md)
|
||||
- [Worker Frame Protocol](./WorkerFrameProtocol.md)
|
||||
@@ -360,6 +360,15 @@ Before launch, validate:
|
||||
- worker file version or product version is acceptable,
|
||||
- worker is expected to be x86.
|
||||
|
||||
`WorkerProcessLauncher` implements the first validation layer now: it resolves
|
||||
the worker executable path, requires a `.exe`, validates the Windows Portable
|
||||
Executable header, and verifies the configured processor architecture. It passes
|
||||
only `--session-id`, `--pipe-name`, and `--protocol-version` on the command
|
||||
line. The per-session nonce is set through `MXGATEWAY_WORKER_NONCE` so the
|
||||
command line remains safe to log. Startup failures and startup timeouts kill and
|
||||
dispose the worker process and the pre-created pipe reservation before the
|
||||
session manager observes the failure.
|
||||
|
||||
## Worker IPC
|
||||
|
||||
The gateway creates the pipe server before launching the worker.
|
||||
|
||||
@@ -47,6 +47,8 @@ Detailed follow-up docs:
|
||||
security, observability, and test strategy.
|
||||
- `docs/WorkerFrameProtocol.md` covers the gateway-side named-pipe frame
|
||||
reader/writer and `WorkerEnvelope` validation rules.
|
||||
- `docs/WorkerProcessLauncher.md` covers worker executable validation, process
|
||||
launch arguments, nonce handling, and startup cleanup behavior.
|
||||
- `docs/mxaccess-worker-instance-design.md` covers each .NET Framework 4.8 x86
|
||||
MXAccess worker instance, including STA ownership, message pumping, COM
|
||||
lifetime, command dispatch, event sinks, conversion, and shutdown.
|
||||
|
||||
@@ -3,6 +3,7 @@ using MxGateway.Server.Configuration;
|
||||
using MxGateway.Server.Diagnostics;
|
||||
using MxGateway.Server.Metrics;
|
||||
using MxGateway.Server.Security.Authentication;
|
||||
using MxGateway.Server.Workers;
|
||||
|
||||
namespace MxGateway.Server;
|
||||
|
||||
@@ -27,6 +28,7 @@ public static class GatewayApplication
|
||||
builder.Services.AddSqliteAuthStore();
|
||||
builder.Services.AddHealthChecks();
|
||||
builder.Services.AddSingleton<GatewayMetrics>();
|
||||
builder.Services.AddWorkerProcessLauncher();
|
||||
|
||||
return builder;
|
||||
}
|
||||
|
||||
@@ -0,0 +1,14 @@
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public interface IWorkerProcess : IDisposable
|
||||
{
|
||||
int Id { get; }
|
||||
|
||||
bool HasExited { get; }
|
||||
|
||||
int? ExitCode { get; }
|
||||
|
||||
ValueTask WaitForExitAsync(CancellationToken cancellationToken);
|
||||
|
||||
void Kill(bool entireProcessTree);
|
||||
}
|
||||
@@ -0,0 +1,8 @@
|
||||
using System.Diagnostics;
|
||||
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public interface IWorkerProcessFactory
|
||||
{
|
||||
IWorkerProcess Start(ProcessStartInfo startInfo);
|
||||
}
|
||||
@@ -0,0 +1,8 @@
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public interface IWorkerProcessLauncher
|
||||
{
|
||||
Task<WorkerProcessHandle> LaunchAsync(
|
||||
WorkerProcessLaunchRequest request,
|
||||
CancellationToken cancellationToken = default);
|
||||
}
|
||||
@@ -0,0 +1,9 @@
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public interface IWorkerStartupProbe
|
||||
{
|
||||
Task WaitUntilReadyAsync(
|
||||
IWorkerProcess process,
|
||||
WorkerProcessLaunchRequest request,
|
||||
CancellationToken cancellationToken);
|
||||
}
|
||||
@@ -0,0 +1,27 @@
|
||||
using System.Diagnostics;
|
||||
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
internal sealed class SystemWorkerProcess(Process process) : IWorkerProcess
|
||||
{
|
||||
public int Id => process.Id;
|
||||
|
||||
public bool HasExited => process.HasExited;
|
||||
|
||||
public int? ExitCode => process.HasExited ? process.ExitCode : null;
|
||||
|
||||
public async ValueTask WaitForExitAsync(CancellationToken cancellationToken)
|
||||
{
|
||||
await process.WaitForExitAsync(cancellationToken).ConfigureAwait(false);
|
||||
}
|
||||
|
||||
public void Kill(bool entireProcessTree)
|
||||
{
|
||||
process.Kill(entireProcessTree);
|
||||
}
|
||||
|
||||
public void Dispose()
|
||||
{
|
||||
process.Dispose();
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,22 @@
|
||||
using System.Diagnostics;
|
||||
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public sealed class SystemWorkerProcessFactory : IWorkerProcessFactory
|
||||
{
|
||||
public IWorkerProcess Start(ProcessStartInfo startInfo)
|
||||
{
|
||||
Process process = new()
|
||||
{
|
||||
StartInfo = startInfo,
|
||||
};
|
||||
|
||||
if (!process.Start())
|
||||
{
|
||||
process.Dispose();
|
||||
throw new InvalidOperationException("Worker process failed to start.");
|
||||
}
|
||||
|
||||
return new SystemWorkerProcess(process);
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,80 @@
|
||||
using System.Buffers.Binary;
|
||||
using MxGateway.Server.Configuration;
|
||||
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
internal static class WorkerExecutableValidator
|
||||
{
|
||||
private const ushort ImageFileMachineI386 = 0x014c;
|
||||
private const ushort ImageFileMachineAmd64 = 0x8664;
|
||||
private const int DosHeaderSignatureOffset = 0;
|
||||
private const int PeHeaderOffsetPointer = 0x3c;
|
||||
private const int PeSignatureSize = 4;
|
||||
private const int MachineOffsetFromPeHeader = PeSignatureSize;
|
||||
private const int MinimumHeaderSize = 0x40;
|
||||
|
||||
public static void Validate(
|
||||
string executablePath,
|
||||
WorkerArchitecture requiredArchitecture)
|
||||
{
|
||||
ushort machine = ReadMachineType(executablePath);
|
||||
ushort expectedMachine = requiredArchitecture switch
|
||||
{
|
||||
WorkerArchitecture.X86 => ImageFileMachineI386,
|
||||
WorkerArchitecture.X64 => ImageFileMachineAmd64,
|
||||
_ => throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidExecutable,
|
||||
"Worker executable required architecture is unsupported."),
|
||||
};
|
||||
|
||||
if (machine != expectedMachine)
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidExecutable,
|
||||
$"Worker executable architecture does not match required {requiredArchitecture} architecture.");
|
||||
}
|
||||
}
|
||||
|
||||
private static ushort ReadMachineType(string executablePath)
|
||||
{
|
||||
byte[] header = new byte[MinimumHeaderSize];
|
||||
using FileStream stream = File.OpenRead(executablePath);
|
||||
if (stream.Read(header) < header.Length)
|
||||
{
|
||||
throw InvalidExecutable("Worker executable is too small to contain a valid PE header.");
|
||||
}
|
||||
|
||||
if (header[DosHeaderSignatureOffset] != 'M' || header[DosHeaderSignatureOffset + 1] != 'Z')
|
||||
{
|
||||
throw InvalidExecutable("Worker executable does not contain an MZ header.");
|
||||
}
|
||||
|
||||
int peHeaderOffset = BinaryPrimitives.ReadInt32LittleEndian(header.AsSpan(PeHeaderOffsetPointer, sizeof(int)));
|
||||
if (peHeaderOffset < MinimumHeaderSize)
|
||||
{
|
||||
throw InvalidExecutable("Worker executable PE header offset is invalid.");
|
||||
}
|
||||
|
||||
byte[] peHeaderBytes = new byte[PeSignatureSize + sizeof(ushort)];
|
||||
stream.Position = peHeaderOffset;
|
||||
if (stream.Read(peHeaderBytes) < peHeaderBytes.Length)
|
||||
{
|
||||
throw InvalidExecutable("Worker executable PE header is missing.");
|
||||
}
|
||||
|
||||
if (peHeaderBytes[0] != 'P' || peHeaderBytes[1] != 'E' || peHeaderBytes[2] != 0 || peHeaderBytes[3] != 0)
|
||||
{
|
||||
throw InvalidExecutable("Worker executable does not contain a PE header.");
|
||||
}
|
||||
|
||||
return BinaryPrimitives.ReadUInt16LittleEndian(
|
||||
peHeaderBytes.AsSpan(MachineOffsetFromPeHeader, sizeof(ushort)));
|
||||
}
|
||||
|
||||
private static WorkerProcessLaunchException InvalidExecutable(string message)
|
||||
{
|
||||
return new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidExecutable,
|
||||
message);
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,30 @@
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public sealed class WorkerProcessCommandLine
|
||||
{
|
||||
public WorkerProcessCommandLine(
|
||||
string executablePath,
|
||||
IReadOnlyList<string> arguments)
|
||||
{
|
||||
ExecutablePath = executablePath;
|
||||
Arguments = arguments;
|
||||
}
|
||||
|
||||
public string ExecutablePath { get; }
|
||||
|
||||
public IReadOnlyList<string> Arguments { get; }
|
||||
|
||||
public override string ToString()
|
||||
{
|
||||
return string.Join(
|
||||
" ",
|
||||
new[] { Quote(ExecutablePath) }.Concat(Arguments.Select(Quote)));
|
||||
}
|
||||
|
||||
private static string Quote(string value)
|
||||
{
|
||||
return value.Contains(' ', StringComparison.Ordinal)
|
||||
? $"\"{value}\""
|
||||
: value;
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,28 @@
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public sealed class WorkerProcessHandle : IDisposable
|
||||
{
|
||||
public WorkerProcessHandle(
|
||||
IWorkerProcess process,
|
||||
WorkerProcessCommandLine commandLine,
|
||||
DateTimeOffset launchedAt)
|
||||
{
|
||||
Process = process;
|
||||
ProcessId = process.Id;
|
||||
CommandLine = commandLine;
|
||||
LaunchedAt = launchedAt;
|
||||
}
|
||||
|
||||
public IWorkerProcess Process { get; }
|
||||
|
||||
public int ProcessId { get; }
|
||||
|
||||
public WorkerProcessCommandLine CommandLine { get; }
|
||||
|
||||
public DateTimeOffset LaunchedAt { get; }
|
||||
|
||||
public void Dispose()
|
||||
{
|
||||
Process.Dispose();
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,13 @@
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public enum WorkerProcessLaunchErrorCode
|
||||
{
|
||||
Unknown = 0,
|
||||
InvalidRequest = 1,
|
||||
ExecutableNotFound = 2,
|
||||
InvalidExecutable = 3,
|
||||
InvalidWorkingDirectory = 4,
|
||||
StartFailed = 5,
|
||||
StartupTimeout = 6,
|
||||
StartupFailed = 7,
|
||||
}
|
||||
@@ -0,0 +1,23 @@
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public sealed class WorkerProcessLaunchException : Exception
|
||||
{
|
||||
public WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode errorCode,
|
||||
string message)
|
||||
: base(message)
|
||||
{
|
||||
ErrorCode = errorCode;
|
||||
}
|
||||
|
||||
public WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode errorCode,
|
||||
string message,
|
||||
Exception innerException)
|
||||
: base(message, innerException)
|
||||
{
|
||||
ErrorCode = errorCode;
|
||||
}
|
||||
|
||||
public WorkerProcessLaunchErrorCode ErrorCode { get; }
|
||||
}
|
||||
@@ -0,0 +1,8 @@
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public sealed record WorkerProcessLaunchRequest(
|
||||
string SessionId,
|
||||
string PipeName,
|
||||
uint ProtocolVersion,
|
||||
string Nonce,
|
||||
IDisposable? PipeReservation = null);
|
||||
@@ -0,0 +1,262 @@
|
||||
using System.Diagnostics;
|
||||
using Microsoft.Extensions.Options;
|
||||
using MxGateway.Server.Configuration;
|
||||
using MxGateway.Server.Metrics;
|
||||
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public sealed class WorkerProcessLauncher : IWorkerProcessLauncher
|
||||
{
|
||||
public const string WorkerNonceEnvironmentVariableName = "MXGATEWAY_WORKER_NONCE";
|
||||
|
||||
private readonly IWorkerProcessFactory _processFactory;
|
||||
private readonly IWorkerStartupProbe _startupProbe;
|
||||
private readonly GatewayMetrics _metrics;
|
||||
private readonly TimeProvider _timeProvider;
|
||||
private readonly WorkerOptions _workerOptions;
|
||||
|
||||
public WorkerProcessLauncher(
|
||||
IOptions<GatewayOptions> gatewayOptions,
|
||||
IWorkerProcessFactory processFactory,
|
||||
IWorkerStartupProbe startupProbe,
|
||||
GatewayMetrics metrics,
|
||||
TimeProvider? timeProvider = null)
|
||||
{
|
||||
ArgumentNullException.ThrowIfNull(gatewayOptions);
|
||||
ArgumentNullException.ThrowIfNull(processFactory);
|
||||
ArgumentNullException.ThrowIfNull(startupProbe);
|
||||
ArgumentNullException.ThrowIfNull(metrics);
|
||||
|
||||
_workerOptions = gatewayOptions.Value.Worker;
|
||||
_processFactory = processFactory;
|
||||
_startupProbe = startupProbe;
|
||||
_metrics = metrics;
|
||||
_timeProvider = timeProvider ?? TimeProvider.System;
|
||||
}
|
||||
|
||||
public async Task<WorkerProcessHandle> LaunchAsync(
|
||||
WorkerProcessLaunchRequest request,
|
||||
CancellationToken cancellationToken = default)
|
||||
{
|
||||
try
|
||||
{
|
||||
return await LaunchCoreAsync(request, cancellationToken).ConfigureAwait(false);
|
||||
}
|
||||
catch
|
||||
{
|
||||
request.PipeReservation?.Dispose();
|
||||
throw;
|
||||
}
|
||||
}
|
||||
|
||||
private async Task<WorkerProcessHandle> LaunchCoreAsync(
|
||||
WorkerProcessLaunchRequest request,
|
||||
CancellationToken cancellationToken)
|
||||
{
|
||||
ValidateRequest(request);
|
||||
|
||||
DateTimeOffset startedAt = _timeProvider.GetUtcNow();
|
||||
ProcessStartInfo startInfo = CreateStartInfo(request, out WorkerProcessCommandLine commandLine);
|
||||
|
||||
IWorkerProcess process;
|
||||
try
|
||||
{
|
||||
process = _processFactory.Start(startInfo);
|
||||
}
|
||||
catch (Exception exception) when (exception is not WorkerProcessLaunchException)
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.StartFailed,
|
||||
"Worker process failed to start.",
|
||||
exception);
|
||||
}
|
||||
|
||||
try
|
||||
{
|
||||
using CancellationTokenSource startupTimeout = CancellationTokenSource.CreateLinkedTokenSource(cancellationToken);
|
||||
startupTimeout.CancelAfter(TimeSpan.FromSeconds(_workerOptions.StartupTimeoutSeconds));
|
||||
|
||||
await _startupProbe
|
||||
.WaitUntilReadyAsync(process, request, startupTimeout.Token)
|
||||
.ConfigureAwait(false);
|
||||
|
||||
_metrics.WorkerStarted(_timeProvider.GetUtcNow() - startedAt);
|
||||
|
||||
return new WorkerProcessHandle(process, commandLine, startedAt);
|
||||
}
|
||||
catch (OperationCanceledException exception) when (!cancellationToken.IsCancellationRequested)
|
||||
{
|
||||
KillAndDispose(process, "StartupTimeout");
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.StartupTimeout,
|
||||
"Worker process did not complete startup before the configured timeout.",
|
||||
exception);
|
||||
}
|
||||
catch (OperationCanceledException)
|
||||
{
|
||||
KillAndDispose(process, "LaunchCanceled");
|
||||
throw;
|
||||
}
|
||||
catch (Exception exception) when (exception is not WorkerProcessLaunchException)
|
||||
{
|
||||
KillAndDispose(process, "StartupFailed");
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.StartupFailed,
|
||||
"Worker process failed during startup.",
|
||||
exception);
|
||||
}
|
||||
catch (WorkerProcessLaunchException)
|
||||
{
|
||||
KillAndDispose(process, "StartupFailed");
|
||||
throw;
|
||||
}
|
||||
}
|
||||
|
||||
private ProcessStartInfo CreateStartInfo(
|
||||
WorkerProcessLaunchRequest request,
|
||||
out WorkerProcessCommandLine commandLine)
|
||||
{
|
||||
string executablePath = ResolveExecutablePath();
|
||||
string workingDirectory = ResolveWorkingDirectory(executablePath);
|
||||
string[] arguments =
|
||||
[
|
||||
"--session-id",
|
||||
request.SessionId,
|
||||
"--pipe-name",
|
||||
request.PipeName,
|
||||
"--protocol-version",
|
||||
request.ProtocolVersion.ToString(System.Globalization.CultureInfo.InvariantCulture),
|
||||
];
|
||||
|
||||
ProcessStartInfo startInfo = new()
|
||||
{
|
||||
FileName = executablePath,
|
||||
WorkingDirectory = workingDirectory,
|
||||
UseShellExecute = false,
|
||||
CreateNoWindow = true,
|
||||
ErrorDialog = false,
|
||||
};
|
||||
|
||||
foreach (string argument in arguments)
|
||||
{
|
||||
startInfo.ArgumentList.Add(argument);
|
||||
}
|
||||
|
||||
startInfo.Environment[WorkerNonceEnvironmentVariableName] = request.Nonce;
|
||||
|
||||
commandLine = new WorkerProcessCommandLine(executablePath, arguments);
|
||||
|
||||
return startInfo;
|
||||
}
|
||||
|
||||
private string ResolveExecutablePath()
|
||||
{
|
||||
string executablePath;
|
||||
try
|
||||
{
|
||||
executablePath = Path.GetFullPath(_workerOptions.ExecutablePath);
|
||||
}
|
||||
catch (Exception exception) when (exception is ArgumentException or NotSupportedException or PathTooLongException)
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidExecutable,
|
||||
"Worker executable path is not a valid filesystem path.",
|
||||
exception);
|
||||
}
|
||||
|
||||
if (!string.Equals(Path.GetExtension(executablePath), ".exe", StringComparison.OrdinalIgnoreCase))
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidExecutable,
|
||||
"Worker executable path must point to a .exe file.");
|
||||
}
|
||||
|
||||
if (!File.Exists(executablePath))
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.ExecutableNotFound,
|
||||
"Worker executable does not exist.");
|
||||
}
|
||||
|
||||
WorkerExecutableValidator.Validate(executablePath, _workerOptions.RequiredArchitecture);
|
||||
|
||||
return executablePath;
|
||||
}
|
||||
|
||||
private string ResolveWorkingDirectory(string executablePath)
|
||||
{
|
||||
if (string.IsNullOrWhiteSpace(_workerOptions.WorkingDirectory))
|
||||
{
|
||||
return Path.GetDirectoryName(executablePath) ?? Environment.CurrentDirectory;
|
||||
}
|
||||
|
||||
string workingDirectory;
|
||||
try
|
||||
{
|
||||
workingDirectory = Path.GetFullPath(_workerOptions.WorkingDirectory);
|
||||
}
|
||||
catch (Exception exception) when (exception is ArgumentException or NotSupportedException or PathTooLongException)
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidWorkingDirectory,
|
||||
"Worker working directory is not a valid filesystem path.",
|
||||
exception);
|
||||
}
|
||||
|
||||
if (!Directory.Exists(workingDirectory))
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidWorkingDirectory,
|
||||
"Worker working directory does not exist.");
|
||||
}
|
||||
|
||||
return workingDirectory;
|
||||
}
|
||||
|
||||
private void KillAndDispose(IWorkerProcess process, string reason)
|
||||
{
|
||||
try
|
||||
{
|
||||
if (!process.HasExited)
|
||||
{
|
||||
process.Kill(entireProcessTree: true);
|
||||
_metrics.WorkerKilled(reason);
|
||||
}
|
||||
}
|
||||
finally
|
||||
{
|
||||
process.Dispose();
|
||||
}
|
||||
}
|
||||
|
||||
private static void ValidateRequest(WorkerProcessLaunchRequest request)
|
||||
{
|
||||
if (string.IsNullOrWhiteSpace(request.SessionId))
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidRequest,
|
||||
"Worker launch requires a session id.");
|
||||
}
|
||||
|
||||
if (string.IsNullOrWhiteSpace(request.PipeName))
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidRequest,
|
||||
"Worker launch requires a pipe name.");
|
||||
}
|
||||
|
||||
if (request.ProtocolVersion == 0)
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidRequest,
|
||||
"Worker launch requires a non-zero protocol version.");
|
||||
}
|
||||
|
||||
if (string.IsNullOrWhiteSpace(request.Nonce))
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.InvalidRequest,
|
||||
"Worker launch requires a nonce.");
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,19 @@
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public sealed class WorkerProcessStartedProbe : IWorkerStartupProbe
|
||||
{
|
||||
public Task WaitUntilReadyAsync(
|
||||
IWorkerProcess process,
|
||||
WorkerProcessLaunchRequest request,
|
||||
CancellationToken cancellationToken)
|
||||
{
|
||||
if (process.HasExited)
|
||||
{
|
||||
throw new WorkerProcessLaunchException(
|
||||
WorkerProcessLaunchErrorCode.StartupFailed,
|
||||
$"Worker process exited before startup completed with exit code {process.ExitCode}.");
|
||||
}
|
||||
|
||||
return Task.CompletedTask;
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,13 @@
|
||||
namespace MxGateway.Server.Workers;
|
||||
|
||||
public static class WorkerServiceCollectionExtensions
|
||||
{
|
||||
public static IServiceCollection AddWorkerProcessLauncher(this IServiceCollection services)
|
||||
{
|
||||
services.AddSingleton<IWorkerProcessFactory, SystemWorkerProcessFactory>();
|
||||
services.AddSingleton<IWorkerStartupProbe, WorkerProcessStartedProbe>();
|
||||
services.AddSingleton<IWorkerProcessLauncher, WorkerProcessLauncher>();
|
||||
|
||||
return services;
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,307 @@
|
||||
using System.Diagnostics;
|
||||
using Microsoft.Extensions.Options;
|
||||
using MxGateway.Contracts;
|
||||
using MxGateway.Server.Configuration;
|
||||
using MxGateway.Server.Metrics;
|
||||
using MxGateway.Server.Workers;
|
||||
|
||||
namespace MxGateway.Tests.Gateway.Workers;
|
||||
|
||||
public sealed class WorkerProcessLauncherTests
|
||||
{
|
||||
private const string SessionId = "session-1";
|
||||
private const string PipeName = "mxaccess-gateway-123-session-1";
|
||||
private const string Nonce = "super-secret-nonce";
|
||||
|
||||
[Fact]
|
||||
public async Task LaunchAsync_WithValidWorker_StartsProcessWithBootstrapArgumentsAndNonceEnvironment()
|
||||
{
|
||||
using TestDirectory directory = TestDirectory.Create();
|
||||
string executablePath = directory.CreateWorkerExecutable(machine: 0x014c);
|
||||
FakeWorkerProcess process = new(processId: 1234);
|
||||
FakePipeReservation pipeReservation = new();
|
||||
FakeWorkerProcessFactory processFactory = new(process);
|
||||
GatewayMetrics metrics = new();
|
||||
WorkerProcessLauncher launcher = CreateLauncher(executablePath, processFactory, new SucceedingStartupProbe(), metrics);
|
||||
|
||||
using WorkerProcessHandle handle = await launcher.LaunchAsync(CreateRequest(pipeReservation));
|
||||
|
||||
Assert.Equal(1234, handle.ProcessId);
|
||||
Assert.Same(process, handle.Process);
|
||||
Assert.NotNull(processFactory.LastStartInfo);
|
||||
Assert.Equal(Path.GetFullPath(executablePath), processFactory.LastStartInfo.FileName);
|
||||
Assert.False(processFactory.LastStartInfo.UseShellExecute);
|
||||
Assert.True(processFactory.LastStartInfo.CreateNoWindow);
|
||||
Assert.Equal(
|
||||
["--session-id", SessionId, "--pipe-name", PipeName, "--protocol-version", "1"],
|
||||
processFactory.LastStartInfo.ArgumentList);
|
||||
Assert.Equal(Nonce, processFactory.LastStartInfo.Environment[WorkerProcessLauncher.WorkerNonceEnvironmentVariableName]);
|
||||
Assert.DoesNotContain(Nonce, handle.CommandLine.ToString(), StringComparison.Ordinal);
|
||||
Assert.DoesNotContain(Nonce, string.Join(" ", handle.CommandLine.Arguments), StringComparison.Ordinal);
|
||||
Assert.False(pipeReservation.DisposeCalled);
|
||||
Assert.Equal(1, metrics.GetSnapshot().WorkersRunning);
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public async Task LaunchAsync_WhenStartupProbeFails_KillsAndDisposesWorker()
|
||||
{
|
||||
using TestDirectory directory = TestDirectory.Create();
|
||||
string executablePath = directory.CreateWorkerExecutable(machine: 0x014c);
|
||||
FakeWorkerProcess process = new(processId: 1234);
|
||||
FakePipeReservation pipeReservation = new();
|
||||
GatewayMetrics metrics = new();
|
||||
WorkerProcessLauncher launcher = CreateLauncher(
|
||||
executablePath,
|
||||
new FakeWorkerProcessFactory(process),
|
||||
new FailingStartupProbe(),
|
||||
metrics);
|
||||
|
||||
WorkerProcessLaunchException exception =
|
||||
await Assert.ThrowsAsync<WorkerProcessLaunchException>(
|
||||
async () => await launcher.LaunchAsync(CreateRequest(pipeReservation)));
|
||||
|
||||
Assert.Equal(WorkerProcessLaunchErrorCode.StartupFailed, exception.ErrorCode);
|
||||
Assert.True(process.KillCalled);
|
||||
Assert.True(process.DisposeCalled);
|
||||
Assert.True(pipeReservation.DisposeCalled);
|
||||
Assert.Equal(1, metrics.GetSnapshot().WorkerKills);
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public async Task LaunchAsync_WhenStartupTimesOut_KillsAndDisposesWorker()
|
||||
{
|
||||
using TestDirectory directory = TestDirectory.Create();
|
||||
string executablePath = directory.CreateWorkerExecutable(machine: 0x014c);
|
||||
FakeWorkerProcess process = new(processId: 1234);
|
||||
GatewayMetrics metrics = new();
|
||||
WorkerProcessLauncher launcher = CreateLauncher(
|
||||
executablePath,
|
||||
new FakeWorkerProcessFactory(process),
|
||||
new WaitingStartupProbe(),
|
||||
metrics,
|
||||
startupTimeoutSeconds: 1);
|
||||
|
||||
WorkerProcessLaunchException exception =
|
||||
await Assert.ThrowsAsync<WorkerProcessLaunchException>(
|
||||
async () => await launcher.LaunchAsync(CreateRequest()));
|
||||
|
||||
Assert.Equal(WorkerProcessLaunchErrorCode.StartupTimeout, exception.ErrorCode);
|
||||
Assert.True(process.KillCalled);
|
||||
Assert.True(process.DisposeCalled);
|
||||
Assert.Equal(1, metrics.GetSnapshot().WorkerKills);
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public async Task LaunchAsync_WhenExecutableDoesNotExist_FailsBeforeStartingProcess()
|
||||
{
|
||||
using TestDirectory directory = TestDirectory.Create();
|
||||
string executablePath = Path.Combine(directory.Path, "missing-worker.exe");
|
||||
FakeWorkerProcessFactory processFactory = new(new FakeWorkerProcess(processId: 1234));
|
||||
WorkerProcessLauncher launcher = CreateLauncher(executablePath, processFactory, new SucceedingStartupProbe());
|
||||
|
||||
WorkerProcessLaunchException exception =
|
||||
await Assert.ThrowsAsync<WorkerProcessLaunchException>(
|
||||
async () => await launcher.LaunchAsync(CreateRequest()));
|
||||
|
||||
Assert.Equal(WorkerProcessLaunchErrorCode.ExecutableNotFound, exception.ErrorCode);
|
||||
Assert.Null(processFactory.LastStartInfo);
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public async Task LaunchAsync_WhenExecutableArchitectureDoesNotMatch_FailsBeforeStartingProcess()
|
||||
{
|
||||
using TestDirectory directory = TestDirectory.Create();
|
||||
string executablePath = directory.CreateWorkerExecutable(machine: 0x8664);
|
||||
FakeWorkerProcessFactory processFactory = new(new FakeWorkerProcess(processId: 1234));
|
||||
WorkerProcessLauncher launcher = CreateLauncher(executablePath, processFactory, new SucceedingStartupProbe());
|
||||
|
||||
WorkerProcessLaunchException exception =
|
||||
await Assert.ThrowsAsync<WorkerProcessLaunchException>(
|
||||
async () => await launcher.LaunchAsync(CreateRequest()));
|
||||
|
||||
Assert.Equal(WorkerProcessLaunchErrorCode.InvalidExecutable, exception.ErrorCode);
|
||||
Assert.Null(processFactory.LastStartInfo);
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public async Task LaunchAsync_WhenWorkerAlreadyExited_FailsAndDisposesWorkerWithoutKill()
|
||||
{
|
||||
using TestDirectory directory = TestDirectory.Create();
|
||||
string executablePath = directory.CreateWorkerExecutable(machine: 0x014c);
|
||||
FakeWorkerProcess process = new(processId: 1234)
|
||||
{
|
||||
HasExited = true,
|
||||
ExitCode = 42,
|
||||
};
|
||||
WorkerProcessLauncher launcher = CreateLauncher(
|
||||
executablePath,
|
||||
new FakeWorkerProcessFactory(process),
|
||||
new WorkerProcessStartedProbe());
|
||||
|
||||
WorkerProcessLaunchException exception =
|
||||
await Assert.ThrowsAsync<WorkerProcessLaunchException>(
|
||||
async () => await launcher.LaunchAsync(CreateRequest()));
|
||||
|
||||
Assert.Equal(WorkerProcessLaunchErrorCode.StartupFailed, exception.ErrorCode);
|
||||
Assert.False(process.KillCalled);
|
||||
Assert.True(process.DisposeCalled);
|
||||
}
|
||||
|
||||
private static WorkerProcessLauncher CreateLauncher(
|
||||
string executablePath,
|
||||
IWorkerProcessFactory processFactory,
|
||||
IWorkerStartupProbe startupProbe,
|
||||
GatewayMetrics? metrics = null,
|
||||
int startupTimeoutSeconds = 30)
|
||||
{
|
||||
GatewayOptions options = new()
|
||||
{
|
||||
Worker = new WorkerOptions
|
||||
{
|
||||
ExecutablePath = executablePath,
|
||||
RequiredArchitecture = WorkerArchitecture.X86,
|
||||
StartupTimeoutSeconds = startupTimeoutSeconds,
|
||||
},
|
||||
};
|
||||
|
||||
return new WorkerProcessLauncher(
|
||||
Options.Create(options),
|
||||
processFactory,
|
||||
startupProbe,
|
||||
metrics ?? new GatewayMetrics());
|
||||
}
|
||||
|
||||
private static WorkerProcessLaunchRequest CreateRequest(IDisposable? pipeReservation = null)
|
||||
{
|
||||
return new WorkerProcessLaunchRequest(
|
||||
SessionId,
|
||||
PipeName,
|
||||
GatewayContractInfo.WorkerProtocolVersion,
|
||||
Nonce,
|
||||
pipeReservation);
|
||||
}
|
||||
|
||||
private sealed class FakeWorkerProcessFactory(IWorkerProcess process) : IWorkerProcessFactory
|
||||
{
|
||||
public ProcessStartInfo? LastStartInfo { get; private set; }
|
||||
|
||||
public IWorkerProcess Start(ProcessStartInfo startInfo)
|
||||
{
|
||||
LastStartInfo = startInfo;
|
||||
return process;
|
||||
}
|
||||
}
|
||||
|
||||
private sealed class FakeWorkerProcess(int processId) : IWorkerProcess
|
||||
{
|
||||
public int Id { get; } = processId;
|
||||
|
||||
public bool HasExited { get; set; }
|
||||
|
||||
public int? ExitCode { get; set; }
|
||||
|
||||
public bool DisposeCalled { get; private set; }
|
||||
|
||||
public bool KillCalled { get; private set; }
|
||||
|
||||
public ValueTask WaitForExitAsync(CancellationToken cancellationToken)
|
||||
{
|
||||
return ValueTask.CompletedTask;
|
||||
}
|
||||
|
||||
public void Kill(bool entireProcessTree)
|
||||
{
|
||||
Assert.True(entireProcessTree);
|
||||
KillCalled = true;
|
||||
HasExited = true;
|
||||
}
|
||||
|
||||
public void Dispose()
|
||||
{
|
||||
DisposeCalled = true;
|
||||
}
|
||||
}
|
||||
|
||||
private sealed class SucceedingStartupProbe : IWorkerStartupProbe
|
||||
{
|
||||
public Task WaitUntilReadyAsync(
|
||||
IWorkerProcess process,
|
||||
WorkerProcessLaunchRequest request,
|
||||
CancellationToken cancellationToken)
|
||||
{
|
||||
return Task.CompletedTask;
|
||||
}
|
||||
}
|
||||
|
||||
private sealed class FailingStartupProbe : IWorkerStartupProbe
|
||||
{
|
||||
public Task WaitUntilReadyAsync(
|
||||
IWorkerProcess process,
|
||||
WorkerProcessLaunchRequest request,
|
||||
CancellationToken cancellationToken)
|
||||
{
|
||||
throw new InvalidOperationException("Fake worker startup failed.");
|
||||
}
|
||||
}
|
||||
|
||||
private sealed class WaitingStartupProbe : IWorkerStartupProbe
|
||||
{
|
||||
public async Task WaitUntilReadyAsync(
|
||||
IWorkerProcess process,
|
||||
WorkerProcessLaunchRequest request,
|
||||
CancellationToken cancellationToken)
|
||||
{
|
||||
await Task.Delay(Timeout.InfiniteTimeSpan, cancellationToken);
|
||||
}
|
||||
}
|
||||
|
||||
private sealed class FakePipeReservation : IDisposable
|
||||
{
|
||||
public bool DisposeCalled { get; private set; }
|
||||
|
||||
public void Dispose()
|
||||
{
|
||||
DisposeCalled = true;
|
||||
}
|
||||
}
|
||||
|
||||
private sealed class TestDirectory : IDisposable
|
||||
{
|
||||
private TestDirectory(string path)
|
||||
{
|
||||
Path = path;
|
||||
}
|
||||
|
||||
public string Path { get; }
|
||||
|
||||
public static TestDirectory Create()
|
||||
{
|
||||
string path = System.IO.Path.Combine(System.IO.Path.GetTempPath(), $"mxgateway-tests-{Guid.NewGuid():N}");
|
||||
Directory.CreateDirectory(path);
|
||||
|
||||
return new TestDirectory(path);
|
||||
}
|
||||
|
||||
public string CreateWorkerExecutable(ushort machine)
|
||||
{
|
||||
string path = System.IO.Path.Combine(Path, "MxGateway.Worker.exe");
|
||||
byte[] bytes = new byte[0x100];
|
||||
bytes[0] = (byte)'M';
|
||||
bytes[1] = (byte)'Z';
|
||||
BitConverter.GetBytes(0x80).CopyTo(bytes, 0x3c);
|
||||
bytes[0x80] = (byte)'P';
|
||||
bytes[0x81] = (byte)'E';
|
||||
bytes[0x82] = 0;
|
||||
bytes[0x83] = 0;
|
||||
BitConverter.GetBytes(machine).CopyTo(bytes, 0x84);
|
||||
File.WriteAllBytes(path, bytes);
|
||||
|
||||
return path;
|
||||
}
|
||||
|
||||
public void Dispose()
|
||||
{
|
||||
Directory.Delete(Path, recursive: true);
|
||||
}
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user